draft-ietf-rift-rift-02.txt   draft-ietf-rift-rift-03.txt 
RIFT Working Group T. Przygienda, Ed. RIFT Working Group The RIFT Authors
Internet-Draft Juniper Networks Internet-Draft "Heaven is under our feet as well as over our heads"
Intended status: Standards Track A. Sharma Intended status: Standards Track October 19, 2018
Expires: December 23, 2018 Comcast Expires: April 22, 2019
P. Thubert
Cisco
A. Atlas
Individual
J. Drake
Juniper Networks
Jun 21, 2018
RIFT: Routing in Fat Trees RIFT: Routing in Fat Trees
draft-ietf-rift-rift-02 draft-ietf-rift-rift-03
Abstract Abstract
This document outlines a specialized, dynamic routing protocol for This document outlines a specialized, dynamic routing protocol for
Clos and fat-tree network topologies. The protocol (1) deals with Clos and fat-tree network topologies. The protocol (1) deals with
automatic construction of fat-tree topologies based on detection of fully automatied construction of fat-tree topologies based on
links, (2) minimizes the amount of routing state held at each level, detection of links, (2) minimizes the amount of routing state held at
(3) automatically prunes the topology distribution exchanges to a each level, (3) automatically prunes and load balances topology
sufficient subset of links, (4) supports automatic disaggregation of flooding exchanges over a sufficient subset of links, (4) supports
prefixes on link and node failures to prevent black-holing and automatic disaggregation of prefixes on link and node failures to
suboptimal routing, (5) allows traffic steering and re-routing prevent black-holing and suboptimal routing, (5) allows traffic
policies, (6) allows non-ECMP forwarding, (7) automatically re- steering and re-routing policies, (6) allows loop-free non-ECMP
balances traffic towards the spines based on bandwidth available and forwarding, (7) automatically re-balances traffic towards the spines
ultimately (8) provides mechanisms to synchronize a limited key-value based on bandwidth available and finally (8) provides mechanisms to
data-store that can be used after protocol convergence to e.g. synchronize a limited key-value data-store that can be used after
bootstrap higher levels of functionality on nodes. protocol convergence to e.g. bootstrap higher levels of
functionality on nodes.
Status of This Memo Status of This Memo
This Internet-Draft is submitted in full conformance with the This Internet-Draft is submitted in full conformance with the
provisions of BCP 78 and BCP 79. provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF). Note that other groups may also distribute Task Force (IETF). Note that other groups may also distribute
working documents as Internet-Drafts. The list of current Internet- working documents as Internet-Drafts. The list of current Internet-
Drafts is at https://datatracker.ietf.org/drafts/current/. Drafts is at https://datatracker.ietf.org/drafts/current/.
Internet-Drafts are draft documents valid for a maximum of six months Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at any and may be updated, replaced, or obsoleted by other documents at any
time. It is inappropriate to use Internet-Drafts as reference time. It is inappropriate to use Internet-Drafts as reference
material or to cite them other than as "work in progress." material or to cite them other than as "work in progress."
This Internet-Draft will expire on December 23, 2018. This Internet-Draft will expire on April 22, 2019.
Copyright Notice Copyright Notice
Copyright (c) 2018 IETF Trust and the persons identified as the Copyright (c) 2018 IETF Trust and the persons identified as the
document authors. All rights reserved. document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal This document is subject to BCP 78 and the IETF Trust's Legal
Provisions Relating to IETF Documents Provisions Relating to IETF Documents
(https://trustee.ietf.org/license-info) in effect on the date of (https://trustee.ietf.org/license-info) in effect on the date of
publication of this document. Please review these documents publication of this document. Please review these documents
carefully, as they describe your rights and restrictions with respect carefully, as they describe your rights and restrictions with respect
to this document. Code Components extracted from this document must to this document. Code Components extracted from this document must
include Simplified BSD License text as described in Section 4.e of include Simplified BSD License text as described in Section 4.e of
the Trust Legal Provisions and are provided without warranty as the Trust Legal Provisions and are provided without warranty as
described in the Simplified BSD License. described in the Simplified BSD License.
Table of Contents Table of Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 4 1. Authors . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.1. Requirements Language . . . . . . . . . . . . . . . . . . 5 2. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 4
2. Reference Frame . . . . . . . . . . . . . . . . . . . . . . . 5 2.1. Requirements Language . . . . . . . . . . . . . . . . . . 6
2.1. Terminology . . . . . . . . . . . . . . . . . . . . . . . 5 3. Reference Frame . . . . . . . . . . . . . . . . . . . . . . . 6
2.2. Topology . . . . . . . . . . . . . . . . . . . . . . . . 8 3.1. Terminology . . . . . . . . . . . . . . . . . . . . . . . 6
3. Requirement Considerations . . . . . . . . . . . . . . . . . 10 3.2. Topology . . . . . . . . . . . . . . . . . . . . . . . . 10
4. RIFT: Routing in Fat Trees . . . . . . . . . . . . . . . . . 12 4. Requirement Considerations . . . . . . . . . . . . . . . . . 11
4.1. Overview . . . . . . . . . . . . . . . . . . . . . . . . 13 5. RIFT: Routing in Fat Trees . . . . . . . . . . . . . . . . . 14
4.2. Specification . . . . . . . . . . . . . . . . . . . . . . 13 5.1. Overview . . . . . . . . . . . . . . . . . . . . . . . . 15
4.2.1. Transport . . . . . . . . . . . . . . . . . . . . . . 13 5.1.1. Properties . . . . . . . . . . . . . . . . . . . . . 15
4.2.2. Link (Neighbor) Discovery (LIE Exchange) . . . . . . 13 5.1.2. Generalized Topology View . . . . . . . . . . . . . . 15
4.2.3. Topology Exchange (TIE Exchange) . . . . . . . . . . 15 5.1.3. Fallen Leaf Problem . . . . . . . . . . . . . . . . . 25
4.2.3.1. Topology Information Elements . . . . . . . . . . 15 5.1.4. Discovering Fallen Leaves . . . . . . . . . . . . . . 27
4.2.3.2. South- and Northbound Representation . . . . . . 16 5.1.5. Addressing the Fallen Leaves Problem . . . . . . . . 28
4.2.3.3. Flooding . . . . . . . . . . . . . . . . . . . . 19 5.2. Specification . . . . . . . . . . . . . . . . . . . . . . 29
4.2.3.4. TIE Flooding Scopes . . . . . . . . . . . . . . . 19 5.2.1. Transport . . . . . . . . . . . . . . . . . . . . . . 29
4.2.3.5. Initial and Periodic Database Synchronization . . 21 5.2.2. Link (Neighbor) Discovery (LIE Exchange) . . . . . . 30
4.2.3.6. Purging . . . . . . . . . . . . . . . . . . . . . 21 5.2.3. Topology Exchange (TIE Exchange) . . . . . . . . . . 32
4.2.3.7. Southbound Default Route Origination . . . . . . 22 5.2.3.1. Topology Information Elements . . . . . . . . . . 32
4.2.3.8. Northbound TIE Flooding Reduction . . . . . . . . 22 5.2.3.2. South- and Northbound Representation . . . . . . 33
4.2.4. Policy-Guided Prefixes . . . . . . . . . . . . . . . 26 5.2.3.3. Flooding . . . . . . . . . . . . . . . . . . . . 35
4.2.4.1. Ingress Filtering . . . . . . . . . . . . . . . . 27 5.2.3.4. TIE Flooding Scopes . . . . . . . . . . . . . . . 36
4.2.4.2. Applying Policy . . . . . . . . . . . . . . . . . 28 5.2.3.5. Initial and Periodic Database Synchronization . . 38
4.2.4.3. Store Policy-Guided Prefix for Route Computation 5.2.3.6. Purging . . . . . . . . . . . . . . . . . . . . . 38
and Regeneration . . . . . . . . . . . . . . . . 29 5.2.3.7. Southbound Default Route Origination . . . . . . 39
4.2.4.4. Re-origination . . . . . . . . . . . . . . . . . 29 5.2.3.8. Northbound TIE Flooding Reduction . . . . . . . . 40
4.2.4.5. Overlap with Disaggregated Prefixes . . . . . . . 30 5.2.4. Reachability Computation . . . . . . . . . . . . . . 44
4.2.5. Reachability Computation . . . . . . . . . . . . . . 30 5.2.4.1. Northbound SPF . . . . . . . . . . . . . . . . . 44
4.2.5.1. Northbound SPF . . . . . . . . . . . . . . . . . 30 5.2.4.2. Southbound SPF . . . . . . . . . . . . . . . . . 45
4.2.5.2. Southbound SPF . . . . . . . . . . . . . . . . . 31 5.2.4.3. East-West Forwarding Within a Level . . . . . . . 45
4.2.5.3. East-West Forwarding Within a Level . . . . . . . 31 5.2.5. Automatic Disaggregation on Link & Node Failures . . 45
5.2.5.1. Positive, Non-transitive Disaggregation . . . . . 45
5.2.5.2. Negative, Transitive Disaggregation for Fallen
Leafs . . . . . . . . . . . . . . . . . . . . . . 49
5.2.6. Attaching Prefixes . . . . . . . . . . . . . . . . . 51
5.2.7. Optional Zero Touch Provisioning (ZTP) . . . . . . . 53
5.2.7.1. Terminology . . . . . . . . . . . . . . . . . . . 54
5.2.7.2. Automatic SystemID Selection . . . . . . . . . . 55
5.2.7.3. Generic Fabric Example . . . . . . . . . . . . . 55
5.2.7.4. Level Determination Procedure . . . . . . . . . . 56
5.2.7.5. Resulting Topologies . . . . . . . . . . . . . . 58
5.2.8. Stability Considerations . . . . . . . . . . . . . . 59
5.3. Further Mechanisms . . . . . . . . . . . . . . . . . . . 60
5.3.1. Overload Bit . . . . . . . . . . . . . . . . . . . . 60
5.3.2. Optimized Route Computation on Leafs . . . . . . . . 60
5.3.3. Mobility . . . . . . . . . . . . . . . . . . . . . . 60
5.3.3.1. Clock Comparison . . . . . . . . . . . . . . . . 62
5.3.3.2. Interaction between Time Stamps and Sequence
Counters . . . . . . . . . . . . . . . . . . . . 62
5.3.3.3. Anycast vs. Unicast . . . . . . . . . . . . . . . 63
5.3.3.4. Overlays and Signaling . . . . . . . . . . . . . 63
5.3.4. Key/Value Store . . . . . . . . . . . . . . . . . . . 64
5.3.4.1. Southbound . . . . . . . . . . . . . . . . . . . 64
5.3.4.2. Northbound . . . . . . . . . . . . . . . . . . . 64
5.3.5. Interactions with BFD . . . . . . . . . . . . . . . . 64
5.3.6. Fabric Bandwidth Balancing . . . . . . . . . . . . . 65
5.3.6.1. Northbound Direction . . . . . . . . . . . . . . 65
5.3.6.2. Southbound Direction . . . . . . . . . . . . . . 67
5.3.7. Label Binding . . . . . . . . . . . . . . . . . . . . 68
5.3.8. Segment Routing Support with RIFT . . . . . . . . . . 68
5.3.8.1. Global Segment Identifiers Assignment . . . . . . 68
5.3.8.2. Distribution of Topology Information . . . . . . 68
5.3.9. Leaf to Leaf Procedures . . . . . . . . . . . . . . . 69
5.3.10. Address Family and Multi Topology Considerations . . 69
5.3.11. Reachability of Internal Nodes in the Fabric . . . . 69
5.3.12. One-Hop Healing of Levels with East-West Links . . . 70
6. Examples . . . . . . . . . . . . . . . . . . . . . . . . . . 70
6.1. Normal Operation . . . . . . . . . . . . . . . . . . . . 70
6.2. Leaf Link Failure . . . . . . . . . . . . . . . . . . . . 71
6.3. Partitioned Fabric . . . . . . . . . . . . . . . . . . . 72
6.4. Northbound Partitioned Router and Optional East-West
Links . . . . . . . . . . . . . . . . . . . . . . . . . . 74
6.5. Multi-Plane Fabric and Negative Disaggregation . . . . . 75
7. Implementation and Operation: Further Details . . . . . . . . 75
7.1. Considerations for Leaf-Only Implementation . . . . . . . 75
7.2. Adaptations to Other Proposed Data Center Topologies . . 76
7.3. Originating Non-Default Route Southbound . . . . . . . . 77
8. Security Considerations . . . . . . . . . . . . . . . . . . . 77
8.1. General . . . . . . . . . . . . . . . . . . . . . . . . . 77
8.2. ZTP . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
8.3. Lifetime . . . . . . . . . . . . . . . . . . . . . . . . 78
9. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 78
10. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 78
11. References . . . . . . . . . . . . . . . . . . . . . . . . . 78
11.1. Normative References . . . . . . . . . . . . . . . . . . 78
11.2. Informative References . . . . . . . . . . . . . . . . . 81
Appendix A. Information Elements Schema . . . . . . . . . . . . 83
A.1. common.thrift . . . . . . . . . . . . . . . . . . . . . . 84
A.2. encoding.thrift . . . . . . . . . . . . . . . . . . . . . 88
Appendix B. Finite State Machines and Precise Operational
Specifications . . . . . . . . . . . . . . . . . . . 94
B.1. LIE FSM . . . . . . . . . . . . . . . . . . . . . . . . . 95
B.2. ZTP FSM . . . . . . . . . . . . . . . . . . . . . . . . . 101
B.3. Flooding Procedures . . . . . . . . . . . . . . . . . . . 105
B.3.1. FloodState Structure per Adjacency . . . . . . . . . 106
B.3.2. TIDEs . . . . . . . . . . . . . . . . . . . . . . . . 107
B.3.2.1. TIDE Generation . . . . . . . . . . . . . . . . . 107
B.3.2.2. TIDE Processing . . . . . . . . . . . . . . . . . 108
B.3.3. TIREs . . . . . . . . . . . . . . . . . . . . . . . . 109
B.3.3.1. TIRE Generation . . . . . . . . . . . . . . . . . 109
B.3.3.2. TIRE Processing . . . . . . . . . . . . . . . . . 110
B.3.4. TIEs Processing on Flood State Adjacency . . . . . . 110
B.3.5. TIEs Processing When LSDB Received Newer Version on
Other Adjacencies . . . . . . . . . . . . . . . . . . 111
Appendix C. Constants . . . . . . . . . . . . . . . . . . . . . 111
C.1. Configurable Protocol Constants . . . . . . . . . . . . . 111
Appendix D. TODO . . . . . . . . . . . . . . . . . . . . . . . . 113
Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 113
4.2.6. Attaching Prefixes . . . . . . . . . . . . . . . . . 32 1. Authors
4.2.7. Attaching Policy-Guided Prefixes . . . . . . . . . . 33
4.2.8. Automatic Disaggregation on Link & Node Failures . . 34
4.2.9. Optional Autoconfiguration . . . . . . . . . . . . . 37
4.2.9.1. Terminology . . . . . . . . . . . . . . . . . . . 38
4.2.9.2. Automatic SystemID Selection . . . . . . . . . . 39
4.2.9.3. Generic Fabric Example . . . . . . . . . . . . . 39
4.2.9.4. Level Determination Procedure . . . . . . . . . . 40
4.2.9.5. Resulting Topologies . . . . . . . . . . . . . . 41
4.2.10. Stability Considerations . . . . . . . . . . . . . . 43
4.3. Further Mechanisms . . . . . . . . . . . . . . . . . . . 44
4.3.1. Overload Bit . . . . . . . . . . . . . . . . . . . . 44
4.3.2. Optimized Route Computation on Leafs . . . . . . . . 44
4.3.3. Mobility . . . . . . . . . . . . . . . . . . . . . . 44
4.3.3.1. Clock Comparison . . . . . . . . . . . . . . . . 46
4.3.3.2. Interaction between Time Stamps and Sequence
Counters . . . . . . . . . . . . . . . . . . . . 46
4.3.3.3. Anycast vs. Unicast . . . . . . . . . . . . . . . 47
4.3.3.4. Overlays and Signaling . . . . . . . . . . . . . 47
4.3.4. Key/Value Store . . . . . . . . . . . . . . . . . . . 48
4.3.4.1. Southbound . . . . . . . . . . . . . . . . . . . 48
4.3.4.2. Northbound . . . . . . . . . . . . . . . . . . . 48
4.3.5. Interactions with BFD . . . . . . . . . . . . . . . . 48
4.3.6. Fabric Bandwidth Balancing . . . . . . . . . . . . . 49
4.3.6.1. Northbound Direction . . . . . . . . . . . . . . 49
4.3.6.2. Southbound Direction . . . . . . . . . . . . . . 51
4.3.7. Label Binding . . . . . . . . . . . . . . . . . . . . 52
4.3.8. Segment Routing Support with RIFT . . . . . . . . . . 52
4.3.8.1. Global Segment Identifiers Assignment . . . . . . 52
4.3.8.2. Distribution of Topology Information . . . . . . 52
4.3.9. Leaf to Leaf Procedures . . . . . . . . . . . . . . . 53
4.3.10. Other End-to-End Services . . . . . . . . . . . . . . 53
4.3.11. Address Family and Multi Topology Considerations . . 53
4.3.12. Reachability of Internal Nodes in the Fabric . . . . 54
4.3.13. One-Hop Healing of Levels with East-West Links . . . 54
5. Examples . . . . . . . . . . . . . . . . . . . . . . . . . . 54
5.1. Normal Operation . . . . . . . . . . . . . . . . . . . . 54
5.2. Leaf Link Failure . . . . . . . . . . . . . . . . . . . . 55
5.3. Partitioned Fabric . . . . . . . . . . . . . . . . . . . 56
5.4. Northbound Partitioned Router and Optional East-West
Links . . . . . . . . . . . . . . . . . . . . . . . . . . 58
6. Implementation and Operation: Further Details . . . . . . . . 59
6.1. Considerations for Leaf-Only Implementation . . . . . . . 60
6.2. Adaptations to Other Proposed Data Center Topologies . . 60
6.3. Originating Non-Default Route Southbound . . . . . . . . 61
7. Security Considerations . . . . . . . . . . . . . . . . . . . 61
8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 62
9. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 62
10. References . . . . . . . . . . . . . . . . . . . . . . . . . 62
10.1. Normative References . . . . . . . . . . . . . . . . . . 62
10.2. Informative References . . . . . . . . . . . . . . . . . 65
Appendix A. Information Elements Schema . . . . . . . . . . . . 66
A.1. common.thrift . . . . . . . . . . . . . . . . . . . . . . 67
A.2. encoding.thrift . . . . . . . . . . . . . . . . . . . . . 72
Appendix B. Finite State Machines . . . . . . . . . . . . . . . 77
B.1. LIE . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
B.2. ZTP . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
Appendix C. Constants . . . . . . . . . . . . . . . . . . . . . 87
C.1. Configurable Protocol Constants . . . . . . . . . . . . . 87
Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 87
1. Introduction This work is a product of a growing list of individuals.
ANISOTROPIC Tony Przygienda, Ed | Alankar Sharma | Pascal Thubert
Juniper Networks | Comcast | Cisco
Bruno Rijsman | Ilya Vershkov | Alia Atlas
Individual | Mellanox | Individual
Don Fedyk | John Drake |
HPE | Juniper |
Table 1: RIFT Authors
2. Introduction
Clos [CLOS] and Fat-Tree [FATTREE] have gained prominence in today's Clos [CLOS] and Fat-Tree [FATTREE] have gained prominence in today's
networking, primarily as result of the paradigm shift towards a networking, primarily as result of the paradigm shift towards a
centralized data-center based architecture that is poised to deliver centralized data-center based architecture that is poised to deliver
a majority of computation and storage services in the future. a majority of computation and storage services in the future.
Today's routing protocols were geared towards a network with an Today's current routing protocols were geared towards a network with
irregular topology and low degree of connectivity originally but an irregular topology and low degree of connectivity originally but
given they were the only available mechanisms, consequently several given they were the only available options, consequently several
attempts to apply those to Clos have been made. Most successfully attempts to apply those protocols to Clos have been made. Most
BGP [RFC4271] [RFC7938] has been extended to this purpose, not as successfully BGP [RFC4271] [RFC7938] has been extended to this
much due to its inherent suitability to solve the problem but rather purpose, not as much due to its inherent suitability but rather
because the perceived capability to modify it "quicker" and the because the perceived capability to easily modify BGP and the
immanent difficulties with link-state [DIJKSTRA] based protocols to immanent difficulties with link-state link-state [DIJKSTRA] based
perform in large scale densely meshed topologies. protocols to optimize topology exchange and converge quickly in large
scale densely meshed topologies. The incumbent protocols
precondition normally extensive configuration or provisioning during
bring up and re-dimensioning which is only viable for a set of
organizations with according networking operation skills and budgets.
For the majority of data center consumers a preferable protocol would
be one that auto-configures itself and deals with failures and
misconfigurations with a minimum of human intervention only. Such a
solution would allow local IP fabric bandwidth to be consumed in a
standardized component fashion, i.e. provision it much faster and
operate it at much lower costs, much like compute or storage is
consumed today.
In looking at the problem through the lens of its requirements an In looking at the problem through the lens of data center
optimal approach does not seem however to be a simple modification of requirements, an optimal approach does not seem however to be a
either a link-state (distributed computation) or distance-vector simple modification of either a link-state (distributed computation)
(diffused computation) approach but rather a mixture of both, or distance-vector (diffused computation) approach but rather a
colloquially best described as "link-state towards the spine" and mixture of both, colloquially best described as "link-state towards
"distance vector towards the leafs". In other words, "bottom" levels the spine" and "distance vector towards the leafs". In other words,
are flooding their link-state information in the "northern" direction "bottom" levels are flooding their link-state information in the
while each switch generates under normal conditions a default route "northern" direction while each node generates under normal
and floods it in the "southern" direction. Obviously, such conditions a default route and floods it in the "southern" direction.
aggregation can blackhole in cases of misconfiguration or failures This type of protocol allows highly desirable aggregation. Alas,
and this has to be addressed somehow. such aggregation could blackhole traffic in cases of misconfiguration
or while failures are being resolved or even cause partial network
partitioning and this has to be addressed. The approach RIFT takes
is described in Section 5.2.5 and is basically based on automatic,
sufficient disaggregation of prefixes to prevent any possible
problems.
For the visually oriented reader, Figure 1 presents a first For the visually oriented reader, Figure 1 presents a first level
simplified view of the resulting information and routes on a RIFT simplified view of the resulting information and routes on a RIFT
fabric. The top of the fabric is holding in its link-state database fabric. The top of the fabric, is holding, in its link-state
the nodes below it and routes to them. In the second row of the database the nodes below it and the routes to them. In the second
database we indicate that a partial information of other nodes in the row of the database we indicate that partial information of other
same level is available as well; the details of how this is achieved nodes in the same level is available as well. The details of how
should be postponed for the moment. Whereas when we look at the this is achieved will be postponed for the moment. When we look at
"bottom" of the fabric we see that the topology of the leafs is the "bottom" of the fabric, the leafs, we see that the topology is
basically empty and they only hold a load balanced default route to basically empty and they only hold a load balanced default route to
the next level. the next level.
The balance of this document details the resulting protocol and fills The balance of this document details the resulting protocol and fills
in the missing details. in the missing details.
. [A,B,C,D] . [A,B,C,D]
. [E] . [E]
. +-----+ +-----+ . +-----+ +-----+
. | E | | F | A/32 @ [C,D] . | E | | F | A/32 @ [C,D]
skipping to change at page 5, line 37 skipping to change at page 6, line 32
. A/32 @ A | | +-----+ | A/32 @ A . A/32 @ A | | +-----+ | A/32 @ A
. B/32 @ B | | | | B/32 @ B . B/32 @ B | | | | B/32 @ B
. | +------+ | . | +------+ |
. | | | | . | | | |
. +-+---+ | | +---+-+ . +-+---+ | | +---+-+
. | A +--+ +-+ B | . | A +--+ +-+ B |
. 0/0 @ [C,D] +-----+ +-----+ 0/0 @ [C,D] . 0/0 @ [C,D] +-----+ +-----+ 0/0 @ [C,D]
Figure 1: RIFT information distribution Figure 1: RIFT information distribution
1.1. Requirements Language 2.1. Requirements Language
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in RFC 2119 [RFC2119]. document are to be interpreted as described in RFC 2119 [RFC2119].
2. Reference Frame 3. Reference Frame
2.1. Terminology 3.1. Terminology
This section presents the terminology used in this document. It is This section presents the terminology used in this document. It is
assumed that the reader is thoroughly familiar with the terms and assumed that the reader is thoroughly familiar with the terms and
concepts used in OSPF [RFC2328] and IS-IS [ISO10589-Second-Edition], concepts used in OSPF [RFC2328] and IS-IS [ISO10589-Second-Edition],
[ISO10589] as well as the according graph theoretical concepts of [ISO10589] as well as the according graph theoretical concepts of
shortest path first (SPF) [DIJKSTRA] computation and directed acyclic shortest path first (SPF) [DIJKSTRA] computation and directed acyclic
graphs (DAG). graphs (DAG).
Level: Clos and Fat Tree networks are trees and 'level' denotes the Level: Clos and Fat Tree networks are topologically partially
set of nodes at the same height in such a network, where the ordered graphs and 'level' denotes the set of nodes at the same
bottom level is level 0. A node has links to nodes one level down height in such a network, where the bottom level (leaf) is the
level with lowest value. A node has links to nodes one level down
and/or one level up. Under some circumstances, a node may have and/or one level up. Under some circumstances, a node may have
links to nodes at the same level. As footnote: Clos terminology links to nodes at the same level. As footnote: Clos terminology
uses often the concept of "stage" but due to the folded nature of uses often the concept of "stage" but due to the folded nature of
the Fat Tree we do not use it to prevent misunderstandings. the Fat Tree we do not use it to prevent misunderstandings.
Superspine/Aggregation or Spine/Edge Levels: Traditional names for Superspine/Aggregation or Spine/Edge Levels: Traditional names in
Level 2, 1 and 0 respectively. Level 0 is often called leaf as 5-stages folded Clos for Level 2, 1 and 0 respectively. Level 0
well. is often called leaf as well. We normalize this language to talk
about leafs, spines and top-of-fabric (ToF).
Point of Delivery (PoD): A self-contained vertical slice of a Clos Point of Delivery (PoD): A self-contained vertical slice or subset
or Fat Tree network containing normally only level 0 and level 1 of a Clos or Fat Tree network containing normally only level 0 and
nodes. It communicates with nodes in other PoDs via the spine. level 1 nodes. A node in a PoD communicates with nodes in other
We number PoDs to distinguish them and use PoD #0 to denote PoDs via the Top-of-Fabric. We number PoDs to distinguish them
"undefined" PoD. and use PoD #0 to denote "undefined" PoD.
Superspine: The set of nodes that provide inter-PoD communication Top of PoD (ToP): The set of nodes that provide intra-PoD
and have no northbound adjacencies. Superspine nodes do not communication and have northbound adjacencies outside of the PoD,
belong to any PoD and are assigned "undefined" PoD value to i.e. are at the "top" of the PoD.
indicate the equivalent of "any" PoD.
Top of Fabric (ToF): The set of nodes that provide inter-PoD
communication and have no northbound adjacencies, i.e. are at the
"very top" of the fabric. ToF nodes do not belong to any PoD and
are assigned "undefined" PoD value to indicate the equivalent of
"any" PoD.
Spine: Any nodes north of leafs and south of top-of-fabric nodes.
Multiple layers of spines in a PoD are possible.
Leaf: A node without southbound adjacencies. Its level is 0 (except Leaf: A node without southbound adjacencies. Its level is 0 (except
cases where it is deriving its level via ZTP and is running cases where it is deriving its level via ZTP and is running
without LEAF_ONLY which will be explained in Section 4.2.9). without LEAF_ONLY which will be explained in Section 5.2.7).
Connected Spine: In case a spine level represents a connected graph Top-of-fabric Plane or Partition: In large fabrics top-of-fabric
(discounting links terminating at different levels), we call it a switches may not have enough ports to aggregate all switches south
"connected spine", in case a spine level consists of multiple of them and with that, the ToF is 'split' into multiple
partitions, we call it a "disconnected" or "partitioned spine". independent planes. Introduction and Section 5.1.2 explains the
In other terms, a spine without East-West links is disconnected concept in more detail. A plane is subset of ToF nodes that see
and is the typical configuration forf Clos and Fat Tree networks. each other through south reflection or E-W links.
Radix: A radix of a switch is basically number of switching ports it
provides. It's sometimes called fanout as well.
North Radix: Ports cabled northbound to higher level nodes.
South Radix: Ports cabled southbound to lower level nodes.
South/Southbound and North/Northbound (Direction): When describing South/Southbound and North/Northbound (Direction): When describing
protocol elements and procedures, we will be using in different protocol elements and procedures, we will be using in different
situations the directionality of the compass. I.e., 'south' or situations the directionality of the compass. I.e., 'south' or
'southbound' mean moving towards the bottom of the Clos or Fat 'southbound' mean moving towards the bottom of the Clos or Fat
Tree network and 'north' and 'northbound' mean moving towards the Tree network and 'north' and 'northbound' mean moving towards the
top of the Clos or Fat Tree network. top of the Clos or Fat Tree network.
Northbound Link: A link to a node one level up or in other words, Northbound Link: A link to a node one level up or in other words,
one level further north. one level further north.
skipping to change at page 7, line 17 skipping to change at page 8, line 27
East-West Link: A link between two nodes at the same level. East- East-West Link: A link between two nodes at the same level. East-
West links are normally not part of Clos or "fat-tree" topologies. West links are normally not part of Clos or "fat-tree" topologies.
Leaf shortcuts (L2L): East-West links at leaf level will need to be Leaf shortcuts (L2L): East-West links at leaf level will need to be
differentiated from East-West links at other levels. differentiated from East-West links at other levels.
Southbound representation: Information sent towards a lower level Southbound representation: Information sent towards a lower level
representing only limited amount of information. representing only limited amount of information.
South Reflection: Often abbreviated just as "reflection" it defines
a mechanism where South Node TIEs are "reflected" back up north to
allow nodes in same level without E-W links to "see" each other.
TIE: This is an acronym for a "Topology Information Element". TIEs TIE: This is an acronym for a "Topology Information Element". TIEs
are exchanged between RIFT nodes to describe parts of a network are exchanged between RIFT nodes to describe parts of a network
such as links and address prefixes. It can be thought of as such as links and address prefixes. A TIE can be thought of as
largely equivalent to ISIS LSPs or OSPF LSA. We will talk about largely equivalent to ISIS LSPs or OSPF LSA. We will talk about
N-TIEs when talking about TIEs in the northbound representation N-TIEs when talking about TIEs in the northbound representation
and S-TIEs for the southbound equivalent. and S-TIEs for the southbound equivalent.
Node TIE: This is an acronym for a "Node Topology Information Node TIE: This is an acronym for a "Node Topology Information
Element", largely equivalent to OSPF Node LSA, i.e. it contains Element", largely equivalent to OSPF Router LSA, i.e. it contains
all neighbors the node discovered and information about node all adjacencies the node discovered and information about node
itself. itself.
Prefix TIE: This is an acronym for a "Prefix Topology Information Prefix TIE: This is an acronym for a "Prefix Topology Information
Element" and it contains all prefixes directly attached to this Element" and it contains all prefixes directly attached to this
node in case of a N-TIE and in case of S-TIE the necessary default node in case of a N-TIE and in case of S-TIE the necessary default
and de-aggregated prefixes the node passes southbound. the node passes southbound.
Policy-Guided Information: Information that is passed in either
southbound direction or north-bound direction by the means of
diffusion and can be filtered via policies. Policy-Guided
Prefixes and KV Ties are examples of Policy-Guided Information.
Key Value TIE: A S-TIE that is carrying a set of key value pairs Key Value TIE: A S-TIE that is carrying a set of key value pairs
[DYNAMO]. It can be used to distribute information in the [DYNAMO]. It can be used to distribute information in the
southbound direction within the protocol. southbound direction within the protocol.
TIDE: Topology Information Description Element, equivalent to CSNP TIDE: Topology Information Description Element, equivalent to CSNP
in ISIS. in ISIS.
TIRE: Topology Information Request Element, equivalent to PSNP in TIRE: Topology Information Request Element, equivalent to PSNP in
ISIS. It can both confirm received and request missing TIEs. ISIS. It can both confirm received and request missing TIEs.
PGP: Policy-Guided Prefixes allow to support traffic engineering
that cannot be achieved by the means of SPF computation or normal
node and prefix S-TIE origination. S-PGPs are propagated in south
direction only and N-PGPs follow northern direction strictly.
De-aggregation/Disaggregation: Process in which a node decides to De-aggregation/Disaggregation: Process in which a node decides to
advertise certain prefixes it received in N-TIEs to prevent black- advertise certain prefixes it received in N-TIEs to prevent black-
holing and suboptimal routing upon link failures. holing and suboptimal routing upon link failures.
LIE: This is an acronym for a "Link Information Element", largely LIE: This is an acronym for a "Link Information Element", largely
equivalent to HELLOs in IGPs and exchanged over all the links equivalent to HELLOs in IGPs and exchanged over all the links
between systems running RIFT to form adjacencies. all the links
between systems running RIFT to form adjacencies. between systems running RIFT to form adjacencies.
FL: Flooding Leader for a specific system has a dedicated role to Flooding Leader (FL): Flooding Leader for a specific system has a
flood TIEs of that system. dedicated role to flood on northbound TIEs sent by this system.
Similar to MPR in OSLR.
FR: Flooding Repeater for a specific system has a dedicated role to
flood TIEs of that system northbound. Similar to MPR in OSLR.
BAD: This is an acronym for Bandwidth Adjusted Distance. RIFT Bandwidth Adjusted Distance (BAD): This is an acronym for Bandwidth
calculates the amount of northbound bandwidth available towards a Adjusted Distance. Each RIFT node calculates the amount of
node compared to other nodes at the same level and adjusts the northbound bandwidth available towards a node compared to other
default route distance accordingly to allow for the lower level to nodes at the same level and modifies the default route distance
adjust their load balancing. accordingly to allow for the lower level to adjust their load
balancing towards spines.
Overloaded: Applies to a node advertising `overload` attribute as Overloaded: Applies to a node advertising `overload` attribute as
set. The semantics closely follow the meaning of the same set. The semantics closely follow the meaning of the same
attribute in [ISO10589-Second-Edition]. attribute in [ISO10589-Second-Edition].
2.2. Topology Interface: A layer 3 entity over which RIFT control packets are
. +--------+ +--------+ exchanged.
. | | | | ^ N
. |Spine 21| |Spine 22| | Adjacency: RIFT tries to form a unique adjacency over an interface
and exchange local configuration and necessary ZTP information.
Neighbor: Once a three way adjacency has been formed a neighborship
relationship contains the neighbor's properties. Multiple
adjacencies can be formed to a neighbor via parallel interfaces
but such adjacencies are NOT sharing a neighbor structure. Saying
"neighbor" is thus equivalent to saying "a three way adjacency".
Cost: The term signifies the weighted distance between two
neighbors.
Distance: Sum of costs (bound by infinite distance) between two
nodes.
Metric: Without going deeper into the mathematic differentiation, a
metric is equivalent to distance.
3.2. Topology
. +--------+ +--------+ ^ N
. |ToF 21| |ToF 22| |
.Level 2 ++-+--+-++ ++-+--+-++ <-*-> E/W .Level 2 ++-+--+-++ ++-+--+-++ <-*-> E/W
. | | | | | | | | | . | | | | | | | | |
. P111/2| |P121 | | | | S v . P111/2| |P121 | | | | S v
. ^ ^ ^ ^ | | | | . ^ ^ ^ ^ | | | |
. | | | | | | | | . | | | | | | | |
. +--------------+ | +-----------+ | | | +---------------+ . +--------------+ | +-----------+ | | | +---------------+
. | | | | | | | | . | | | | | | | |
. South +-----------------------------+ | | ^ . South +-----------------------------+ | | ^
. | | | | | | | All TIEs . | | | | | | | All TIEs
. 0/0 0/0 0/0 +-----------------------------+ | . 0/0 0/0 0/0 +-----------------------------+ |
. v v v | | | | | . v v v | | | | |
. | | +-+ +<-0/0----------+ | | . | | +-+ +<-0/0----------+ | |
. | | | | | | | | . | | | | | | | |
.+-+----++ optional +-+----++ ++----+-+ ++-----++ .+-+----++ optional +-+----++ ++----+-+ ++-----++
.| | E/W link | | | | | | .| | E/W link | | | | | |
.|Node111+----------+Node112| |Node121| |Node122| .|Spin111+----------+Spin112| |Spin121| |Spin122|
.+-+---+-+ ++----+-+ +-+---+-+ ++---+--+ .+-+---+-+ ++----+-+ +-+---+-+ ++---+--+
. | | | South | | | | . | | | South | | | |
. | +---0/0--->-----+ 0/0 | +----------------+ | . | +---0/0--->-----+ 0/0 | +----------------+ |
. 0/0 | | | | | | | . 0/0 | | | | | | |
. | +---<-0/0-----+ | v | +--------------+ | | . | +---<-0/0-----+ | v | +--------------+ | |
. v | | | | | | | . v | | | | | | |
.+-+---+-+ +--+--+-+ +-+---+-+ +---+-+-+ .+-+---+-+ +--+--+-+ +-+---+-+ +---+-+-+
.| | (L2L) | | | | Level 0 | | .| | (L2L) | | | | Level 0 | |
.|Leaf111~~~~~~~~~~~~Leaf112| |Leaf121| |Leaf122| .|Leaf111~~~~~~~~~~~~Leaf112| |Leaf121| |Leaf122|
.+-+-----+ +-+---+-+ +--+--+-+ +-+-----+ .+-+-----+ +-+---+-+ +--+--+-+ +-+-----+
. + + \ / + + . + + \ / + +
. Prefix111 Prefix112 \ / Prefix121 Prefix122 . Prefix111 Prefix112 \ / Prefix121 Prefix122
. multi-homed . multi-homed
. Prefix . Prefix
.+---------- Pod 1 ---------+ +---------- Pod 2 ---------+ .+---------- Pod 1 ---------+ +---------- Pod 2 ---------+
Figure 2: A two level spine-and-leaf topology Figure 2: A three level spine-and-leaf topology
.+--------+ +--------+ +--------+ +--------+
.|ToF A1| |ToF B1| |ToF B2| |ToF A2|
.++-+-----+ ++-+-----+ ++-+-----+ ++-+-----+
. | | | | | | | |
. | | | | | +---------------+
. | | | | | | | |
. | | | +-------------------------+ |
. | | | | | | | |
. | +-----------------------+ | | | |
. | | | | | | | |
. | | +---------+ | +---------+ | |
. | | | | | | | |
. | +---------------------------------+ | |
. | | | | | | | |
.++-+-----+ ++-+-----+ +--+-+---+ +----+-+-+
.|Spine111| |Spine112| |Spine121| |Spine122|
.+-+---+--+ ++----+--+ +-+---+--+ ++---+---+
. | | | | | | | |
. | +--------+ | | +--------+ |
. | | | | | | | |
. | -------+ | | | +------+ | |
. | | | | | | | |
.+-+---+-+ +--+--+-+ +-+---+-+ +---+-+-+
.|Leaf111| |Leaf112| |Leaf121| |Leaf122|
.+-------+ +-------+ +-------+ +-------+
We will use this topology (called commonly a fat tree/network in Figure 3: Topology with multiple planes
modern DC considerations [VAHDAT08] as homonym to the original
definition of the term [FATTREE]) in all further considerations. It
depicts a generic "fat-tree" and the concepts explained in three
levels here carry by induction for further levels and higher degrees
of connectivity. However, this document will deal with designs that
provide only sparser connectivity as well.
3. Requirement Considerations We will use topology in Figure 2 (called commonly a fat tree/network
in modern IP fabric considerations [VAHDAT08] as homonym to the
original definition of the term [FATTREE]) in all further
considerations. This figure depicts a generic "single plane fat-
tree" and the concepts explained using three levels apply by
induction to further levels and higher degrees of connectivity.
Further, this document will deal also with designs that provide only
sparser connectivity and "partitioned spines" as shown in Figure 3
and explained further in Section 5.1.2.
4. Requirement Considerations
[RFC7938] gives the original set of requirements augmented here based [RFC7938] gives the original set of requirements augmented here based
upon recent experience in the operation of fat-tree networks. upon recent experience in the operation of fat-tree networks.
REQ1: The control protocol should discover the physical links REQ1: The control protocol should discover the physical links
automatically and be able to detect cabling that violates automatically and be able to detect cabling that violates
fat-tree topology constraints. It must react accordingly to fat-tree topology constraints. It must react accordingly to
such mis-cabling attempts, at a minimum preventing such mis-cabling attempts, at a minimum preventing
adjacencies between nodes from being formed and traffic from adjacencies between nodes from being formed and traffic from
being forwarded on those mis-cabled links. E.g. connecting being forwarded on those mis-cabled links. E.g. connecting
a leaf to a spine at level 2 should be detected and ideally a leaf to a spine at level 2 should be detected and ideally
prevented. prevented.
REQ2: A node without any configuration beside default values REQ2: A node without any configuration beside default values
should come up at the correct level in any PoD it is should come up at the correct level in any PoD it is
introduced into. Optionally, it must be possible to introduced into. Optionally, it must be possible to
configure nodes to restrict their participation to the configure nodes to restrict their participation to the
PoD(s) targeted at any level. PoD(s) targeted at any level.
REQ3: Optionally, the protocol should allow to provision data REQ3: Optionally, the protocol should allow to provision IP
centers where the individual switches carry no configuration fabrics where the individual switches carry no configuration
information and are all deriving their level from a "seed". information and are all deriving their level from a "seed".
Observe that this requirement may collide with the desire to Observe that this requirement may collide with the desire to
detect cabling misconfiguration and with that only one of detect cabling misconfiguration and with that only one of
the requirements can be fully met in a chosen configuration the requirements can be fully met in a chosen configuration
mode. mode.
REQ4: The solution should allow for minimum size routing REQ4: The solution should allow for minimum size routing
information base and forwarding tables at leaf level for information base and forwarding tables at leaf level for
speed, cost and simplicity reasons. Holding excessive speed, cost and simplicity reasons. Holding excessive
amount of information away from leaf nodes simplifies amount of information away from leaf nodes simplifies
operation and lowers cost of the underlay. operation and lowers cost of the underlay and allows to
scale and introduce proper multi-homing down to the server
level. The routing solution should allow for easy
instantiation of multiple routing planes. Coupled with
mobility defined in Paragraph 17 this should allow for
"light-weight" overlays on an IP fabric with e.g. native
IPv6 mobility support.
REQ5: Very high degree of ECMP must be supported. Maximum ECMP is REQ5: Very high degree of ECMP must be supported. Maximum ECMP is
currently understood as the most efficient routing approach currently understood as the most efficient routing approach
to maximize the throughput of switching fabrics to maximize the throughput of switching fabrics
[MAKSIC2013]. [MAKSIC2013].
REQ6: Non equal cost anycast must be supported to allow for easy REQ6: Non equal cost anycast must be supported to allow for easy
and robust multi-homing of services without regressing to and robust multi-homing of services without regressing to
careful balancing of link costs. careful balancing of link costs.
skipping to change at page 11, line 11 skipping to change at page 13, line 9
REQ8: The solution should allow for access to link states of the REQ8: The solution should allow for access to link states of the
whole topology to enable efficient support for modern whole topology to enable efficient support for modern
control architectures like SPRING [RFC7855] or PCE control architectures like SPRING [RFC7855] or PCE
[RFC4655]. [RFC4655].
REQ9: The solution should easily accommodate opaque data to be REQ9: The solution should easily accommodate opaque data to be
carried throughout the topology to subsets of nodes. This carried throughout the topology to subsets of nodes. This
can be used for many purposes, one of them being a key-value can be used for many purposes, one of them being a key-value
store that allows bootstrapping of nodes based right at the store that allows bootstrapping of nodes based right at the
time of topology discovery. time of topology discovery. Another use is distributing MAC
to L3 address binding from the leafs up north in case of
e.g. DHCP.
REQ10: Nodes should be taken out and introduced into production REQ10: Nodes should be taken out and introduced into production
with minimum wait-times and minimum of "shaking" of the with minimum wait-times and minimum of "shaking" of the
network, i.e. radius of propagation (often called "blast network, i.e. radius of propagation (often called "blast
radius") of changed information should be as small as radius") of changed information should be as small as
feasible. feasible.
REQ11: The protocol should allow for maximum aggregation of carried REQ11: The protocol should allow for maximum aggregation of carried
routing information while at the same time automatically de- routing information while at the same time automatically de-
aggregating the prefixes to prevent black-holing in case of aggregating the prefixes to prevent black-holing in case of
failures. The de-aggregation should support maximum failures. The de-aggregation should support maximum
possible ECMP/N-ECMP remaining after failure. possible ECMP/N-ECMP remaining after failure.
REQ12: Reducing the scope of communication needed throughout the REQ12: Reducing the scope of communication needed throughout the
network on link and state failure, as well as reducing network on link and state failure, as well as reducing
advertisements of repeating, idiomatic or policy-guided advertisements of repeating or idiomatic information in
information in stable state is highly desirable since it stable state is highly desirable since it leads to better
leads to better stability and faster convergence behavior. stability and faster convergence behavior.
REQ13: Once a packet traverses a link in a "southbound" direction, REQ13: Once a packet traverses a link in a "southbound" direction,
it must not take any further "northbound" steps along its it must not take any further "northbound" steps along its
path to delivery to its destination under normal conditions. path to delivery to its destination under normal, i.e.
Taking a path through the spine in cases where a shorter fully converged, conditions. Taking a path through the
path is available is highly undesirable. spine in cases where a shorter path is available is highly
undesirable.
REQ14: Parallel links between same set of nodes must be REQ14: Parallel links between same set of nodes must be
distinguishable for SPF, failure and traffic engineering distinguishable for SPF, failure and traffic engineering
purposes. purposes.
REQ15: The protocol must not rely on interfaces having discernible REQ15: The protocol must not rely on interfaces having discernible
unique addresses, i.e. it must operate in presence of unique addresses, i.e. it must operate in presence of
unnumbered links (even parallel ones) or links of a single unnumbered links (even parallel ones) or links of a single
node having same addresses. node having same addresses.
skipping to change at page 12, line 14 skipping to change at page 14, line 14
REQ17: The control plane should be able to unambiguously determine REQ17: The control plane should be able to unambiguously determine
the current point of attachment (which port on which leaf the current point of attachment (which port on which leaf
node) of a prefix, even in a context of fast mobility, e.g., node) of a prefix, even in a context of fast mobility, e.g.,
when the prefix is a host address on a wireless node that 1) when the prefix is a host address on a wireless node that 1)
may associate to any of multiple access points (APs) that may associate to any of multiple access points (APs) that
are attached to different ports on a same leaf node or to are attached to different ports on a same leaf node or to
different leaf nodes, and 2) may move and reassociate different leaf nodes, and 2) may move and reassociate
several times to a different AP within a sub-second period. several times to a different AP within a sub-second period.
REQ18: The protocol should provide security mechanisms that allow
to restrict nodes, especially leafs without proper
credentials from forming three-way adjacencies.
Following list represents possible requirements and requirements Following list represents possible requirements and requirements
under discussion: under discussion:
PEND1: Supporting anything but point-to-point links is a non- PEND1: Supporting anything but point-to-point links is a non-
requirement. Questions remain: for connecting to the requirement. Questions remain: for connecting to the
leaves, is there a case where multipoint is desirable? One leaves, is there a case where multipoint is desirable? One
could still model it as point-to-point links; it seems there could still model it as point-to-point links; it seems there
is no need for anything more than a NBMA-type construct. is no need for anything more than a NBMA-type construct.
PEND2: What is the maximum scale of number leaf prefixes we need to PEND2: What is the maximum scale of number leaf prefixes we need to
carry. Is 500'000 enough ? carry. 500'000 seems plenty even if we deploy RIFT down to
servers as leafs.
Finally, following are the non-requirements: Finally, following are the non-requirements:
NONREQ1: Broadcast media support is unnecessary. NONREQ1: Broadcast media support is unnecessary. However,
miscabling leading to multiple nodes on a broadcast
segment must be operationally easily recognizable and
operationally easily detectable while not taxing the
protocol excessively.
NONREQ2: Purging is unnecessary given its fragility and complexity NONREQ2: Purging link state elements is unnecessary given its
and today's large memory size on even modest switches and fragility and complexity and today's large memory size on
routers. even modest switches and routers.
NONREQ3: Special support for layer 3 multi-hop adjacencies is not NONREQ3: Special support for layer 3 multi-hop adjacencies is not
part of the protocol specification. Such support can be part of the protocol specification. Such support can be
easily provided by using tunneling technologies the same easily provided by using tunneling technologies the same
way IGPs today are solving the problem. way IGPs today are solving the problem.
4. RIFT: Routing in Fat Trees 5. RIFT: Routing in Fat Trees
Derived from the above requirements we present a detailed outline of Derived from the above requirements we present a detailed outline of
a protocol optimized for Routing in Fat Trees (RIFT) that in most a protocol optimized for Routing in Fat Trees (RIFT) that in most
abstract terms has many properties of a modified link-state protocol abstract terms has many properties of a modified link-state protocol
[RFC2328][ISO10589-Second-Edition] when "pointing north" and path- [RFC2328][ISO10589-Second-Edition] when "pointing north" and path-
vector [RFC4271] protocol when "pointing south". Albeit an unusual vector [RFC4271] protocol when "pointing south". While this is an
combination, it does quite naturally exhibit the desirable properties unusual combination, it does quite naturally exhibit the desirable
we seek. properties we seek.
4.1. Overview 5.1. Overview
The singular property of RIFT is that it floods only northbound 5.1.1. Properties
"flat" link-state information so that each level understands the full
The most singular property of RIFT is that it floods flat link-state
information northbound only so that each level obtains the full
topology of levels south of it. That information is never flooded topology of levels south of it. That information is never flooded
East-West or back South again. In the southbound direction the East-West (we'll talk about exceptions later) or back South again.
protocol operates like a "unidirectional" path vector protocol or In the southbound direction the protocol operates like a "fully
rather a distance vector with implicit split horizon whereas the summarizing, unidirectional" path vector protocol or rather a
information only propagates one hop south and is 're-advertised' by distance vector with implicit split horizon whereas the information
nodes at next lower level. However, we use flooding in the southern propagates one hop south and is 're-advertised' by nodes at next
direction as well to avoid the necessity to build an update per lower level, normally just the default route. However, RIFT uses
neighbor. We leave the East-West direction out for the moment. flooding in the southern direction as well to avoid the necessity to
build an update per adjacency. We omit describing the East-West
direction out for the moment.
Those information flow constraints create a "smooth" information Those information flow constraints create not only an anisotropic
propagation where nodes do not receive the same information from protocol (i.e. the information is not distributed "evenly" or
multiple fronts which would force them to perform a diffused "clumped" but summarized along the N-S gradient) but also a "smooth"
computation to tie-break the same reachability information arriving information propagation where nodes do not receive the same
on arbitrary links and ultimately force hop-by-hop forwarding on information from multiple fronts which would force them to perform a
shortest-paths only. diffused computation to tie-break the same reachability information
arriving on arbitrary links and ultimately force hop-by-hop
forwarding on shortest-paths only.
To account for the "northern" and the "southern" information split To account for the "northern" and the "southern" information split
the link state database is partitioned into "north representation" the link state database is partitioned into "north representation"
and "south representation" TIEs, whereas in simplest terms the N-TIEs and "south representation" TIEs, whereas in simplest terms the N-TIEs
contain a link state topology description of lower levels and and contain a link state topology description of lower levels and and
S-TIEs carry simply default routes. This oversimplified view will be S-TIEs carry simply default routes. This oversimplified view will be
refined gradually in following sections while introducing protocol refined gradually in following sections while introducing protocol
procedures aimed to fulfill the described requirements. procedures aimed to fulfill the described requirements.
4.2. Specification 5.1.2. Generalized Topology View
4.2.1. Transport This section will dwell on the topologies addresses by RIFT including
multi plane fabrics and their related implications. Readers that are
only interested in single plane designs, i.e. all top-of-fabric nodes
being topologically equal and initially connected to all the switches
at the level below them can skip this section and resulting
Section 5.2.5.2 as well.
All protocol elements are carried over UDP. Once QUIC [QUIC] Given the difficulty of visualizing multi plane design which are
achieves the desired stability in deployments it may prove a valuable effectively multi-dimensional switching matrices we will introduce a
candidate for TIE transport. methodology allowing us to visualize the connectivity in a two-
dimensional document and leverage the fact that we are dealing
basically with crossbar fabrics stacked on top of each other where
ports also align "on top of each other" in a regular fashion.
The typical topology for which RIFT is defined is built of a number P
of PoDs, connected together by a number S of spine nodes. A PoD node
has a number of ports called Radix, with half of them (K=Radix/2)
used to connect host devices from the south, and half to connect to
interleaved PoD Top-Level switches to the north. Ratio K can be
chosen differently without loss of generality when port speeds differ
or fabric is oversubscribed but K=R/2 allows for more readable
representation whereby there are as many ports facing north as south
on any intermediate node. We represent a node hence in a schematic
fashion with ports "sticking out" to its north and south rather than
by the usual real-world front faceplate designs of the day.
Figure 4 provides a view of a leaf node as seen from the north, i.e.
showing ports that connect northbound and for lack of a better
symbol, we have chosen to use the "HH" symbol as ASCII visualisation
of a RJ45 jack. Observe that the number of PoDs is not related to
Radix unless the Spine Nodes are constrained to be the same as the
PoD nodes in a particular deployment. We set the radix of the leaf
to K_LEAF, in this example 6 ports.
Top view
+----+
| |
| HH | e.g., Radix = 12, K_LEAF = 6
| |
| HH |
| | -------------------------
| HH ------- Physical Port (Ethernet) ----+
| | ------------------------- |
| HH | |
| | |
| HH | |
| | |
| HH | |
| | |
+----+ |
|| || || || || || ||
+----+ +------------------------------------------------+
| | | |
+----+ +------------------------------------------------+
|| || || || || || ||
Side views
Figure 4: A Leaf Node, K_LEAF=6
The Radix of a node on top of a PoD may be different than that of the
leaf node, though more often than not a same type of node is used for
both, effectively forming a square (K*K). In the general case, we
could have switches with K_TOP southern ports on nodes at the top of
the PoD that is not necessarily the same as K_LEAF; for instance, in
the representations below, we pick a K_LEAF of 6 and a K_TOP of 8.
In order to form a crossbar, we need K_TOP Leaf Nodes as illustrated
in Figure 5.
+----+ +----+ +----+ +----+ +----+ +----+ +----+ +----+
| | | | | | | | | | | | | | | |
| HH | | HH | | HH | | HH | | HH | | HH | | HH | | HH |
| | | | | | | | | | | | | | | |
| HH | | HH | | HH | | HH | | HH | | HH | | HH | | HH |
| | | | | | | | | | | | | | | |
| HH | | HH | | HH | | HH | | HH | | HH | | HH | | HH |
| | | | | | | | | | | | | | | |
| HH | | HH | | HH | | HH | | HH | | HH | | HH | | HH |
| | | | | | | | | | | | | | | |
| HH | | HH | | HH | | HH | | HH | | HH | | HH | | HH |
| | | | | | | | | | | | | | | |
| HH | | HH | | HH | | HH | | HH | | HH | | HH | | HH |
| | | | | | | | | | | | | | | |
+----+ +----+ +----+ +----+ +----+ +----+ +----+ +----+
Figure 5: Southern View of a PoD, K_TOP=8
The K_TOP Leaf Nodes are fully interconnected with the K_LEAF PoD-top
nodes, providing a connectivity that can be represented as a crossbar
as seen from the north and illustrated in Figure 6. The result is
that, in the absence of a breakage, a packet entering the PoD from
North on any port can be routed to any port on the south of the PoD
and vice versa.
E<-*->W
+----+ +----+ +----+ +----+ +----+ +----+ +----+ +----+
| | | | | | | | | | | | | | | |
+----------------------------------------------------------------+
| HH HH HH HH HH HH HH HH |
+----------------------------------------------------------------+
+----------------------------------------------------------------+
| HH HH HH HH HH HH HH HH |
+----------------------------------------------------------------+
+----------------------------------------------------------------+
| HH HH HH HH HH HH HH HH |
+----------------------------------------------------------------+
+----------------------------------------------------------------+
| HH HH HH HH HH HH HH HH |
+----------------------------------------------------------------+
+----------------------------------------------------------------+
| HH HH HH HH HH HH HH HH |<-+
+----------------------------------------------------------------+ |
+----------------------------------------------------------------+ |
| HH HH HH HH HH HH HH HH | |
+----------------------------------------------------------------+ |
| | | | | | | | | | | | | | | | |
+----+ +----+ +----+ +----+ +----+ +----+ +----+ +----+ |
^ |
| |
| ---------- --------------------- |
+----- Leaf Node PoD top Node (Spine) --+
---------- ---------------------
Figure 6: Northern View of a PoD's Spines, K_TOP=8
Side views of this PoD is illustrated in Figure 7 and Figure 8.
Connecting to Spine
|| || || || || || || ||
+----------------------------------------------------------------+ N
| PoD top Node seen sideways | ^
+----------------------------------------------------------------+ |
|| || || || || || || || *
+----+ +----+ +----+ +----+ +----+ +----+ +----+ +----+ |
| | | | | | | | | | | | | | | | v
+----+ +----+ +----+ +----+ +----+ +----+ +----+ +----+ S
|| || || || || || || ||
Connecting to Client nodes
Figure 7: Side View of a PoD, K_TOP=8, K_LEAF=6
Connecting to Spine
|| || || || || ||
+----+ +----+ +----+ +----+ +----+ +----+ N
| | | | | | | | | | | PoD top Nodes ^
+----+ +----+ +----+ +----+ +----+ +----+ |
|| || || || || || *
+------------------------------------------------+ |
| Leaf seen sideways | v
+------------------------------------------------+ S
|| || || || || ||
Connecting to Client nodes
Figure 8: Other side View of a PoD, K_TOP=8, K_LEAF=6, 90o turn in
E-W Plane
Note that a resulting PoD can be abstracted as a bigger node with a
number K of K_POD= K_TOP * K_LEAF, and the design can recurse.
It is critical at this junction that the concept and the picture of
those "crossed crossbars" is clear before progressing further,
otherwise following considerations will be difficult to comprehend.
Further, the PoDs are interconnected with one another through a Top-
of-Fabric at the very top or the north edge of the fabric. The
resulting ToF is NOT partitioned if and only if (IIF) every PoD top
level node (spine) is connected to every ToF Node. This is also
referred to as a single plane configuration. In order to reach a
1::1 connectivity ratio between the ToF and the Leaves, it results
that there are K_TOP ToF nodes, because each port of a ToP node
connects to a different ToF node, and K_LEAF ToP nodes for the same
reason. Consequently, it takes (P * K_LEAF) ports on a ToF node to
connect to each of the K_LEAF ToP nodes of the P PoDs, as illustrated
in Figure 9.
[ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] <-----+
| | | | | | | | |
[=================================] | -----------
| | | | | | | | +----- Top-of-Fabric
[ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] +----- Node -------+
| ----------- |
| v
+-+ +-+ +-+ +-+ +-+ +-+ +-+ +-+ <-----+ +-+
| | | | | | | | | | | | | | | | | |
[ |H| |H| |H| |H| |H| |H| |H| |H| ] | |
[ |H| |H| |H| |H| |H| |H| |H| |H| ] ------------------------- | |
[ |H| |H| |H| |H| |H| |H| |H| |H<--- Physical Port (Ethernet) | |
[ |H| |H| |H| |H| |H| |H| |H| |H| ] ------------------------- | |
[ |H| |H| |H| |H| |H| |H| |H| |H| ] | |
[ |H| |H| |H| |H| |H| |H| |H| |H| ] | |
| | | | | | | | | | | | | | | | | |
[ |H| |H| |H| |H| |H| |H| |H| |H| ] | |
[ |H| |H| |H| |H| |H| |H| |H| |H| ] -------------- | |
[ |H| |H| |H| |H| |H| |H| |H| |H| ] <--- PoD top level | |
[ |H| |H| |H| |H| |H| |H| |H| |H| ] node (Spine) ---+ | |
[ |H| |H| |H| |H| |H| |H| |H| |H| ] -------------- | | |
[ |H| |H| |H| |H| |H| |H| |H| |H| ] | | |
| | | | | | | | | | | | | | | | -+ +- +-+ v | |
[ |H| |H| |H| |H| |H| |H| |H| |H| ] | | --| |--[ ]--| |
[ |H| |H| |H| |H| |H| |H| |H| |H| ] | ----- | --| |--[ ]--| |
[ |H| |H| |H| |H| |H| |H| |H| |H| ] +--- PoD ---+ --| |--[ ]--| |
[ |H| |H| |H| |H| |H| |H| |H| |H| ] | ----- | --| |--[ ]--| |
[ |H| |H| |H| |H| |H| |H| |H| |H| ] | | --| |--[ ]--| |
[ |H| |H| |H| |H| |H| |H| |H| |H| ] | | --| |--[ ]--| |
| | | | | | | | | | | | | | | | -+ +- +-+ | |
+-+ +-+ +-+ +-+ +-+ +-+ +-+ +-+ +-+
Figure 9: Fabric Spines and TOFs in Single Plane Design, 3 PoDs
The top view can be collapsed into a third dimension where the hidden
depth index is representing the PoD number. So we can show one PoD
as a class of PoDs and hence save one dimension in our
representation. The Spine Node expands in the depth and the vertical
dimensions whereas the PoD top level Nodes are constrained in
horizontal dimension. A port in the 2-D representation represents
effectively the class of all the ports at the same position in all
the PoDs that are projected in its position along the depth axis.
This is shown in Figure 10.
/ / / / / / / / / / / / / / / /
/ / / / / / / / / / / / / / / /
/ / / / / / / / / / / / / / / /
/ / / / / / / / / / / / / / / / ]
+-+ +-+ +-+ +-+ +-+ +-+ +-+ +-+ ]]
| | | | | | | | | | | | | | | | ] ---------------------------
[ |H| |H| |H| |H| |H| |H| |H| |H| ] <-- PoD top level node (Spine)
[ |H| |H| |H| |H| |H| |H| |H| |H| ] ---------------------------
[ |H| |H| |H| |H| |H| |H| |H| |H| ]]]]
[ |H| |H| |H| |H| |H| |H| |H| |H| ]]] ^^
[ |H| |H| |H| |H| |H| |H| |H| |H| ]] // PoDs
[ |H| |H| |H| |H| |H| |H| |H| |H| ] // (in depth)
| |/| |/| |/| |/| |/| |/| |/| |/ //
+-+ +-+ +-+/+-+/+-+ +-+ +-+ +-+ //
^
| ----------------
+----- Top-of-Fabric Node
----------------
Figure 10: Collapsed Northern View of a Fabric for Any Number of PoDs
This type of deployment introduces a "single plane limit" where the
bound is the available radix of the ToF nodes, which limits (P *
K_LEAF). Nevertheless, a distinct advantage of a connected or
unpartitioned Top-of-Fabric is that all failures can be resolved by
simple, non-transitive, positive disaggregation described in
Section 5.2.5.1 that propagates only within one level of the fabric.
In other words unpartitoned ToF nodes can always reach nodes below or
withdraw the routes from PoDs they cannot reach unambiguously. To be
more precise, all failures which still allow all the ToF nodes to see
each other via south reflection as explained in Section 5.2.5.
In order to scale beyond the "single plane limit", the Top-of-Fabric
can be partitioned by a number N of identically wired planes, N being
an integer divider of K_LEAF. The 1::1 ratio and the desired
symmetry are still served, this time with (K_TOP * N) ToF nodes, each
of (P * K_LEAF / N) ports. N=1 represents a non-partitioned Spine
and N=K_LEAF is a maximally partitioned Spine. Further, if R is any
divisor of K_LEAF, then (N=K_LEAF/R) is a feasible number of planes
and R a redundancy factor. If proves convenient for deployments to
use a radix for the leaf nodes that is a power of 2 so they can pick
a number of planes that is a lower power of 2. The example in
Figure 11 splits the Spine in 2 planes with a redundancy factor R=3,
meaning that there are 3 non-intersecting paths between any leaf node
and any ToF node. A ToF node must have in this case at least 3*P
ports, and be directly connected to 3 of the 6 PoD-ToP nodes (spines)
in each PoD.
+----+ +----+ +----+ +----+ +----+ +----+ +----+ +----+
+-| |--| |--| |--| |--| |--| |--| |--| |-+
| | HH | | HH | | HH | | HH | | HH | | HH | | HH | | HH | |
+-| |--| |--| |--| |--| |--| |--| |--| |-+
+-| |--| |--| |--| |--| |--| |--| |--| |-+
| | HH | | HH | | HH | | HH | | HH | | HH | | HH | | HH | |
+-| |--| |--| |--| |--| |--| |--| |--| |-+
+-| |--| |--| |--| |--| |--| |--| |--| |-+
| | HH | | HH | | HH | | HH | | HH | | HH | | HH | | HH | |
+-| |--| |--| |--| |--| |--| |--| |--| |-+
+----+ +----+ +----+ +----+ +----+ +----+ +----+ +----+
Plane 1
----------- . ------------ . ------------ . ------------ . --------
Plane 2
+----+ +----+ +----+ +----+ +----+ +----+ +----+ +----+
+-| |--| |--| |--| |--| |--| |--| |--| |-+
| | HH | | HH | | HH | | HH | | HH | | HH | | HH | | HH | |
+-| |--| |--| |--| |--| |--| |--| |--| |-+
+-| |--| |--| |--| |--| |--| |--| |--| |-+
| | HH | | HH | | HH | | HH | | HH | | HH | | HH | | HH | |
+-| |--| |--| |--| |--| |--| |--| |--| |-+
+-| |--| |--| |--| |--| |--| |--| |--| |-+
| | HH | | HH | | HH | | HH | | HH | | HH | | HH | | HH | |
+-| |--| |--| |--| |--| |--| |--| |--| |-+
+----+ +----+ +----+ +----+ +----+ +----+ +----+ +----+
^
|
| ----------------
+----- Top-of-Fabric node
"across" depth
----------------
Figure 11: Northern View of a Multi-Plane ToF Level, K_LEAF=6, N=2
At the extreme end of the spectrum, it is even possible to fully
partition the spine with N = K_LEAF and R=1, while maintaining
connectivity between each leaf node and each Top-of-Fabric node. In
that case the ToF node connects to a single Port per PoD, so it
appears as a single port in the projected view represented in
Figure 12 and the number of ports required on the Spine Node is more
or equal to P, the number of PoDs.
Plane 1
+----+ +----+ +----+ +----+ +----+ +----+ +----+ +----+ -+
+-| |--| |--| |--| |--| |--| |--| |--| |-+ |
| | HH | | HH | | HH | | HH | | HH | | HH | | HH | | HH | | |
+-| |--| |--| |--| |--| |--| |--| |--| |-+ |
+----+ +----+ +----+ +----+ +----+ +----+ +----+ +----+ |
----------- . ------------ . ------------ . ------------ . -------- |
+----+ +----+ +----+ +----+ +----+ +----+ +----+ +----+ |
+-| |--| |--| |--| |--| |--| |--| |--| |-+ |
| | HH | | HH | | HH | | HH | | HH | | HH | | HH | | HH | | |
+-| |--| |--| |--| |--| |--| |--| |--| |-+ |
+----+ +----+ +----+ +----+ +----+ +----+ +----+ +----+ |
----------- . ------------ . ------------ . ------------ . -------- |
+----+ +----+ +----+ +----+ +----+ +----+ +----+ +----+ |
+-| |--| |--| |--| |--| |--| |--| |--| |-+ |
| | HH | | HH | | HH | | HH | | HH | | HH | | HH | | HH | | |
+-| |--| |--| |--| |--| |--| |--| |--| |-+ |
+----+ +----+ +----+ +----+ +----+ +----+ +----+ +----+ |
----------- . ------------ . ------------ . ------------ . -------- +<-+
+----+ +----+ +----+ +----+ +----+ +----+ +----+ +----+ | |
+-| |--| |--| |--| |--| |--| |--| |--| |-+ | |
| | HH | | HH | | HH | | HH | | HH | | HH | | HH | | HH | | | |
+-| |--| |--| |--| |--| |--| |--| |--| |-+ | |
+----+ +----+ +----+ +----+ +----+ +----+ +----+ +----+ | |
----------- . ------------ . ------------ . ------------ . -------- | |
+----+ +----+ +----+ +----+ +----+ +----+ +----+ +----+ | |
+-| |--| |--| |--| |--| |--| |--| |--| |-+ | |
| | HH | | HH | | HH | | HH | | HH | | HH | | HH | | HH | | | |
+-| |--| |--| |--| |--| |--| |--| |--| |-+ | |
+----+ +----+ +----+ +----+ +----+ +----+ +----+ +----+ | |
----------- . ------------ . ------------ . ------------ . -------- | |
+----+ +----+ +----+ +----+ +----+ +----+ +----+ +----+ | |
+-| |--| |--| |--| |--| |--| |--| |--| |-+ | |
| | HH | | HH | | HH | | HH | | HH | | HH | | HH | | HH | | | |
+-| |--| |--| |--| |--| |--| |--| |--| |-+ | |
+----+ +----+ +----+ +----+ +----+ +----+ +----+ +----+ -+ |
Plane 6 ^ |
| |
| ---------------- -------------- |
+----- ToF Node Class of PoDs ---+
---------------- -------------
Figure 12: Northern View of a Maximally Partitioned ToF Level, R=1
5.1.3. Fallen Leaf Problem
As mentioned earlier, RIFT exhibits an anisotropic behavior tailored
for fabrics with a North / South orientation and a high level of
interleaving paths. A non-partitioned fabric makes a total loss of
connectivity between a Top-of-Fabric node at the north and a leaf
node at the south a very rare but yet possible occasion that is fully
healed by positive disaggregation described in Section 5.2.5.1. In
large fabrics or fabrics built from switches with low radix, the ToF
ends often being partioned in planes which makes the occurrence of
having a given leaf being only reachable from a subset of the ToF
nodes more likely to happen. This makes some further considerations
necessary.
We define a "Fallen Leaf" as a leaf that can be reached by only a
subset of Top-of-Fabric nodes but cannot be reached by all due to
missing connectivity. If R is the redundancy factor, then it takes
at least R breakages to reach a "Fallen Leaf" situation.
In a maximally partitioned fabric, the redundancy factor R is 1, so
any breakage may cause one or more fallen leaves, but not all cases
require disaggregation. The following cases do not require
particular action:
If the breakage is a southern link from a leaf Node going down,
then connectivity to any node attached to the link is lost. There
is no need to disaggregate since the connectivity is lost for all
spine nodes in a same fashion.
If the breakage is a leaf Node going down, then connectivity
through that leaf is lost for all nodes. There is no need to
disaggregate since the connectivity is lost for all spine nodes in
a same fashion.
If the breakage is a ToF Node going down, then northern traffic is
routed via alternate ToF nodes in the same plane and there is no
need to disaggregate routes.
In a general manner, the mechanism of non-transitive positive
disaggregation is sufficient when the disaggregating ToF nodes
collectively connect to all the ToP nodes in the broken plane. This
happens in the following case:
If the breakage is the last northern link from a ToP node to a ToF
node going down, then the fallen leaf problem affects only The ToF
node, and the connectivity to all the nodes in the PoD is lost
from that ToF node. This can be observed by other ToF nodes
within the plane where the ToP node is located and positively
disaggregated within that plane.
On the other hand, there is a need to disaggregate the routes to
Fallen Leaves in a transitive fashion all the way to the other leaves
in the following cases:
If the breakage is the last northern link from a Leaf node within
a plane - there is only one such link in a maximally partitioned
fabric - that goes down, then connectivity to all unicast prefixes
attached to the Leaf node is lost within the plane where the link
is located. Southern Reflection by a Leaf Node - e.g., between
ToP nodes if the PoD has only 2 levels - happens in between
planes, allowing the ToP nodes to detect the problem within the
PoD where it occurs and positively disaggregate. The breakage can
be observed by the ToF nodes in the same plane through the
flooding of N-TIEs from the ToP nodes, but the ToF nodes need to
be aware of all the affected prefixes for the negative
disaggregation to be fully effective. The problem can also be
observed by the ToF nodes in the other planes through the flooding
of N-TIEs from the affected Leaf nodes, together with non-node
N-TIEs which indicate the affected prefixes. To be effective in
that case, the positive disaggregation must reach down to the
nodes that make the plane selection, which are typically the
ingress Leaf nodes, and the information is not useful for routing
in the intermediate levels.
If the breakage is a ToP node in a maximally partitioned fabric -
in which case it is the only ToP node serving that plane in that
PoD - that goes down, then the connectivity to all the nodes in
the PoD is lost within the plane where the ToP node is located -
all leaves fall. Since the Southern Reflection between the ToF
nodes happens only within a plane, ToF nodes in other planes
cannot discover the case of fallen leaves in a different plane,
and cannot determine beyond their local plane whether a Leaf node
that was initially reachable has become unreachable. As above,
the breakage can be observed by the ToF nodes in the plane where
the breakage happened, and then again, the ToF nodes in the plane
need to be aware of all the affected prefixes for the negative
disaggregation to be fully effective. The problem can also be
observed by the ToF nodes in the other planes through the flooding
of N-TIEs from the affected Leaf nodes, if there are only 3 levels
and the ToP nodes are directly connected to the Leaf nodes, and
then again it can only be effective it is propagated transitively
to the Leaf, and useless above that level.
For the sake of readability let us roll the abstractions back to a
simplest example and observe that in Figure 3 the loss of link Spine
122 to Leaf 122 will make Leaf 122 a fallen leaf for Top-of-Fabric
plane B. Worse, if the cabling was never present in first place,
plane B will not even be able to know that such a fallen leaf exists.
Hence partitioning without further treatment results in two grave
problems:
o Leaf111 trying to route to Leaf122 MUST choose Spine 111 in plane
A as its next hop since plane B will inevitably blackhole the
packet when forwarding using default routes or do excessive bow
tie'ing, i.e. this information must be in its routing table.
o any kind of "flooding" or distance vector trying to deal with the
problem by distributing host routes will be able to converge only
using paths through leafs, i.e. the flooding of information on
Leaf122 will go up to Top-of-Fabric A and then "loopback" over
other leafs to ToF B leading in extreme cases to traffic for
Leaf122 when presented to plane B taking an "inverted fabric" path
where leafs start to serve as TOFs.
5.1.4. Discovering Fallen Leaves
As we illustrate later and without further proof here, to deal with
fallen leafs in multi-plane designs RIFT requires all the ToF nodes
to share the same topology database. This happens naturally in
single plane design but needs additional considerations in multi-
plane fabrics. To satisfy this RIFT in multi-plane designs relies at
the ToF Level on ring interconnection of switches in multiple planes.
Other solutions are possible but they either need more cabling or end
up having much longer flooding path or single points of failure.
In more detail, by reserving two ports on each Top-of-Fabric node it
is possible to connect them together in an interplane bi-directional
ring as illustrated in Figure 13 (where we show a bi-directional ring
connecting switches across planes). The rings will exchange full
topology information between planes and with that allow consequently
by the means of transitive, negative disaggregation described in
Section 5.2.5.2 to efficiently fix any possible fallen leaf scenario.
Somewhat as a side-effect, the exchange of information fulfills the
requirement to present full view of the fabric topology at the Top-
of-Fabric level without the need to collate it from multiple points
by additional complexity of technologies like [RFC7752].
+----+ +----+ +----+ +----+ +----+ +----+ +--------+
| | | | | | | | | | | | | |
| | | | | | | |
+-o--+ +-o--+ +-o--+ +-o--+ +-o--+ +-o--+ +-o--+ |
+-| |--| |--| |--| |--| |--| |--| |-+ |
| | HH | | HH | | HH | | HH | | HH | | HH | | HH | | | Plane A
+-| |--| |--| |--| |--| |--| |--| |-+ |
+-o--+ +-o--+ +-o--+ +-o--+ +-o--+ +-o--+ +-o--+ |
| | | | | | | |
+-o--+ +-o--+ +-o--+ +-o--+ +-o--+ +-o--+ +-o--+ |
+-| |--| |--| |--| |--| |--| |--| |-+ |
| | HH | | HH | | HH | | HH | | HH | | HH | | HH | | | Plane B
+-| |--| |--| |--| |--| |--| |--| |-+ |
+-o--+ +-o--+ +-o--+ +-o--+ +-o--+ +-o--+ +-o--+ |
| | | | | | | |
... |
| | | | | | | |
+-o--+ +-o--+ +-o--+ +-o--+ +-o--+ +-o--+ +-o--+ |
+-| |--| |--| |--| |--| |--| |--| |-+ |
| | HH | | HH | | HH | | HH | | HH | | HH | | HH | | | Plane X
+-| |--| |--| |--| |--| |--| |--| |-+ |
+-o--+ +-o--+ +-o--+ +-o--+ +-o--+ +-o--+ +-o--+ |
| | | | | | | |
| | | | | | | | | | | | | |
+----+ +----+ +----+ +----+ +----+ +----+ +--------+
Figure 13: Connecting Top-of-Fabric Nodes Across Planes by Two Rings
5.1.5. Addressing the Fallen Leaves Problem
One consequence of the Fallen Leaf problem is that some prefixes
attached to the fallen leaf become unreachable from some of the ToF
nodes. RIFT proposes two methods to address this issue, the positive
and the negative disaggregation. Both methods flood S-TIEs to
advertise the impacted prefix(es).
When used for the operation of disaggregation, a positive S-TIE, as
usual, indicates reachability to a prefix of given length and all
addresses subsumed by it. In contrast, a negative route
advertisement indicates that the origin cannot route to the
advertised prefix.
The positive disaggregation is originated by a router that can still
reach the advertised prefix, and the operation is not transitive,
meaning that the receiver does not generate its own flooding south as
a consequence of receiving positive disaggregation advertisements
from an higher level node. The effect of a positive disaggregation
is that the traffic to the impacted prefix will follow the prefix
longest match and will be limited to the northbound routers that
advertised the more specific route.
In contrast, the negative disaggregation is transitive, and is
propagated south when all the possible routes northwards are barred.
A negative route advertisement is only actionable when the negative
prefix is aggregated by a positive route advertisement for a shorter
prefix. In that case, the negative advertisement carves an exception
to the positive route in the routing table (one could think of
"punching a hole"), making the positive prefix reachable through the
originator with the special consideration of the negative prefix
removing certain next hop neighbors.
When the ToF is not partitioned, the collective southern flooding of
the positive disaggregation by the ToF nodes that can still reach the
impacted prefix is in general enough to cover all the switches at the
next level south, typically the ToP nodes. If all those switches are
aware of the disaggregation, they collectively create a ceiling that
intercepts all the traffic north and forwards it to the ToF nodes
that advertised the more specific route. In that case, the positive
disaggregation alone is sufficient to solve the fallen leaf problem.
On the other hand, when the fabric is partitioned in planes, the
positive disaggregation from ToF nodes in different planes do not
reach the ToP switches in the affected plane and cannot solve the
fallen leaves problem. In other words, a breakage in a plane can
only be solved in that plane. Also, the selection of the plane for a
packet typically occurs at the leaf level and the disaggregation must
be transitive and reach all the leaves. In that case, the negative
disaggregation is necessary. The details on the RIFT approach to
deal with fallen leafs in an optimal way is specified in
Section 5.2.5.2.
5.2. Specification
5.2.1. Transport
All packet formats are defined in Thrift models in Appendix A. All packet formats are defined in Thrift models in Appendix A.
Future versions may include a [PROTOBUF] schema. The serialized model is carried in an envelope within a UDP frame
that provides security and allows validation/modification of several
important fields without de-serialization for performance and
security reasons.
4.2.2. Link (Neighbor) Discovery (LIE Exchange) The ultimate transport envelope, especially the placement of nonces
is under active discussion.
+--------+----------+-------------+----------+-------------+---------+----------------+
| UDP | TIE | Fingerprint | Key ID | Security | Model | Serialized |
| Header | Lifetime | Type | | Fingerprint | Version | RIFT Model ... |
| | | (e.g. SHA) | | | | Object |
+--------+----------+-------------+----------+-------------+---------+----------------+
Figure 14: Security Envelope
5.2.2. Link (Neighbor) Discovery (LIE Exchange)
LIE exchange happens over well-known administratively locally scoped LIE exchange happens over well-known administratively locally scoped
and configured or otherwise well-known IPv4 multicast address and configured or otherwise well-known IPv4 multicast address
[RFC2365] or link-local multicast scope [RFC4291] for IPv6 [RFC8200] [RFC2365] or link-local multicast scope [RFC4291] for IPv6 [RFC8200]
using a configured or otherwise a well-known destination UDP port using a configured or otherwise a well-known destination UDP port
defined in Appendix C.1. LIEs SHOULD be sent with a TTL of 1 to defined in Appendix C.1. LIEs SHOULD be sent with a TTL of 1 to
prevent RIFT information reaching beyond a single L3 next-hop in the prevent RIFT information reaching beyond a single L3 next-hop in the
topology. LIEs SHOULD be sent with network control precendence. topology. LIEs SHOULD be sent with network control precedence.
Originating port of the LIE has no further significance. LIEs are Originating port of the LIE has no further significance other than
exchanged over all links running RIFT. An implementation MAY listen identifying the origination point. LIEs are exchanged over all links
and send LIEs on IPv4 and/or IPv6 multicast addresses. LIEs on same running RIFT. An implementation MAY listen and send LIEs on IPv4
link are considered part of the same negotiation independent on the and/or IPv6 multicast addresses. LIEs on same link are considered
address family they arrive on. Observe further that the LIE source part of the same negotiation independent on the address family they
address may not identify the peer uniquely in unnumbered or link- arrive on. Observe further that the LIE source address may not
local address cases so the transmission should occur over the same identify the peer uniquely in unnumbered or link-local address cases
interface the LIEs have been received on. A node can use any of the so the response transmission MUST occur over the same interface the
neighbor's LIE source addresses to send TIEs. LIEs have been received on. A node CAN use any of the adjacency's
source addresses it saw in LIEs on the specific interface during
adjacency formation to send TIEs. That implies that an
implementation MUST be ready to accept TIEs on all addresses it used
as source of LIE frames.
Unless Section 4.2.9 is used, each node is provisioned with the level Observe further that the protocol does NOT support selective
disabling of address families or any local address changes in three
way state, i.e. if a link has entered three way IPv4 and/or IPv6 with
a neighbor on an adjacency and it wants to stop supporting one of the
families or change any of its local addresses, it has to tear down
and rebuild the adjacency. It also has to remove any information it
stored about adjacency's' LIE source addresses seen.
All RIFT routers MUST support IPv4 forwarding and MAY support IPv6
forwarding. A three way adjacency over IPv6 addresses implies
support for IPv4 forwarding.
Unless Section 5.2.7 is used, each node is provisioned with the level
at which it is operating and its PoD (or otherwise a default level at which it is operating and its PoD (or otherwise a default level
and "undefined" PoD are assumed; meaning that leafs do not need to be and "undefined" PoD are assumed; meaning that leafs do not need to be
configured at all if initial configuration values are all left at 0). configured at all if initial configuration values are all left at 0).
Nodes in the spine are configured with "any" PoD which has the same Nodes in the spine are configured with "any" PoD which has the same
value "undefined" PoD hence we will talk about "undefined/any" PoD. value "undefined" PoD hence we will talk about "undefined/any" PoD.
This information is propagated in the LIEs exchanged. This information is propagated in the LIEs exchanged.
Further definitions of leaf flags are found in Section 5.2.7 given
they have implications in terms of level and adjacency forming here.
A node tries to form a three way adjacency if and only if A node tries to form a three way adjacency if and only if
(definitions of LEAF_ONLY are found in Section 4.2.9)
1. the node is in the same PoD or either the node or the neighbor 1. the node is in the same PoD or either the node or the neighbor
advertises "undefined/any" PoD membership (PoD# = 0) AND advertises "undefined/any" PoD membership (PoD# = 0) AND
2. the neighboring node is running the same MAJOR schema version AND 2. the neighboring node is running the same MAJOR schema version AND
3. the neighbor is not member of some PoD while the node has a 3. the neighbor is not member of some PoD while the node has a
northbound adjacency already joining another PoD AND northbound adjacency already joining another PoD AND
4. the neighboring node uses a valid System ID AND 4. the neighboring node uses a valid System ID AND
skipping to change at page 14, line 46 skipping to change at page 31, line 38
5. the neighboring node uses a different System ID than the node 5. the neighboring node uses a different System ID than the node
itself itself
6. the advertised MTUs match on both sides AND 6. the advertised MTUs match on both sides AND
7. both nodes advertise defined level values AND 7. both nodes advertise defined level values AND
8. [ 8. [
i) the node is at level 0 and has no three way adjacencies i) the node is at level 0 and has no three way adjacencies
already to nodes with level higher than the neighboring node already to HAT nodes with level different than the adjacent
OR node OR
ii) the node is not at level 0 and the neighboring node is at ii) the node is not at level 0 and the neighboring node is at
level 0 OR level 0 OR
iii) both nodes are at level 0 AND both indicate support for iii) both nodes are at level 0 AND both indicate support for
Section 4.3.9 OR Section 5.3.9 OR
iii) neither node is at level 0 and the neighboring node is at iv) neither node is at level 0 and the neighboring node is at
most one level away most one level away
]. ].
Rule in Paragraph 3 MAY be optionally disregarded by a node if PoD The rule in Paragraph 3 MAY be optionally disregarded by a node if
detection is undesirable or has to be disregarded. PoD detection is undesirable or has to be disregarded.
A node configured with "undefined" PoD membership MUST, after A node configured with "undefined" PoD membership MUST, after
building first northbound three way adjacencies to a node being in a building first northbound three way adjacencies to a node being in a
defined PoD, advertise that PoD as part of its LIEs. In case that defined PoD, advertise that PoD as part of its LIEs. In case that
adjacency is lost, from all available northbound three way adjacency is lost, from all available northbound three way
adjacencies the node with the highest System ID and defined PoD is adjacencies the node with the highest System ID and defined PoD is
chosen. That way the northmost defined PoD value (normally the top chosen. That way the northmost defined PoD value (normally the top
spines in a PoD) can diffuse southbound towards the leafs "forcing" spines in a PoD) can diffuse southbound towards the leafs "forcing"
the PoD value on any node with "undefined" PoD. the PoD value on any node with "undefined" PoD.
LIEs arriving with a TTL larger than 1 MUST be ignored. LIEs arriving with a TTL larger than 1 MUST be ignored.
A node SHOULD NOT send out LIEs without defined level in the header A node SHOULD NOT send out LIEs without defined level in the header
but in certain scenarios it may be beneficial for trouble-shooting but in certain scenarios it may be beneficial for trouble-shooting
purposes. purposes.
LIE exchange uses three way handshake mechanism [RFC5303]. Precise LIE exchange uses three way handshake mechanism which is a cleaned up
finite state machines will be provided in later versions of this version of [RFC5303]. Observe that for easier comprehension the
specification. LIE packets contain nonces and may contain an SHA-1 terminology of one/two and three-way states does NOT align with OSPF
[RFC6234] over nonces and some of the LIE data which prevents or ISIS FSMs albeit they use roughly same mechanisms. LIE packets
corruption and replay attacks. TIE flooding reuses those nonces to reflect nonces and may contain an SHA-1 [RFC6234] over nonces and
prevent mismatches and can use those for security purposes in case it some of the LIE data which prevents corruption and replay attacks.
is using QUIC [QUIC]. Section 7 will address the precise security
mechanisms in the future.
4.2.3. Topology Exchange (TIE Exchange) 5.2.3. Topology Exchange (TIE Exchange)
4.2.3.1. Topology Information Elements 5.2.3.1. Topology Information Elements
Topology and reachability information in RIFT is conveyed by the Topology and reachability information in RIFT is conveyed by the
means of TIEs which have good amount of commonalities with LSAs in means of TIEs which have good amount of commonalities with LSAs in
OSPF. OSPF.
TIE exchange mechanism uses port indicated by each node in the LIE The TIE exchange mechanism uses the port indicated by each node in
exchange and the interface on which the adjacency has been formed as the LIE exchange and the interface on which the adjacency has been
destination. It SHOULD use TTL of 1 as well. formed as destination. It SHOULD use TTL of 1 as well.
TIEs contain sequence numbers, lifetimes and a type. Each type has a TIEs contain sequence numbers, lifetimes and a type. Each type has a
large identifying number space and information is spread across large identifying number space and information is spread across
possibly many TIEs of a certain type by the means of a hash function possibly many TIEs of a certain type by the means of a hash function
that a node or deployment can individually determine. One extreme that a node or deployment can individually determine. One extreme
point of the design space is a prefix per TIE which leads to BGP-like design choice is a prefix per TIE which leads to more BGP-like
behavior vs. dense packing into few TIEs leading to more traditional behavior where small increments are only advertised on route changes
IGP trade-off with fewer TIEs. An implementation may even rehash at vs. deploying with dense prefix packing into few TIEs leading to more
the cost of significant amount of re-advertisements of TIEs. traditional IGP trade-off with fewer TIEs. An implementation may
even rehash prefix to TIE mapping at any time at the cost of
significant amount of re-advertisements of TIEs.
More information about the TIE structure can be found in the schema More information about the TIE structure can be found in the schema
in Appendix A. in Appendix A.
4.2.3.2. South- and Northbound Representation 5.2.3.2. South- and Northbound Representation
As a central concept to RIFT, each node represents itself differently A central concept of RIFT is that each node represents itself
depending on the direction in which it is advertising information. differently depending on the direction in which it is advertising
More precisely, a spine node represents two different databases to information. More precisely, a spine node represents two different
its neighbors depending whether it advertises TIEs to the north or to databases over its adjacencies depending whether it advertises TIEs
the south/sideways. We call those differing TIE databases either to the north or to the south/sideways. We call those differing TIE
south- or northbound (S-TIEs and N-TIEs) depending on the direction databases either south- or northbound (S-TIEs and N-TIEs) depending
of distribution. on the direction of distribution.
The N-TIEs hold all of the node's adjacencies, local prefixes and The N-TIEs hold all of the node's adjacencies and local prefixes
northbound policy-guided prefixes while the S-TIEs hold only all of while the S-TIEs hold only all of the node's adjacencies, the default
the node's adjacencies, the default prefix with necessary prefix with necessary disaggregated prefixes and local prefixes. We
disaggregated prefixes, local prefixes and southbound policy-guided will explain this in detail further in Section 5.2.5.
prefixes. We will explain this in detail further in Section 4.2.8
and Section 4.2.4.
The TIE types are symmetric in both directions and Table 1 provides a The TIE types are symmetric in both directions and Table 2 provides a
quick reference to the different TIE types including direction and quick reference to main TIE types including direction and their
their function. function.
+----------+--------------------------------------------------------+ +----------+--------------------------------------------------------+
| TIE-Type | Content | | TIE-Type | Content |
+----------+--------------------------------------------------------+ +----------+--------------------------------------------------------+
| node | node properties, adjacencies and information helping | | node | node properties, adjacencies and information helping |
| N-TIE | in complex disaggregation scenarios | | N-TIE | in complex disaggregation scenarios |
+----------+--------------------------------------------------------+ +----------+--------------------------------------------------------+
| node | same content as node N-TIE except the information to | | node | same content as node N-TIE except the information to |
| S-TIE | help disaggregation | | S-TIE | help disaggregation |
+----------+--------------------------------------------------------+ +----------+--------------------------------------------------------+
| Prefix | contains nodes' directly reachable prefixes | | Prefix | contains nodes' directly reachable prefixes |
| N-TIE | | | N-TIE | |
+----------+--------------------------------------------------------+ +----------+--------------------------------------------------------+
| Prefix | contains originated defaults and de-aggregated | | Prefix | contains originated defaults and de-aggregated |
| S-TIE | prefixes | | S-TIE | prefixes |
+----------+--------------------------------------------------------+ +----------+--------------------------------------------------------+
| PGP | contains nodes north PGPs |
| N-TIE | |
+----------+--------------------------------------------------------+
| PGP | contains nodes south PGPs |
| S-TIE | |
+----------+--------------------------------------------------------+
| KV | contains nodes northbound KVs | | KV | contains nodes northbound KVs |
| N-TIE | | | N-TIE | |
+----------+--------------------------------------------------------+ +----------+--------------------------------------------------------+
| KV | contains nodes southbound KVs | | KV | contains nodes southbound KVs |
| S-TIE | | | S-TIE | |
+----------+--------------------------------------------------------+ +----------+--------------------------------------------------------+
Table 1: TIE Types Table 2: TIE Types
As an example illustrating a databases holding both representations, As an example illustrating a databases holding both representations,
consider the topology in Figure 2 with the optional link between node consider the topology in Figure 2 with the optional link between
111 and node 112 (so that the flooding on an East-West link can be spine 111 and spine 112 (so that the flooding on an East-West link
shown). This example assumes unnumbered interfaces. First, here are can be shown). This example assumes unnumbered interfaces. First,
the TIEs generated by some nodes. For simplicity, the key value here are the TIEs generated by some nodes. For simplicity, the key
elements and the PGP elements which may be included in their S-TIEs value elements which may be included in their S-TIEs or N-TIEs are
or N-TIEs are not shown. not shown.
Spine21 S-TIEs: Spine21 S-TIEs:
Node S-TIE: Node S-TIE:
NodeElement(level=2, neighbors((Node111, level 1, cost 1), NodeElement(level=2, neighbors((Spine 111, level 1, cost 1),
(Node112, level 1, cost 1), (Node121, level 1, cost 1), (Spine 112, level 1, cost 1), (Spine 121, level 1, cost 1),
(Node122, level 1, cost 1))) (Spine 122, level 1, cost 1)))
Prefix S-TIE: Prefix S-TIE:
SouthPrefixesElement(prefixes(0/0, cost 1), (::/0, cost 1)) SouthPrefixesElement(prefixes(0/0, cost 1), (::/0, cost 1))
Node111 S-TIEs: Spine 111 S-TIEs:
Node S-TIE: Node S-TIE:
NodeElement(level=1, neighbors((Spine21, level 2, cost 1, links(...)), NodeElement(level=1, neighbors((Spine21, level 2, cost 1, links(...)),
(Spine22, level 2, cost 1, links(...)), (Spine22, level 2, cost 1, links(...)),
(Node112, level 1, cost 1, links(...)), (Spine 112, level 1, cost 1, links(...)),
(Leaf111, level 0, cost 1, links(...)), (Leaf111, level 0, cost 1, links(...)),
(Leaf112, level 0, cost 1, links(...)))) (Leaf112, level 0, cost 1, links(...))))
Prefix S-TIE: Prefix S-TIE:
SouthPrefixesElement(prefixes(0/0, cost 1), (::/0, cost 1)) SouthPrefixesElement(prefixes(0/0, cost 1), (::/0, cost 1))
Node111 N-TIEs: Spine 111 N-TIEs:
Node N-TIE: Node N-TIE:
NodeElement(level=1, NodeElement(level=1,
neighbors((Spine21, level 2, cost 1, links(...)), neighbors((Spine21, level 2, cost 1, links(...)),
(Spine22, level 2, cost 1, links(...)), (Spine22, level 2, cost 1, links(...)),
(Node112, level 1, cost 1, links(...)), (Spine 112, level 1, cost 1, links(...)),
(Leaf111, level 0, cost 1, links(...)), (Leaf111, level 0, cost 1, links(...)),
(Leaf112, level 0, cost 1, links(...)))) (Leaf112, level 0, cost 1, links(...))))
Prefix N-TIE: Prefix N-TIE:
NorthPrefixesElement(prefixes(Node111.loopback) NorthPrefixesElement(prefixes(Spine 111.loopback)
Node121 S-TIEs: Spine 121 S-TIEs:
Node S-TIE: Node S-TIE:
NodeElement(level=1, neighbors((Spine21,level 2,cost 1), NodeElement(level=1, neighbors((Spine21,level 2,cost 1),
(Spine22, level 2, cost 1), (Leaf121, level 0, cost 1), (Spine22, level 2, cost 1), (Leaf121, level 0, cost 1),
(Leaf122, level 0, cost 1))) (Leaf122, level 0, cost 1)))
Prefix S-TIE: Prefix S-TIE:
SouthPrefixesElement(prefixes(0/0, cost 1), (::/0, cost 1)) SouthPrefixesElement(prefixes(0/0, cost 1), (::/0, cost 1))
Node121 N-TIEs: Spine 121 N-TIEs:
Node N-TIE: Node N-TIE:
NodeElement(level=1, NodeElement(level=1,
neighbors((Spine21, level 2, cost 1, links(...)), neighbors((Spine21, level 2, cost 1, links(...)),
(Spine22, level 2, cost 1, links(...)), (Spine22, level 2, cost 1, links(...)),
(Leaf121, level 0, cost 1, links(...)), (Leaf121, level 0, cost 1, links(...)),
(Leaf122, level 0, cost 1, links(...)))) (Leaf122, level 0, cost 1, links(...))))
Prefix N-TIE: Prefix N-TIE:
NorthPrefixesElement(prefixes(Node121.loopback) NorthPrefixesElement(prefixes(Spine 121.loopback)
Leaf112 N-TIEs: Leaf112 N-TIEs:
Node N-TIE: Node N-TIE:
NodeElement(level=0, NodeElement(level=0,
neighbors((Node111, level 1, cost 1, links(...)), neighbors((Spine 111, level 1, cost 1, links(...)),
(Node112, level 1, cost 1, links(...)))) (Spine 112, level 1, cost 1, links(...))))
Prefix N-TIE: Prefix N-TIE:
NorthPrefixesElement(prefixes(Leaf112.loopback, Prefix112, NorthPrefixesElement(prefixes(Leaf112.loopback, Prefix112,
Prefix_MH)) Prefix_MH))
Figure 3: example TIES generated in a 2 level spine-and-leaf topology Figure 15: example TIES generated in a 2 level spine-and-leaf
topology
4.2.3.3. Flooding 5.2.3.3. Flooding
The mechanism used to distribute TIEs is the well-known (albeit The mechanism used to distribute TIEs is the well-known (albeit
modified in several respects to address fat tree requirements) modified in several respects to address fat tree requirements)
flooding mechanism used by today's link-state protocols. Although flooding mechanism used by today's link-state protocols. Although
flloding is initially more demanding to implement it avoids many flooding is initially more demanding to implement it avoids many
problems with update style used in diffused computation such as path problems with update style used in diffused computation such as path
vector protocols. Since flooding tends to present an unscalable vector protocols. Since flooding tends to present an unscalable
burden in large, densely meshed topologies (fat trees being burden in large, densely meshed topologies (fat trees being
unfortunately such a topology) we provide as solution a close to unfortunately such a topology) we provide as solution a close to
optimal global flood reduction and load balancing optimization in optimal global flood reduction and load balancing optimization in
Section 4.2.3.8. Section 5.2.3.8.
As described before, TIEs themselves are transported over UDP with As described before, TIEs themselves are transported over UDP with
the ports indicated in the LIE exchanges and using the destination the ports indicated in the LIE exchanges and using the destination
address (for unnumbered IPv4 interfaces same considerations apply as address on which the LIE adjacency has been formed. For unnumbered
in equivalent OSPF case) on which the LIE adjacency has been formed. IPv4 interfaces same considerations apply as in equivalent OSPF case.
On reception of a TIE with an undefined level value in the packet On reception of a TIE with an undefined level value in the packet
header the node SHOULD issue a warning and indiscriminately discard header the node SHOULD issue a warning and indiscriminately discard
the packet. the packet.
Precise finite state machines and procedures will be provided in Precise finite state machines and procedures can be found in
later versions of this specification. Appendix B.3.
4.2.3.4. TIE Flooding Scopes 5.2.3.4. TIE Flooding Scopes
In a somewhat analogous fashion to link-local, area and domain In a somewhat analogous fashion to link-local, area and domain
flooding scopes, RIFT defines several complex "flooding scopes" flooding scopes, RIFT defines several complex "flooding scopes"
depending on the direction and type of TIE propagated. depending on the direction and type of TIE propagated.
Every N-TIE is flooded northbound, providing a node at a given level Every N-TIE is flooded northbound, providing a node at a given level
with the complete topology of the Clos or Fat Tree network underneath with the complete topology of the Clos or Fat Tree network underneath
it, including all specific prefixes. This means that a packet it, including all specific prefixes. This means that a packet
received from a node at the same or lower level whose destination is received from a node at the same or lower level whose destination is
covered by one of those specific prefixes may be routed directly covered by one of those specific prefixes may be routed directly
towards the node advertising that prefix rather than sending the towards the node advertising that prefix rather than sending the
packet to a node at a higher level. packet to a node at a higher level.
A node's Node S-TIEs, consisting of all node's adjacencies and prefix A node's Node S-TIEs, consisting of all node's adjacencies and prefix
S-TIEs limited to those related to default IP prefix and S-TIEs limited to those related to default IP prefix and
disaggregated prefixes, are flooded southbound in order to allow the disaggregated prefixes, are flooded southbound in order to allow the
nodes one level down to see connectivity of the higher level as well nodes one level down to see connectivity of the higher level as well
as reachability to the rest of the fabric. In order to allow a E-W as reachability to the rest of the fabric. In order to allow an E-W
disconnected node in a given level to receive the S-TIEs of other disconnected node in a given level to receive the S-TIEs of other
nodes at its level, every *NODE* S-TIE is "reflected" northbound to nodes at its level, every *NODE* S-TIE is "reflected" northbound to
level from which it was received. It should be noted that East-West level from which it was received. It should be noted that East-West
links are included in South TIE flooding; those TIEs need to be links are included in South TIE flooding; those TIEs need to be
flooded to satisfy algorithms in Section 4.2.5. In that way nodes at flooded to satisfy algorithms in Section 5.2.4. In that way nodes at
same level can learn about each other without a lower level, e.g. in same level can learn about each other without a lower level, e.g. in
case of leaf level. The precise flooding scopes are given in case of leaf level. The precise flooding scopes are given in
Table 2. Those rules govern as well what SHOULD be included in TIDEs Table 3. Those rules govern as well what SHOULD be included in TIDEs
towards neighbors. East-West flooding scopes are identical to South on the adjacency. East-West flooding scopes are identical to South
flooding scopes. flooding scopes except in case of ToF East-West links (rings).
Node S-TIE "reflection" allows to support disaggregation on failures Node S-TIE "south reflection" allows to support positive
describes in Section 4.2.8 and flooding reduction in Section 4.2.3.8. disaggregation on failures describes in Section 5.2.5 and flooding
reduction in Section 5.2.3.8.
+--------------+----------------------------+-----------------------+ +-----------+---------------------+---------------+-----------------+
| Packet Type | South | North | | Type / | South | North | East-West |
| vs. Peer | | | | Direction | | | |
| Direction | | | +-----------+---------------------+---------------+-----------------+
+--------------+----------------------------+-----------------------+ | node | flood if level of | flood if | flood only if |
| node S-TIE | flood self-originated only | flood if TIE | | S-TIE | originator is equal | level of | this node is |
| | | originator's level is | | | to this node | originator is | not ToF |
| | | higher than own level | | | | higher than | |
+--------------+----------------------------+-----------------------+ | | | this node | |
| non-node | flood self-originated only | flood only if TIE | +-----------+---------------------+---------------+-----------------+
| S-TIE | | originator is equal | | non-node | flood self- | flood only if | flood only if |
| | | peer | | S-TIE | originated only | neighbor is | self-originated |
+--------------+----------------------------+-----------------------+ | | | originator of | and this node |
| all N-TIEs | never flood | flood always | | | | TIE | is not ToF |
+--------------+----------------------------+-----------------------+ +-----------+---------------------+---------------+-----------------+
| TIDE | include TIEs in flooding | include TIEs in | | all | never flood | flood always | flood only if |
| | scope | flooding scope | | N-TIEs | | | this node is |
+--------------+----------------------------+-----------------------+ | | | | ToF |
| TIRE | include all N-TIEs and all | include only if TIE | +-----------+---------------------+---------------+-----------------+
| | peer's self-originated | originator is equal | | TIDE | include at least | include at | if this node is |
| | TIEs and all node S-TIEs | peer | | | all non-self | least all | ToF then |
+--------------+----------------------------+-----------------------+ | | originated N-TIE | node S-TIEs | include all |
| | headers and self- | and all | N-TIEs, |
| | originated S-TIE | S-TIEs | otherwise only |
| | headers and node | originated by | self-originated |
| | S-TIEs of nodes at | peer and all | TIEs |
| | same level | N-TIEs | |
+-----------+---------------------+---------------+-----------------+
| TIRE as | request all N-TIEs | request all | if this node is |
| Request | and all peer's | S-TIEs | ToF then apply |
| | self-originated | | North scope |
| | TIEs and all node | | rules, |
| | S-TIEs | | otherwise South |
| | | | scope rules |
+-----------+---------------------+---------------+-----------------+
| TIRE as | Ack all received | Ack all | Ack all |
| Ack | TIEs | received TIEs | received TIEs |
+-----------+---------------------+---------------+-----------------+
Table 2: Flooding Scopes Table 3: Flooding Scopes
If the TIDE includes additional TIE headers beside the ones
specified, the receiving neighbor must apply according filter to the
received TIDE strictly and MUST NOT request the extra TIE headers
that were not allowed by the flooding scope rules in its direction.
As an example to illustrate these rules, consider using the topology As an example to illustrate these rules, consider using the topology
in Figure 2, with the optional link between node 111 and node 112, in Figure 2, with the optional link between spine 111 and spine 112,
and the associated TIEs given in Figure 3. The flooding from and the associated TIEs given in Figure 15. The flooding from
particular nodes of the TIEs is given in Table 3. particular nodes of the TIEs is given in Table 4.
+------------+----------+-------------------------------------------+ +-------------+----------+------------------------------------------+
| Router | Neighbor | TIEs | | Router | Neighbor | TIEs |
| floods to | | | | floods to | | |
+------------+----------+-------------------------------------------+ +-------------+----------+------------------------------------------+
| Leaf111 | Node112 | Leaf111 N-TIEs, Node111 node S-TIE | | Leaf111 | Spine | Leaf111 N-TIEs, Spine 111 node S-TIE |
| Leaf111 | Node111 | Leaf111 N-TIEs, Node112 node S-TIE | | | 112 | |
| | | | | Leaf111 | Spine | Leaf111 N-TIEs, Spine 112 node S-TIE |
| Node111 | Leaf111 | Node111 S-TIEs | | | 111 | |
| Node111 | Leaf112 | Node111 S-TIEs | | | | |
| Node111 | Node112 | Node111 S-TIEs | | Spine 111 | Leaf111 | Spine 111 S-TIEs |
| Node111 | Spine21 | Node111 N-TIEs, Leaf111 N-TIEs, Leaf112 | | Spine 111 | Leaf112 | Spine 111 S-TIEs |
| | | N-TIEs, Spine22 node S-TIE | | Spine 111 | Spine | Spine 111 S-TIEs |
| Node111 | Spine22 | Node111 N-TIEs, Leaf111 N-TIEs, Leaf112 | | | 112 | |
| | | N-TIEs, Spine21 node S-TIE | | Spine 111 | Spine21 | Spine 111 N-TIEs, Leaf111 N-TIEs, |
| | | | | | | Leaf112 N-TIEs, Spine22 node S-TIE |
| ... | ... | ... | | Spine 111 | Spine22 | Spine 111 N-TIEs, Leaf111 N-TIEs, |
| Spine21 | Node111 | Spine21 S-TIEs | | | | Leaf112 N-TIEs, Spine21 node S-TIE |
| Spine21 | Node112 | Spine21 S-TIEs | | | | |
| Spine21 | Node121 | Spine21 S-TIEs | | ... | ... | ... |
| Spine21 | Node122 | Spine21 S-TIEs | | Spine21 | Spine | Spine21 S-TIEs |
| ... | ... | ... | | | 111 | |
+------------+----------+-------------------------------------------+ | Spine21 | Spine | Spine21 S-TIEs |
| | 112 | |
| Spine21 | Spine | Spine21 S-TIEs |
| | 121 | |
| Spine21 | Spine | Spine21 S-TIEs |
| | 122 | |
| ... | ... | ... |
+-------------+----------+------------------------------------------+
Table 3: Flooding some TIEs from example topology Table 4: Flooding some TIEs from example topology
4.2.3.5. Initial and Periodic Database Synchronization 5.2.3.5. Initial and Periodic Database Synchronization
The initial exchange of RIFT is modeled after ISIS with TIDE being The initial exchange of RIFT is modeled after ISIS with TIDE being
equivalent to CSNP and TIRE playing the role of PSNP. The content of equivalent to CSNP and TIRE playing the role of PSNP. The content of
TIDEs and TIREs is governed by Table 2. TIDEs and TIREs is governed by Table 3.
4.2.3.6. Purging 5.2.3.6. Purging
RIFT does not purge information that has been distributed by the RIFT does not purge information that has been distributed by the
protocol. Purging mechanisms in other routing protocols have proven protocol. Purging mechanisms in other routing protocols have proven
to be complex and fragile over many years of experience. Abundant to be complex and fragile over many years of experience. Abundant
amounts of memory are available today even on low-end platforms. The amounts of memory are available today even on low-end platforms. The
information will age out and all computations will deliver correct information will age out and all computations will deliver correct
results if a node leaves the network due to the new information results if a node leaves the network due to the new information
distributed by its adjacent nodes. distributed by its adjacent nodes.
Once a RIFT node issues a TIE with an ID, it MUST preserve the ID as Once a RIFT node issues a TIE with an ID, it MUST preserve the ID as
long as feasible (also when the protocol restarts), even if the TIE long as feasible (also when the protocol restarts), even if the TIE
looses all content. The re-advertisement of empty TIE fulfills the looses all content. The re-advertisement of empty TIE fulfills the
purpose of purging any information advertised in previous versions. purpose of purging any information advertised in previous versions.
The originator is free to not re-originate the according empty TIE The originator is free to not re-originate the according empty TIE
again or originate an empty TIE with relatively short lifetime to again or originate an empty TIE with relatively short lifetime to
prevent large number of long-lived empty stubs polluting the network. prevent large number of long-lived empty stubs polluting the network.
Each node MUST timeout and clean up the according empty TIEs
Each node will timeout and clean up the according empty TIEs
independently. independently.
Upon restart a node MUST, as any link-state implementation, be Upon restart a node MUST, as any link-state implementation, be
prepared to receive TIEs with its own system ID and supercede them prepared to receive TIEs with its own system ID and supercede them
with equivalent, newly generated, empty TIEs with a higher sequence with equivalent, newly generated, empty TIEs with a higher sequence
number. As above, the lifetime can be relatively short since it only number. As above, the lifetime can be relatively short since it only
needs to exceed the necessary propagation and processing delay by all needs to exceed the necessary propagation and processing delay by all
the nodes that are within the TIE's flooding scope. the nodes that are within the TIE's flooding scope.
4.2.3.7. Southbound Default Route Origination 5.2.3.7. Southbound Default Route Origination
Under certain conditions nodes issue a default route in their South Under certain conditions nodes issue a default route in their South
Prefix TIEs with metrics as computed in Section 4.3.6.1. Prefix TIEs with costs as computed in Section 5.3.6.1.
A node X that A node X that
1. is NOT overloaded AND 1. is NOT overloaded AND
2. has southbound or East-West adjacencies 2. has southbound or East-West adjacencies
originates in its south prefix TIE such a default route IIF originates in its south prefix TIE such a default route IIF
1. all other nodes at X's' level are overloaded OR 1. all other nodes at X's' level are overloaded OR
2. all other nodes at X's' level have NO northbound adjacencies OR 2. all other nodes at X's' level have NO northbound adjacencies OR
3. X has computed reachability to a default route during N-SPF. 3. X has computed reachability to a default route during N-SPF.
The term "all other nodes at X's' level" describes obviously just the The term "all other nodes at X's' level" describes obviously just the
nodes at the same level in the POD with a viable lower level nodes at the same level in the PoD with a viable lower level
(otherwise the node S-TIEs cannot be reflected and the nodes in e.g. (otherwise the node S-TIEs cannot be reflected and the nodes in e.g.
POD 1 and POD 2 are "invisible" to each other). PoD 1 and PoD 2 are "invisible" to each other).
A node originating a southbound default route MUST install a default A node originating a southbound default route MUST install a default
discard route if it did not compute a default route during N-SPF. discard route if it did not compute a default route during N-SPF.
4.2.3.8. Northbound TIE Flooding Reduction 5.2.3.8. Northbound TIE Flooding Reduction
Section 1.4 of the Optimized Link State Routing Protocol [RFC3626] Section 1.4 of the Optimized Link State Routing Protocol [RFC3626]
(OLSR) introduces the concept of a "multipoint relay" (MPR) that (OLSR) introduces the concept of a "multipoint relay" (MPR) that
minimize the overhead of flooding messages in the network by reducing minimize the overhead of flooding messages in the network by reducing
redundant retransmissions in the same region. redundant retransmissions in the same region.
A similar technique is applied to RIFT to control northbound A similar technique is applied to RIFT to control northbound
flooding. Important observations first: flooding. Important observations first:
1. a node MUST flood self-originated N-TIE to all the reachable 1. a node MUST flood self-originated N-TIE to all the reachable
skipping to change at page 24, line 9 skipping to change at page 41, line 18
without any need for synchronization amongst nodes. In a "PoD" without any need for synchronization amongst nodes. In a "PoD"
structure, where the Level L+2 is partitioned in silos of equivalent structure, where the Level L+2 is partitioned in silos of equivalent
grandparents that are only reachable from respective parents, this grandparents that are only reachable from respective parents, this
means treating each silo as a fully connected Clos Network and solve means treating each silo as a fully connected Clos Network and solve
the problem within the silo. the problem within the silo.
In terms of signaling, a node has enough information to select its In terms of signaling, a node has enough information to select its
set of FRs; this information is derived from the node's parents' Node set of FRs; this information is derived from the node's parents' Node
S-TIEs, which indicate the parent's reachable northbound adjacencies S-TIEs, which indicate the parent's reachable northbound adjacencies
to its own parents, i.e. the node's grandparents. An optional to its own parents, i.e. the node's grandparents. An optional
boolean information `you_are_not_flood_repeater` in a LIE packet to a boolean information `you_are_flood_repeater` in a LIE packet to a
parent is set to indicate that the parent is not an FR and that it parent is set to indicate whether the parent lost its flood repeater
SHOULD NOT reflood N-TIEs. leadership and with that SHOULD NOT reflood N-TIEs.
This specification proposes a simple default algorithm that SHOULD be This specification proposes a simple default algorithm that SHOULD be
implemented and used by default on every RIFT node. implemented and used by default on every RIFT node.
o let |NA(Node) be the set of Northbound adjacencies of node Node o let |NA(Node) be the set of Northbound adjacencies of node Node
and CN(Node) be the cardinality of |NA(Node); and CN(Node) be the cardinality of |NA(Node);
o let |SA(Node) be the set of Southbound adjacencies of node Node o let |SA(Node) be the set of Southbound adjacencies of node Node
and CS(Node) be the cardinality of |SA(Node); and CS(Node) be the cardinality of |SA(Node);
skipping to change at page 24, line 45 skipping to change at page 42, line 5
reachability of A(P, G); reachability of A(P, G);
o let R be a redundancy constant integer; a value of 2 or higher for o let R be a redundancy constant integer; a value of 2 or higher for
R is RECOMMENDED; R is RECOMMENDED;
o let S be a similarity constant integer; a value in range 0 .. 2 o let S be a similarity constant integer; a value in range 0 .. 2
for S is RECOMMENDED, the value of 1 SHOULD be used. Two for S is RECOMMENDED, the value of 1 SHOULD be used. Two
cardinalities are considered as equivalent if their absolute cardinalities are considered as equivalent if their absolute
difference is less than or equal to S, i.e. |a-b|<=S. difference is less than or equal to S, i.e. |a-b|<=S.
o let RND be a 64-bit random number generated by the system once on
startup.
The algorithm consists of the following steps: The algorithm consists of the following steps:
1. derive a 16-bits pseudo-random unsigned integer PR(N) from N's 1. derive a 64-bits number by XOR'ing 'N's system ID with RND and
system ID by splitting it in 16-bits-long words W1, W2, ..., Wn then
and then XOR'ing the circularly shifted resulting words together,
and casting the resulting representation: 2. derive a 16-bits pseudo-random unsigned integer PR(N) from the
resulting 64-bits number by splitting it in 16-bits-long words
W1, W2, ..., Wn and then XOR'ing the circularly shifted resulting
words together, and casting the resulting representation:
1. (unsigned integer) (W1<<1 xor (W2<<2) xor ... xor (Wn<<n) ); 1. (unsigned integer) (W1<<1 xor (W2<<2) xor ... xor (Wn<<n) );
2. sort the parents by decreasing number of northbound adjacencies: and then
3. sort the parents by decreasing number of northbound adjacencies:
sort |P(N) by decreasing CN(P), for all P in |P(N), as ordered sort |P(N) by decreasing CN(P), for all P in |P(N), as ordered
array |A(N); array |A(N) and then
3. partition |A(N) in subarrays |A_k(N) of parents with equivalent 4. partition |A(N) in subarrays |A_k(N) of parents with equivalent
cardinality of northbound adjacencies (in other words with cardinality of northbound adjacencies (in other words with
equivalent number of grandparents they can reach): equivalent number of grandparents they can reach):
1. set k=0; // k is the ID of the subarrray 1. set k=0; // k is the ID of the subarrray
2. set i=0; 2. set i=0;
3. while i < CN(N) do 3. while i < CN(N) do
1. set k=k+1; 1. set k=k+1;
skipping to change at page 25, line 37 skipping to change at page 43, line 5
2. set i=i+1; 2. set i=i+1;
/* At this point j is the index in |A(N) of the first member /* At this point j is the index in |A(N) of the first member
of |A_k(N) and (i-j) is C_k(N) defined as the cardinality of |A_k(N) and (i-j) is C_k(N) defined as the cardinality
of |A_k(N) */ of |A_k(N) */
/* At this point k is the total number of subarrays, initialized /* At this point k is the total number of subarrays, initialized
for the shuffling operation below */ for the shuffling operation below */
4. shuffle individually each subarrays |A_k(N) of cardinality C_k(N) 5. shuffle individually each subarrays |A_k(N) of cardinality C_k(N)
within |A(N) using a Fisher-Yates method that depends on N's within |A(N) using a Fisher-Yates method that depends on N's
System ID: System ID:
1. while k > 0 do 1. while k > 0 do
1. for i from C_k(N)-1 to 1 decrementing by 1 do 1. for i from C_k(N)-1 to 1 decrementing by 1 do
1. set j to PR(N) modulo i; 1. set j to PR(N) modulo i;
2. exchange |A_k[j] and |A_k[i]; 2. exchange |A_k[j] and |A_k[i];
2. set k=k-1; 2. set k=k-1;
5. for each grandparent, initialize a counter with the number of its 6. for each grandparent, initialize a counter with the number of its
Southbound adjacencies : Southbound adjacencies :
1. for each G in |G(N) set c(G) = CS(G); 1. for each G in |G(N) set c(G) = CS(G);
6. finally keep as FRs only parents that are needed to maintain the and
7. finally keep as FRs only parents that are needed to maintain the
number of adjacencies between the FRs and any grandparent G equal number of adjacencies between the FRs and any grandparent G equal
or above the redundancy constant R: or above the redundancy constant R:
1. for each P in reshuffled |A(N); 1. for each P in reshuffled |A(N);
1. if there exists an adjacency ADJ(P, G) in |NA(P) such 1. if there exists an adjacency ADJ(P, G) in |NA(P) such
that c(G) <= R then that c(G) <= R then
1. place P in FR set; 1. place P in FR set;
skipping to change at page 26, line 30 skipping to change at page 43, line 49
1. for all adjacencies ADJ(P, G) in |NA(P) 1. for all adjacencies ADJ(P, G) in |NA(P)
1. decrement c(G); 1. decrement c(G);
The algorithm MUST be re-evaluated by a node on every change of local The algorithm MUST be re-evaluated by a node on every change of local
adjacencies or reception of a parent S-TIE with changed adjacencies. adjacencies or reception of a parent S-TIE with changed adjacencies.
A node MAY apply a hysteresis to prevent excessive amount of A node MAY apply a hysteresis to prevent excessive amount of
computation during periods of network instability just like in case computation during periods of network instability just like in case
of reachability computation. of reachability computation.
4.2.4. Policy-Guided Prefixes A node SHOULD send out LIEs that grant leadership before LIEs that
revoke it on leadership changes to prevent transient behavior where
In a fat tree, it can be sometimes desirable to guide traffic to the full coverage of grand parents is not guaranteed. Albeit the
particular destinations or keep specific flows to certain paths. In condition will correct in positively stable manner due to LIE
RIFT, this is done by using policy-guided prefixes with their retransmission and periodic TIDEs, it can slow down flooding
associated communities. Each community is an abstract value whose convergence on leadership changes.
meaning is determined by configuration. It is assumed that the
fabric is under a single administrative control so that the meaning
and intent of the communities is understood by all the nodes in the
fabric. Any node can originate a policy-guided prefix.
Since RIFT uses distance vector concepts in a southbound direction,
it is straightforward to add a policy-guided prefix to an S-TIE. For
easier troubleshooting, the approach taken in RIFT is that a node's
southbound policy-guided prefixes are sent in its S-TIE and the
receiver does inbound filtering based on the associated communities
(an egress policy is imaginable but would lead to different S-TIEs
per neighbor possibly which is not considered in RIFT protocol
procedures). A southbound policy-guided prefix can only use links in
the south direction. If an PGP S-TIE is received on an East-West or
northbound link, it must be discarded by ingress filtering.
Conceptually, a southbound policy-guided prefix guides traffic from
the leaves up to at most the north-most level. It is also necessary
to to have northbound policy-guided prefixes to guide traffic from
the north-most level down to the appropriate leaves. Therefore, RIFT
includes northbound policy-guided prefixes in its N PGP-TIE and the
receiver does inbound filtering based on the associated communities.
A northbound policy-guided prefix can only use links in the northern
direction. If an N PGP TIE is received on an East-West or southbound
link, it must be discarded by ingress filtering.
By separating southbound and northbound policy-guided prefixes and
requiring that the cost associated with a PGP is strictly
monotonically increasing at each hop, the path cannot loop. Because
the costs are strictly increasing, it is not possible to have a loop
between a northbound PGP and a southbound PGP. If East-West links
were to be allowed, then looping could occur and issues such as
counting to infinity would become an issue to be solved. If complete
generality of path - such as including East-West links and using both
north and south links in arbitrary sequence - then a Path Vector
protocol or a similar solution must be considered.
If a node has received the same prefix, after ingress filtering, as a
PGP in an S-TIE and in an N-TIE, then the node determines which
policy-guided prefix to use based upon the advertised cost.
A policy-guided prefix is always preferred to a regular prefix, even
if the policy-guided prefix has a larger cost. Appendix A provides
normative indication of prefix preferences.
The set of policy-guided prefixes received in a TIE is subject to
ingress filtering and then re-originated to be sent out in the
receiver's appropriate TIE. Both the ingress filtering and the re-
origination use the communities associated with the policy-guided
prefixes to determine the correct behavior. The cost on re-
advertisement MUST increase in a strictly monotonic fashion.
4.2.4.1. Ingress Filtering
When a node X receives a PGP S-TIE or a PGP N-TIE that is originated
from a node Y which does not have an adjacency with X, all PGPs in
such a TIE MUST be filtered. Similarly, if node Y is at the same
level as node X, then X MUST filter out PGPs in such S- and N-TIEs to
prevent loops.
Next, policy can be applied to determine which policy-guided prefixes
to accept. Since ingress filtering is chosen rather than egress
filtering and per-neighbor PGPs, policy that applies to links is done
at the receiver. Because the RIFT adjacency is between nodes and
there may be parallel links between the two nodes, the policy-guided
prefix is considered to start with the next-hop set that has all
links to the originating node Y.
A policy-guided prefix has or is assigned the following attributes:
cost: This is initialized to the cost received
community_list: This is initialized to the list of the communities
received.
next_hop_set: This is initialized to the set of links to the
originating node Y.
4.2.4.2. Applying Policy
The specific action to apply based upon a community is deployment
specific. Here are some examples of things that can be done with
communities. The length of a community is a 64 bits number and it
can be written as a single field M or as a multi-field (S = M[0-31],
T = M[32-63]) in these examples. For simplicity, the policy-guided
prefix is referred to as P, the processing node as X and the
originator as Y.
Prune Next-Hops: Community Required: For each next-hop in
P.next_hop_set, if the next-hop does not have the community, prune
that next-hop from P.next_hop_set.
Prune Next-Hops: Avoid Community: For each next-hop in
P.next_hop_set, if the next-hop has the community, prune that
next-hop from P.next_hop_set.
Drop if Community: If node X has community M, discard P.
Drop if not Community: If node X does not have the community M,
discard P.
Prune to ifIndex T: For each next-hop in P.next_hop_set, if the
next-hop's ifIndex is not the value T specified in the community
(S,T), then prune that next-hop from P.next_hop_set.
Add Cost T: For each appearance of community S in P.community_list,
if the node X has community S, then add T to P.cost.
Accumulate Min-BW T: Let bw be the sum of the bandwidth for
P.next_hop_set. If that sum is less than T, then replace (S,T)
with (S, bw).
Add Community T if Node matches S: If the node X has community S,
then add community T to P.community_list.
4.2.4.3. Store Policy-Guided Prefix for Route Computation and
Regeneration
Once a policy-guided prefix has completed ingress filtering and
policy, it is almost ready to store and use. It is still necessary
to adjust the cost of the prefix to account for the link from the
computing node X to the originating neighbor node Y.
There are three different policies that can be used:
Minimum Equal-Cost: Find the loWest cost C next-hops in
P.next_hop_set and prune to those. Add C to P.cost.
Minimum Unequal-Cost: Find the loWest cost C next-hop in
P.next_hop_set. Add C to P.cost.
Maximum Unequal-Cost: Find the highest cost C next-hop in
P.next_hop_set. Add C to P.cost.
The default policy is Minimum Unequal-Cost but well-known communities
can be defined to get the other behaviors.
Regardless of the policy used, a node MUST store a PGP cost that is
at least 1 greater than the PGP cost received. This enforces the
strictly monotonically increasing condition that avoids loops.
Two databases of PGPs - from N-TIEs and from S-TIEs are stored. When
a PGP is inserted into the appropriate database, the usual tie-
breaking on cost is performed. Observe that the node retains all PGP
TIEs due to normal flooding behavior and hence loss of the best
prefix will lead to re-evaluation of TIEs present and re-
advertisement of a new best PGP.
4.2.4.4. Re-origination
A node must re-originate policy-guided prefixes and retransmit them.
The node has its database of southbound policy-guided prefixes to
send in its S-TIE and its database of northbound policy-guided
prefixes to send in its N-TIE.
Of course, a leaf does not need to re-originate southbound policy-
guided prefixes.
4.2.4.5. Overlap with Disaggregated Prefixes
PGPs may overlap with prefixes introduced by automatic de-
aggregation. The topic is under further discussion. The break in
connectivity that leads to infeasibility of a PGP is mirrored in
adjacency tear-down and according removal of such PGPs.
Nevertheless, the underlying link-state flooding will be likely
reacting significantly faster than a hop-by-hop redistribution and
with that the preference for PGPs may cause intermittent black-holes.
4.2.5. Reachability Computation 5.2.4. Reachability Computation
A node has three sources of relevant information. A node knows the A node has three sources of relevant information. A node knows the
full topology south from the received N-TIEs. A node has the set of full topology south from the received N-TIEs. A node has the set of
prefixes with associated distances and bandwidths from received prefixes with associated distances and bandwidths from received
S-TIEs. A node can also have a set of PGPs. S-TIEs.
To compute reachability, a node runs conceptually a northbound and a To compute reachability, a node runs conceptually a northbound and a
southbound SPF. We call that N-SPF and S-SPF. southbound SPF. We call that N-SPF and S-SPF.
Since neither computation can "loop" (with due considerations given Since neither computation can "loop", it is possible to compute non-
to PGPs), it is possible to compute non-equal-cost or even k-shortest equal-cost or even k-shortest paths [EPPSTEIN] and "saturate" the
paths [EPPSTEIN] and "saturate" the fabric to the extent desired. fabric to the extent desired.
4.2.5.1. Northbound SPF 5.2.4.1. Northbound SPF
N-SPF uses northbound and East-West adjacencies in the computing N-SPF uses northbound and East-West adjacencies in the computing
node's node N-TIEs (since if the node is a leaf it may not have node's node N-TIEs (since if the node is a leaf it may not have
generated a node S-TIE) when starting Dijkstra. Observe that N-SPF generated a node S-TIE) when starting Dijkstra. Observe that N-SPF
is really just a one hop variety since Node S-TIEs are not re-flooded is really just a one hop variety since Node S-TIEs are not re-flooded
southbound beyond a single level (or East-West) and with that the southbound beyond a single level (or East-West) and with that the
computation cannot progress beyond adjacent nodes. computation cannot progress beyond adjacent nodes.
Once progressing, we are using the next level's node S-TIEs to find Once progressing, we are using the next level's node S-TIEs to find
according adjacencies to verify backlink connectivity. Just as in according adjacencies to verify backlink connectivity. Just as in
skipping to change at page 30, line 51 skipping to change at page 44, line 43
together to confirm bidirectional connectivity. together to confirm bidirectional connectivity.
Default route found when crossing an E-W link is used IIF Default route found when crossing an E-W link is used IIF
1. the node itself does NOT have any northbound adjacencies AND 1. the node itself does NOT have any northbound adjacencies AND
2. the adjacent node has one or more northbound adjacencies 2. the adjacent node has one or more northbound adjacencies
This rule forms a "one-hop default route split-horizon" and prevents This rule forms a "one-hop default route split-horizon" and prevents
looping over default routes while allowing for "one-hop protection" looping over default routes while allowing for "one-hop protection"
of nodes that lost all northbound adjacencies. of nodes that lost all northbound adjacencies except at Top-of-Fabric
where the links are used exclusively to flood topology information in
multi-plane designs.
Other south prefixes found when crossing E-W link MAY be used IIF Other south prefixes found when crossing E-W link MAY be used IIF
1. no north neighbors are advertising same or supersuming non- 1. no north neighbors are advertising same or supersuming non-
default prefix AND default prefix AND
2. the node does not originate a non-default supersuming prefix 2. the node does not originate a non-default supersuming prefix
itself. itself.
i.e. the E-W link can be used as the gateway of last resort for a i.e. the E-W link can be used as the gateway of last resort for a
specific prefix only. Using south prefixes across E-W link can be specific prefix only. Using south prefixes across E-W link can be
beneficial e.g. on automatic de-aggregation in pathological fabric beneficial e.g. on automatic de-aggregation in pathological fabric
partitioning scenarios. partitioning scenarios.
A detailed example can be found in Section 5.4. A detailed example can be found in Section 6.4.
4.2.5.2. Southbound SPF 5.2.4.2. Southbound SPF
S-SPF uses only the southbound adjacencies in the node S-TIEs, i.e. S-SPF uses only the southbound adjacencies in the node S-TIEs, i.e.
progresses towards nodes at lower levels. Observe that E-W progresses towards nodes at lower levels. Observe that E-W
adjacencies are NEVER used in the computation. This enforces the adjacencies are NEVER used in the computation. This enforces the
requirement that a packet traversing in a southbound direction must requirement that a packet traversing in a southbound direction must
never change its direction. never change its direction.
S-SPF uses northbound adjacencies in node N-TIEs to verify backlink S-SPF uses northbound adjacencies in node N-TIEs to verify backlink
connectivity. connectivity.
4.2.5.3. East-West Forwarding Within a Level 5.2.4.3. East-West Forwarding Within a Level
Ultimately, it should be observed that in presence of a "ring" of E-W Ultimately, it should be observed that in presence of a "ring" of E-W
links in a level neither SPF will provide a "ring protection" scheme links in a level neither SPF will provide a "ring protection" scheme
since such a computation would have to deal necessarily with breaking since such a computation would have to deal necessarily with breaking
of "loops" in generic Dijkstra sense; an application for which RIFT of "loops" in generic Dijkstra sense; an application for which RIFT
is not intended. It is outside the scope of this document how an is not intended. It is outside the scope of this document how an
underlay can be used to provide a full-mesh connectivity between underlay can be used to provide a full-mesh connectivity between
nodes in the same level that would allow for N-SPF to provide nodes in the same level that would allow for N-SPF to provide
protection for a single node loosing all its northbound adjacencies protection for a single node loosing all its northbound adjacencies
(as long as any of the other nodes in the level are northbound (as long as any of the other nodes in the level are northbound
connected). connected).
Using south prefixes over horizontal links is optional and can Using south prefixes over horizontal links is optional and can
protect against pathological fabric partitioning cases that leave protect against pathological fabric partitioning cases that leave
only paths to destinations that would necessitate multiple changes of only paths to destinations that would necessitate multiple changes of
forwarding direction between north and south. forwarding direction between north and south.
4.2.6. Attaching Prefixes 5.2.5. Automatic Disaggregation on Link & Node Failures
After the SPF is run, it is necessary to attach according prefixes.
For S-SPF, prefixes from an N-TIE are attached to the originating
node with that node's next-hop set and a distance equal to the
prefix's cost plus the node's minimized path distance. The RIFT
route database, a set of (prefix, type=spf, path_distance, next-hop
set), accumulates these results. Obviously, the prefix retains its
type which is used to tie-break between the same prefix advertised
with different types.
In case of N-SPF prefixes from each S-TIE need to also be added to
the RIFT route database. The N-SPF is really just a stub so the
computing node needs simply to determine, for each prefix in an S-TIE
that originated from adjacent node, what next-hops to use to reach
that node. Since there may be parallel links, the next-hops to use
can be a set; presence of the computing node in the associated Node
S-TIE is sufficient to verify that at least one link has
bidirectional connectivity. The set of minimum cost next-hops from
the computing node X to the originating adjacent node is determined.
Each prefix has its cost adjusted before being added into the RIFT
route database. The cost of the prefix is set to the cost received
plus the cost of the minimum cost next-hop to that neighbor. Then
each prefix can be added into the RIFT route database with the
next_hop_set; ties are broken based upon type first and then
distance. RIFT route preferences are normalized by the according
thrift model type.
An exemplary implementation for node X follows:
for each S-TIE
if S-TIE.level > X.level
next_hop_set = set of minimum cost links to the S-TIE.originator
next_hop_cost = minimum cost link to S-TIE.originator
end if
for each prefix P in the S-TIE
P.cost = P.cost + next_hop_cost
if P not in route_database:
add (P, type=DistVector, P.cost, next_hop_set) to route_database
end if
if (P in route_database) and
(route_database[P].type is not PolicyGuided):
if route_database[P].cost > P.cost):
update route_database[P] with (P, DistVector, P.cost, next_hop_set)
else if route_database[P].cost == P.cost
update route_database[P] with (P, DistVector, P.cost,
merge(next_hop_set, route_database[P].next_hop_set))
else
// Not preferred route so ignore
end if
end if
end for
end for
Figure 4: Adding Routes from S-TIE Prefixes
4.2.7. Attaching Policy-Guided Prefixes
Each policy-guided prefix P has its cost and next_hop_set already
stored in the associated database, as specified in Section 4.2.4.3;
the cost stored for the PGP is already updated to considering the
cost of the link to the advertising neighbor. By definition, a
policy-guided prefix is preferred to a regular prefix.
for each policy-guided prefix P:
if P not in route_database:
add (P, type=PolicyGuided, P.cost, next_hop_set)
end if
if P in route_database :
if (route_database[P].type is not PolicyGuided) or
(route_database[P].cost > P.cost):
update route_database[P] with (P, PolicyGuided, P.cost, next_hop_set)
else if route_database[P].cost == P.cost
update route_database[P] with (P, PolicyGuided, P.cost,
merge(next_hop_set, route_database[P].next_hop_set))
else
// Not preferred route so ignore
end if
end if
end for
Figure 5: Adding Routes from Policy-Guided Prefixes
4.2.8. Automatic Disaggregation on Link & Node Failures 5.2.5.1. Positive, Non-transitive Disaggregation
Under normal circumstances, node's S-TIEs contain just the Under normal circumstances, node's S-TIEs contain just the
adjacencies, a default route and policy-guided prefixes. However, if adjacencies and a default route. However, if a node detects that its
a node detects that its default IP prefix covers one or more prefixes default IP prefix covers one or more prefixes that are reachable
that are reachable through it but not through one or more other nodes through it but not through one or more other nodes at the same level,
at the same level, then it MUST explicitly advertise those prefixes then it MUST explicitly advertise those prefixes in an S-TIE.
in an S-TIE. Otherwise, some percentage of the northbound traffic
for those prefixes would be sent to nodes without according
reachability, causing it to be black-holed. Even when not black-
holing, the resulting forwarding could 'backhaul' packets through the
higher level spines, clearly an undesirable condition affecting the
blocking probabilities of the fabric.
We refer to the process of advertising additional prefixes as 'de- Otherwise, some percentage of the northbound traffic for those
aggregation' or 'dis-aggregation'. prefixes would be sent to nodes without according reachability,
causing it to be black-holed. Even when not black-holing, the
resulting forwarding could 'backhaul' packets through the higher
level spines, clearly an undesirable condition affecting the blocking
probabilities of the fabric.
We refer to the process of advertising additional prefixes southbound
as 'positive de-aggregation' or 'positive dis-aggregation'.
A node determines the set of prefixes needing de-aggregation using A node determines the set of prefixes needing de-aggregation using
the following steps: the following steps:
1. A DAG computation in the southern direction is performed first, 1. A DAG computation in the southern direction is performed first,
i.e. the N-TIEs are used to find all of prefixes it can reach and i.e. the N-TIEs are used to find all of prefixes it can reach and
the set of next-hops in the lower level for each. Such a the set of next-hops in the lower level for each of them. Such a
computation can be easily performed on a fat tree by e.g. setting computation can be easily performed on a fat tree by e.g. setting
all link costs in the southern direction to 1 and all northern all link costs in the southern direction to 1 and all northern
directions to infinity. We term set of those prefixes |R, and directions to infinity. We term set of those prefixes |R, and
for each prefix, r, in |R, we define its set of next-hops to for each prefix, r, in |R, we define its set of next-hops to
be |H(r). Observe that policy-guided prefixes are NOT affected be |H(r).
since their distribution scope is controlled by configuration.
2. The node uses reflected S-TIEs to find all nodes at the same 2. The node uses reflected S-TIEs to find all nodes at the same
level in the same PoD and the set of southbound adjacencies for level in the same PoD and the set of southbound adjacencies for
each. The set of nodes at the same level is termed |N and for each. The set of nodes at the same level is termed |N and for
each node, n, in |N, we define its set of southbound adjacencies each node, n, in |N, we define its set of southbound adjacencies
to be |A(n). to be |A(n).
3. For a given r, if the intersection of |H(r) and |A(n), for any n, 3. For a given r, if the intersection of |H(r) and |A(n), for any n,
is null then that prefix r must be explicitly advertised by the is null then that prefix r must be explicitly advertised by the
node in an S-TIE. node in an S-TIE.
skipping to change at page 35, line 28 skipping to change at page 46, line 48
node's southbound adjacencies. In accordance with the normal node's southbound adjacencies. In accordance with the normal
flooding rules for an S-TIE, a node at the lower level that flooding rules for an S-TIE, a node at the lower level that
receives this S-TIE will not propagate it south-bound. Neither receives this S-TIE will not propagate it south-bound. Neither
is it necessary for the receiving node to reflect the is it necessary for the receiving node to reflect the
disaggregated prefixes back over its adjacencies to nodes at the disaggregated prefixes back over its adjacencies to nodes at the
level from which it was received. level from which it was received.
To summarize the above in simplest terms: if a node detects that its To summarize the above in simplest terms: if a node detects that its
default route encompasses prefixes for which one of the other nodes default route encompasses prefixes for which one of the other nodes
in its level has no possible next-hops in the level below, it has to in its level has no possible next-hops in the level below, it has to
disaggregate it to prevent black-holing or suboptimal routing. Hence disaggregate it to prevent black-holing or suboptimal routing through
a node X needs to determine if it can reach a different set of south such nodes. Hence a node X needs to determine if it can reach a
neighbors than other nodes at the same level, which are connected to different set of south neighbors than other nodes at the same level,
it via at least one common south or East-West neighbor. If it can, which are connected to it via at least one common south neighbor. If
then prefix disaggregation may be required. If it can't, then no it can, then prefix disaggregation may be required. If it can't,
prefix disaggregation is needed. An example of disaggregation is then no prefix disaggregation is needed. An example of
provided in Section 5.3. disaggregation is provided in Section 6.3.
A possible algorithm is described last: A possible algorithm is described last:
1. Create partial_neighbors = (empty), a set of neighbors with 1. Create partial_neighbors = (empty), a set of neighbors with
partial connectivity to the node X's level from X's perspective. partial connectivity to the node X's level from X's perspective.
Each entry is a list of south neighbor of X and a list of nodes Each entry is a list of south neighbor of X and a list of nodes
of X.level that can't reach that neighbor. of X.level that can't reach that neighbor.
2. A node X determines its set of southbound neighbors 2. A node X determines its set of southbound neighbors
X.south_neighbors. X.south_neighbors.
skipping to change at page 36, line 10 skipping to change at page 47, line 29
X.south_neighbors but the nodes share at least one southern X.south_neighbors but the nodes share at least one southern
neighbor, for each neighbor N in X.south_neighbors but not in neighbor, for each neighbor N in X.south_neighbors but not in
Y.south_neighbors, add (N, (Y)) to partial_neighbors if N isn't Y.south_neighbors, add (N, (Y)) to partial_neighbors if N isn't
there or add Y to the list for N. there or add Y to the list for N.
4. If partial_neighbors is empty, then node X does not to 4. If partial_neighbors is empty, then node X does not to
disaggregate any prefixes. If node X is advertising disaggregate any prefixes. If node X is advertising
disaggregated prefixes in its S-TIE, X SHOULD remove them and re- disaggregated prefixes in its S-TIE, X SHOULD remove them and re-
advertise its according S-TIEs. advertise its according S-TIEs.
A node X computes its SPF based upon the received N-TIEs. This A node X computes reachability to all nodes below it based upon the
results in a set of routes, each categorized by (prefix, received N-TIEs first. This results in a set of routes, each
path_distance, next-hop-set). Alternately, for clarity in the categorized by (prefix, path_distance, next-hop-set). Alternately,
following procedure, these can be organized by next-hop-set as ( for clarity in the following procedure, these can be organized by
(next-hops), {(prefix, path_distance)}). If partial_neighbors isn't next-hop-set as ( (next-hops), {(prefix, path_distance)}). If
empty, then the following procedure describes how to identify partial_neighbors isn't empty, then the following procedure describes
prefixes to disaggregate. how to identify prefixes to disaggregate.
disaggregated_prefixes = {empty } disaggregated_prefixes = { empty }
nodes_same_level = { empty } nodes_same_level = { empty }
for each S-TIE for each S-TIE
if (S-TIE.level == X.level and if (S-TIE.level == X.level and
X shares at least one S-neighbor with X) X shares at least one S-neighbor with X)
add S-TIE.originator to nodes_same_level add S-TIE.originator to nodes_same_level
end if end if
end for end for
for each next-hop-set NHS for each next-hop-set NHS
isolated_nodes = nodes_same_level isolated_nodes = nodes_same_level
skipping to change at page 36, line 48 skipping to change at page 48, line 35
add (prefix, distance) to disaggregated_prefixes add (prefix, distance) to disaggregated_prefixes
end for end for
end if end if
end for end for
copy disaggregated_prefixes to X's S-TIE copy disaggregated_prefixes to X's S-TIE
if X's S-TIE is different if X's S-TIE is different
schedule S-TIE for flooding schedule S-TIE for flooding
end if end if
Figure 6: Computation to Disaggregate Prefixes Figure 16: Computation of Disaggregated Prefixes
Each disaggregated prefix is sent with the accurate path_distance. Each disaggregated prefix is sent with the according path_distance.
This allows a node to send the same S-TIE to each south neighbor. This allows a node to send the same S-TIE to each south neighbor.
The south neighbor which is connected to that prefix will thus have a The south neighbor which is connected to that prefix will thus have a
shorter path. shorter path.
Finally, to summarize the less obvious points partially omitted in Finally, to summarize the less obvious points partially omitted in
the algorithms to keep them more tractable: the algorithms to keep them more tractable:
1. all neighbor relationships MUST perform backlink checks. 1. all neighbor relationships MUST perform backlink checks.
2. overload bits as introduced in Section 4.3.1 have to be respected 2. overload bits as introduced in Section 5.3.1 have to be respected
during the computation. during the computation.
3. all the lower level nodes are flooded the same disaggregated 3. all the lower level nodes are flooded the same disaggregated
prefixes since we don't want to build an S-TIE per node and prefixes since we don't want to build an S-TIE per node and
complicate things unnecessarily. The PoD containing the prefix complicate things unnecessarily. The PoD containing the prefix
will prefer southbound anyway. will prefer southbound anyway.
4. disaggregated prefixes do NOT have to propagate to lower levels. 4. positively disaggregated prefixes do NOT have to propagate to
With that the disturbance in terms of new flooding is contained lower levels. With that the disturbance in terms of new flooding
to a single level experiencing failures only. is contained to a single level experiencing failures.
5. disaggregated prefix S-TIEs are not "reflected" by the lower 5. disaggregated prefix S-TIEs are not "reflected" by the lower
level, i.e. nodes within same level do NOT need to be aware level, i.e. nodes within same level do NOT need to be aware
which node computed the need for disaggregation. which node computed the need for disaggregation.
6. The fabric is still supporting maximum load balancing properties 6. The fabric is still supporting maximum load balancing properties
while not trying to send traffic northbound unless necessary. while not trying to send traffic northbound unless necessary.
Ultimately, complex partitions of superspine on sparsely connected To close this section it is worth to observe that in a single plane
fabrics can lead to necessity of transitive disaggregation through ToF this disaggregation prevents blackholing up to (K_LEAF * P) link
multiple levels. The topic will be described and standardized in failures in terms of Section 5.1.2 or in other terms, it takes at
later versions of this document. minimum that many link failures to partition the ToF into multiple
planes.
4.2.9. Optional Autoconfiguration 5.2.5.2. Negative, Transitive Disaggregation for Fallen Leafs
As explained in Section 5.1.3 failures in multi-plane Top-of-Fabric
or more than (K_LEAF * P) links failing in single plane design can
generate fallen leafs. Such scenario cannot be addressed by positive
disaggregation only and needs a further mechanism.
5.2.5.2.1. Cabling of Multiple Top-of-Fabric Planes
Let us return in this section to designs with multiple planes as
shown in Figure 3. Figure 17 highlights how the ToF is cabled in
case of two planes by the means of dual-rings to distribute all the
N-TIEs within both planes. For people familiar with traditional
link-state routing protocols ToF level can be considered equivalent
to area 0 in OSPF or level-2 in ISIS which need to be "connected" as
well for the protocol to operate correctly.
. ++==========++ ++==========++
. II II II II
.+----++--+ +----++--+ +----++--+ +----++--+
.|ToF A1| |ToF B1| |ToF B2| |ToF A2|
.++-+-++--+ ++-+-++--+ ++-+-++--+ ++-+-++--+
. | | II | | II | | II | | II
. | | ++==========++ | | ++==========++
. | | | | | | | |
.
. ~~~ Highlighted ToF of the previous multi-plane figure ~~
Figure 17: Topologically connected planes
As described in Section 5.1.3 failures in multi-plane fabrics can
lead to blackholes which normal positive disaggregation cannot fix.
The mechanism of negative, transitive disaggregation incorporated in
RIFT provides the according solution.
5.2.5.2.2. Transitive Advertisement of Negative Disaggregates
A ToF node that discovers that it cannot reach a fallen leaf
disaggregates all the prefixes of such leafs. It uses for that
purpose negative prefix S-TIEs that are, as usual, flooded southwards
with the scope defined in Section 5.2.3.4.
Transitively, a node explicitly loses connectivity to a prefix when
none of its children advertises it and when the prefix is negatively
disaggregated by all of its parents. When that happens, the node
originates the negative prefix further down south. Since the
mechanism applies recursively south the negative prefix may propagate
transitively all the way down to the leaf. This is necessary since
leafs connected to multiple planes by means of disjoint paths may
have to choose the correct plane already at the very bottom of the
fabric to make sure that they don't send traffic towards another leaf
using a plane where it is "fallen" at which in point a blackhole is
unavoidable.
When the connectivity is restored, a node that disaggregated a prefix
withdraws the negative disaggregation by the usual mechanism of re-
advertising TIEs omitting the negative prefix.
5.2.5.2.3. Computation of Negative Disaggregates
The document omitted so far the description of the computation
necessary to generate the correct set of negative prefixes. Negative
prefixes can in fact be advertised due to two different triggers. We
describe them consecutively.
The first origination reason is a computation that uses all the node
N-TIEs to build the set of all reachable nodes by reachability
computation over the complete graph. The computation uses the node
itself as root. This is compared with the result of the normal
southbound SPF as described in Section 5.2.4.2. The difference are
the fallen leafs and all their attached prefixes are advertised as
negative prefixes southbound if the node does not see the prefix
being reachable within southbound SPF.
The second mechanism hinges on the understanding how the negative
prefixes are used within the computation as described in Figure 18.
When attaching the negative prefixes at certain point in time the
negative prefix may find itself with all the viable nodes from the
shorter match nexthop being pruned. In other words, all its
northbound neighbors provided a negative prefix advertisement. This
is the trigger to advertise this negative prefix transitively south
and normally caused by the node being in a plane where the prefix
belongs to a fabric leaf that has "fallen" in this plane. Obviously,
when one of the northbound switches withdraws its negative
advertisement, the node has to withdraw its transitively provided
negative prefix as well.
5.2.6. Attaching Prefixes
After the SPF is run, it is necessary to attach according prefixes.
For S-SPF, prefixes from an N-TIE are attached to the originating
node with that node's next-hop set and a distance equal to the
prefix's cost plus the node's minimized path distance. The RIFT
route database, a set of (prefix, type=spf, path_distance, next-hop
set), accumulates these results. Obviously, the prefix retains its
type which is used to tie-break between the same prefix advertised
with different types.
In case of N-SPF prefixes from each S-TIE need to also be added to
the RIFT route database. The N-SPF is really just a stub so the
computing node needs simply to determine, for each prefix in an S-TIE
that originated from adjacent node, what next-hops to use to reach
that node. Since there may be parallel links, the next-hops to use
can be a set; presence of the computing node in the associated Node
S-TIE is sufficient to verify that at least one link has
bidirectional connectivity. The set of minimum cost next-hops from
the computing node X to the originating adjacent node is determined.
Each prefix has its cost adjusted before being added into the RIFT
route database. The cost of the prefix is set to the cost received
plus the cost of the minimum distance next-hop to that neighbor while
taking into account its attributes such as mobility per Section 5.3.3
necessary. Then each prefix can be added into the RIFT route
database with the next_hop_set; ties are broken based upon type first
and then distance and further attributes. RIFT route preferences are
normalized by the according thrift model type.
An exemplary implementation for node X follows:
for each S-TIE
if S-TIE.level > X.level
next_hop_set = set of minimum cost links to the S-TIE.originator
next_hop_cost = minimum cost link to S-TIE.originator
end if
for each prefix P in the S-TIE
P.cost = P.cost + next_hop_cost
if P not in route_database:
add (P, type=DistVector, P.cost, next_hop_set) to route_database
end if
if (P in route_database):
if route_database[P].cost > P.cost):
update route_database[P] with (P, DistVector, P.cost, next_hop_set)
else if route_database[P].cost == P.cost
update route_database[P] with (P, DistVector, P.cost,
merge(next_hop_set, route_database[P].next_hop_set))
else
// Not preferred route so ignore
end if
end if
end for
end for
Figure 18: Adding Routes from S-TIE Positive and Negative Prefixes
After the positive prefixes are attached and tie-broken, negative
prefixes are attached and used in case of northbound computation,
ideally from the shortest length to the longest. The nexthop
adjacencies for a negative prefix are inherited from the longest
prefix that aggregates it, and subsequently adjacencies to nodes that
advertised negative for this prefix are removed. As an example, if a
hypothetical RIFT routing table contains A.1/16 @ [A,B], A.1.1/24 @
[C,D] it will on reception of negative A.1.1.1/32 from D include the
entry A.1.1.1/32 @ [C] resulting from computation inheriting A.1.1/24
nexthops (C and D) and pruning all the nodes that advertised this
negative prefix (which is D in this case).
The rule of inheritance MUST be maintained when the nexthop list for
a prefix is modified, as the modification may affect the entries for
matching negative prefixes of immediate longer prefix length. For
instance, if a nexthop is added, then by inheritance it must be added
to all the negative routes of immediate longer prefixes length unless
it is pruned due to a negative advertisement for the same next hop.
Similarily, if a nexthop is deleted for a given prefix, then it is
deleted for all the immediately aggregated negative routes. This
will recurse in the case of nested negative prefix aggregations.
The rule of inheritance must also be maintained when a new prefix of
intermediate length is inserted, or when the immediately aggregating
prefix is deleted from the routing table, making an even shorter
aggregating prefix the one from which the negative routes now inherit
their adjacencies. As the aggregating prefix changes, all the
negative routes must be recomputed, and then again the process may
recurse in case of nested negative prefix aggregations.
Observe that despite seeming quite computationally expensive the
operations are only necessary if the only available advertisements
for a prefix are negative since tie-breaking always prefers positive
information for the prefix which stops any kind of recursion since
positive information never inherits next hops.
5.2.7. Optional Zero Touch Provisioning (ZTP)
Each RIFT node can optionally operate in zero touch provisioning Each RIFT node can optionally operate in zero touch provisioning
(ZTP) mode, i.e. it has no configuration (unless it is a superspine (ZTP) mode, i.e. it has no configuration (unless it is a Top-of-
at the top of the topology or the must operate in the topology as Fabric at the top of the topology or the must operate in the topology
leaf and/or support leaf-2-leaf procedures) and it will fully as leaf and/or support leaf-2-leaf procedures) and it will fully
configure itself after being attached to the topology. Configured configure itself after being attached to the topology. Configured
nodes and nodes operating in ZTP can be mixed and will form a valid nodes and nodes operating in ZTP can be mixed and will form a valid
topology if achievable. This section describes the necessary topology if achievable.
concepts and procedures.
4.2.9.1. Terminology The derivation of the level of each node happens based on offers
received from its neighbors whereas each node (with possibly
exceptions of configured leafs) tries to attach at the highest
possible point in the fabric. This guarantees that even if the
diffusion front reaches a node from "below" faster than from "above",
it will greedily abandon already negotiated level derived from nodes
topologically below it and properly peers with nodes above.
The fabric is very conciously numbered from the top to allow for PoDs
of different heights and minimize number of provisioning necessary,
in this case just a TOP_OF_FABRIC flag.
This section describes the necessary concepts and procedures for ZTP
operation.
5.2.7.1. Terminology
The interdependencies between the different flags and the configured
level can be somewhat vexing at first and it may take multiple reads
of the glossary to make sense.
Automatic Level Derivation: Procedures which allow nodes without Automatic Level Derivation: Procedures which allow nodes without
level configured to derive it automatically. Only applied if level configured to derive it automatically. Only applied if
CONFIGURED_LEVEL is undefined. CONFIGURED_LEVEL is undefined.
UNDEFINED_LEVEL: An imaginary value that indicates that the level UNDEFINED_LEVEL: An imaginary value that indicates that the level
has not beeen determined and has not been configured. Schemas has not beeen determined and has not been configured. Schemas
normally indicate that by a missing optional value without an normally indicate that by a missing optional value without an
available defined default. available defined default.
LEAF_ONLY: An optional configuration flag that can be configured on LEAF_ONLY: An optional configuration flag that can be configured on
a node to make sure it never leaves the "bottom of the hierarchy". a node to make sure it never leaves the "bottom of the hierarchy".
SUPERSPINE_FLAG and CONFIGURED_LEVEL cannot be defined at the same TOP_OF_FABRIC flag and CONFIGURED_LEVEL cannot be defined at the
time as this flag. It implies CONFIGURED_LEVEL value of 0. same time as this flag. It implies CONFIGURED_LEVEL value of 0.
TOP_OF_FABRIC flag: Configuration flag provided to all Top-of-Fabric
nodes. LEAF_FLAG and CONFIGURED_LEVEL cannot be defined at the
same time as this flag. It implies a CONFIGURED_LEVEL value. In
fact, it is basically a shortcut for configuring same level at all
Top-of-Fabric nodes which is unavoidable since an initial 'seed'
is needed for other ZTP nodes to derive their level in the
topology. The flag plays an important role in fabrics with
multiple planes to enable successful negative disaggregation
(Section 5.2.5.2).
CONFIGURED_LEVEL: A level value provided manually. When this is CONFIGURED_LEVEL: A level value provided manually. When this is
defined (i.e. it is not an UNDEFINED_LEVEL) the node is not defined (i.e. it is not an UNDEFINED_LEVEL) the node is not
participating in ZTP. SUPERSPINE_FLAG is ignored when this value participating in ZTP. TOP_OF_FABRIC flag is ignored when this
is defined. LEAF_ONLY can be set only if this value is undefined value is defined. LEAF_ONLY can be set only if this value is
or set to 0. undefined or set to 0.
DERIVED_LEVEL: Level value computed via automatic level derivation DERIVED_LEVEL: Level value computed via automatic level derivation
when CONFIGURED_LEVEL is equal to UNDEFINED_LEVEL. when CONFIGURED_LEVEL is equal to UNDEFINED_LEVEL.
LEAF_2_LEAF: An optional flag that can be configured on a node to LEAF_2_LEAF: An optional flag that can be configured on a node to
make sure it supports procedures defined in Section 4.3.9. make sure it supports procedures defined in Section 5.3.9. In a
SUPERSPINE_FLAG is ignored when set at the same time as this flag. strict sense it is a capability that implies LEAF_ONLY and the
LEAF_2_LEAF implies LEAF_ONLY and the according restrictions. according restrictions. TOP_OF_FABRIC flag is ignored when set at
the same time as this flag.
LEVEL_VALUE: In ZTP case the original definition of "level" in LEVEL_VALUE: In ZTP case the original definition of "level" in
Section 2.1 is both extended and relaxed. First, level is defined Section 3.1 is both extended and relaxed. First, level is defined
now as LEVEL_VALUE and is the first defined value of now as LEVEL_VALUE and is the first defined value of
CONFIGURED_LEVEL followed by DERIVED_LEVEL. Second, it is CONFIGURED_LEVEL followed by DERIVED_LEVEL. Second, it is
possible for nodes to be more than one level apart to form possible for nodes to be more than one level apart to form
adjacencies if any of the nodes is at least LEAF_ONLY. adjacencies if any of the nodes is at least LEAF_ONLY.
Valid Offered Level (VOL): A neighbor's level received on a valid Valid Offered Level (VOL): A neighbor's level received on a valid
LIE (i.e. passing all checks for adjacency formation while LIE (i.e. passing all checks for adjacency formation while
disregarding all clauses involving level values) persisting for disregarding all clauses involving level values) persisting for
the duration of the holdtime interval on the LIE. Observe that the duration of the holdtime interval on the LIE. Observe that
offers from nodes offering level value of 0 do not constitute VOLs offers from nodes offering level value of 0 do not constitute VOLs
(since no valid DERIVED_LEVEL can be obtained from those). Offers (since no valid DERIVED_LEVEL can be obtained from those and
from LIEs with `not_a_ztp_offer` being true are not VOLs either. consequently `not_a_ztp_offer` MUST be ignored). Offers from LIEs
with `not_a_ztp_offer` being true are not VOLs either. If a node
maintains parallel adjacencies to the neighbor, VOL on each
adjacency is considered as equivalent, i.e. the newest VOL from
any such adjacency updates the VOL received from the same node.
Highest Available Level (HAL): Highest defined level value seen from Highest Available Level (HAL): Highest defined level value seen from
all VOLs received. all VOLs received.
Highest Available Level Systems (HALS): Set of nodes offering HAL
VOLs.
Highest Adjacency Three Way (HAT): Highest neigbhor level of all the Highest Adjacency Three Way (HAT): Highest neigbhor level of all the
formed three way adjacencies for the node. formed three way adjacencies for the node.
SUPERSPINE_FLAG: Configuration flag provided to all superspines. 5.2.7.2. Automatic SystemID Selection
LEAF_FLAG and CONFIGURED_LEVEL cannot be defined at the same time
as this flag. It implies a CONFIGURED_LEVEL value. In fact, it
is basically a shortcut for configuring same level at all
superspine nodes which is unavoidable since an initial 'seed' is
needed for other ZTP nodes to derive their level in the topology.
4.2.9.2. Automatic SystemID Selection
RIFT identifies each node via a SystemID which is a 64 bits wide RIFT identifies each node via a SystemID which is a 64 bits wide
integer. It is relatively simple to derive a, for all practical integer. It is relatively simple to derive a, for all practical
purposes collision free, value for each node on startup. For that purposes collision free, value for each node on startup. For that
purpose, a node MUST use as system ID EUI-64 MA-L format where the purpose, a node MUST use as system ID EUI-64 MA-L format [EUI64]
organizationally governed 24 bits can be used to generate system IDs where the organizationally governed 24 bits can be used to generate
for multiple RIFT instances running on the system. system IDs for multiple RIFT instances running on the system.
The router MUST ensure that such identifier is not changing very As matter of operational concern, the router MUST ensure that such
frequently (at least not without sending all its TIEs with fairly identifier is not changing very frequently (or at least not without
short lifetimes) since otherwise the network may be left with large sending all its TIEs with fairly short lifetimes) since otherwise the
amounts of stale TIEs in other nodes (though this is not necessarily network may be left with large amounts of stale TIEs in other nodes
a serious problem if the procedures suggested in Section 7 are (though this is not necessarily a serious problem if the procedures
implemented). described in Section 8 are implemented).
4.2.9.3. Generic Fabric Example 5.2.7.3. Generic Fabric Example
ZTP forces us to think about miscabled or unusually cabled fabric and ZTP forces us to think about miscabled or unusually cabled fabric and
how such a topology can be forced into a "lattice" structure which a how such a topology can be forced into a "lattice" structure which a
fabric represents (with further restrictions). Let us consider a fabric represents (with further restrictions). Let us consider a
necessary and sufficient physical cabling in Figure 7. We assume all necessary and sufficient physical cabling in Figure 19. We assume
nodes being in the same PoD. all nodes being in the same PoD.
. +---+ . +---+
. | A | s = SUPERSPINE_FLAG . | A | s = TOP_OF_FABRIC
. | s | l = LEAF_ONLY . | s | l = LEAF_ONLY
. ++-++ l2l = LEAF_2_LEAF . ++-++ l2l = LEAF_2_LEAF
. | | . | |
. +--+ +--+ . +--+ +--+
. | | . | |
. +--++ ++--+ . +--++ ++--+
. | E | | F | . | E | | F |
. | +-+ | +-----------+ . | +-+ | +-----------+
. ++--+ | ++-++ | . ++--+ | ++-++ |
. | | | | | . | | | | |
. | +-------+ | | . | +-------+ | |
. | | | | | . | | | | |
. | | +----+ | | . | | +----+ | |
. | | | | | . | | | | |
. ++-++ ++-++ | . ++-++ ++-++ |
. | I +-----+ J | | . | I +-----+ J | |
. | | | +-+ | . | | | +-+ |
. ++-++ +--++ | | . ++-++ +--++ | |
. | | | | | . | | | | |
. +---------+ | +------+ | . +---------+ | +------+ |
. | | | | | . | | | | |
. +-----------------+ | | . +-----------------+ | |
. | | | | | . | | | | |
. ++-++ ++-++ | . ++-++ ++-++ |
. | X +-----+ Y +-+ . | X +-----+ Y +-+
. |l2l| | l | . |l2l| | l |
. +---+ +---+ . +---+ +---+
Figure 7: Generic ZTP Cabling Considerations Figure 19: Generic ZTP Cabling Considerations
First, we must anchor the "top" of the cabling and that's what the First, we must anchor the "top" of the cabling and that's what the
SUPERSPINE_FLAG at node A is for. Then things look smooth until we TOP_OF_FABRIC flag at node A is for. Then things look smooth until
have to decide whether node Y is at the same level as I, J or at the we have to decide whether node Y is at the same level as I, J or at
same level as Y and consequently, X is south of it. This is the same level as Y and consequently, X is south of it. This is
unresolvable here until we "nail down the bottom" of the topology. unresolvable here until we "nail down the bottom" of the topology.
To achieve that we choose to use in this example the leaf flags. We To achieve that we choose to use in this example the leaf flags. We
will see further then whether Y chooses to form adjacencies to F or will see further then whether Y chooses to form adjacencies to F or
I, J successively. I, J successively.
4.2.9.4. Level Determination Procedure 5.2.7.4. Level Determination Procedure
A node starting up with UNDEFINED_VALUE (i.e. without a A node starting up with UNDEFINED_VALUE (i.e. without a
CONFIGURED_LEVEL or any leaf or superspine flag) MUST follow those CONFIGURED_LEVEL or any leaf or TOP_OF_FABRIC flag) MUST follow those
additional procedures: additional procedures:
1. It advertises its LEVEL_VALUE on all LIEs (observe that this can 1. It advertises its LEVEL_VALUE on all LIEs (observe that this can
be UNDEFINED_LEVEL which in terms of the schema is simply an be UNDEFINED_LEVEL which in terms of the schema is simply an
omitted optional value). omitted optional value).
2. It chooses on an ongoing basis from all VOLs the value of 2. It computes HAL as numerically highest available level in all
MAX(HAL-1,0) as its DERIVED_LEVEL. The node then starts to VOLs.
advertise this derived level.
3. A node that lost all adjacencies with HAL value MUST hold down 3. It chooses then MAX(HAL-1,0) as its DERIVED_LEVEL. The node then
starts to advertise this derived level.
4. A node that lost all adjacencies with HAL value MUST hold down
computation of new DERIVED_LEVEL for a short period of time computation of new DERIVED_LEVEL for a short period of time
unless it has no VOLs from southbound adjacencies. After the unless it has no VOLs from southbound adjacencies. After the
holddown expired, it MUST discard all received offers, recompute holddown expired, it MUST discard all received offers, recompute
DERIVED_LEVEL and announce it to all neighbors. DERIVED_LEVEL and announce it to all neighbors.
4. A node MUST reset any adjacency that has changed the level it is 5. A node MUST reset any adjacency that has changed the level it is
offering and is in three way state. offering and is in three way state.
5. A node that changed its defined level value MUST readvertise its 6. A node that changed its defined level value MUST readvertise its
own TIEs (since the new `PacketHeader` will contain a different own TIEs (since the new `PacketHeader` will contain a different
level than before). Sequence number of each TIE MUST be level than before). Sequence number of each TIE MUST be
increased. increased.
6. After a level has been derived the node MUST set the 7. After a level has been derived the node MUST set the
`not_a_ztp_offer` on LIEs towards all systems extending a VOL for `not_a_ztp_offer` on LIEs towards all systems extending a VOL for
HAL. HAL.
A node starting with LEVEL_VALUE being 0 (i.e. it assumes a leaf A node starting with LEVEL_VALUE being 0 (i.e. it assumes a leaf
function by being configured with the appropriate flags or has a function by being configured with the appropriate flags or has a
CONFIGURED_LEVEL of 0) MUST follow those additional procedures: CONFIGURED_LEVEL of 0) MUST follow those additional procedures:
1. It computes HAT per procedures above but does NOT use it to 1. It computes HAT per procedures above but does NOT use it to
compute DERIVED_LEVEL. HAT is used to limit adjacency formation compute DERIVED_LEVEL. HAT is used to limit adjacency formation
per Section 4.2.2. per Section 5.2.2.
Precise finite state machines will be provided in later versions of It MAY also follow modified procedures:
this specification.
4.2.9.5. Resulting Topologies 1. It may pick a different strategy to choose VOL, e.g. use the VOL
value with highest number of VOLs. Such strategies are only
possible since the node always remains "at the bottom of the
fabric" while another layer could "invert" the fabric by picking
its prefered VOL in a different fashion than always trying to
achieve the highest viable level.
The procedures defined in Section 4.2.9.4 will lead to the RIFT 5.2.7.5. Resulting Topologies
topology and levels depicted in Figure 8.
The procedures defined in Section 5.2.7.4 will lead to the RIFT
topology and levels depicted in Figure 20.
. +---+ . +---+
. | As| . | As|
. | 24| . | 24|
. ++-++ . ++-++
. | | . | |
. +--+ +--+ . +--+ +--+
. | | . | |
. +--++ ++--+ . +--++ ++--+
. | E | | F | . | E | | F |
skipping to change at page 42, line 33 skipping to change at page 58, line 38
. | 22| | 22| | . | 22| | 22| |
. ++--+ +--++ | . ++--+ +--++ |
. | | | . | | |
. +---------+ | | . +---------+ | |
. | | | . | | |
. ++-++ +---+ | . ++-++ +---+ |
. | X | | Y +-+ . | X | | Y +-+
. | 0 | | 0 | . | 0 | | 0 |
. +---+ +---+ . +---+ +---+
Figure 8: Generic ZTP Topology Autoconfigured Figure 20: Generic ZTP Topology Autoconfigured
In case we imagine the LEAF_ONLY restriction on Y is removed the In case we imagine the LEAF_ONLY restriction on Y is removed the
outcome would be very different however and result in Figure 9. This outcome would be very different however and result in Figure 21.
demonstrates basically that auto configuration prevents miscabling This demonstrates basically that auto configuration makes miscabling
detection and with that can lead to undesirable effects in cases detection hard and with that can lead to undesirable effects in cases
where leafs are not "nailed" by the accordingly configured flags and where leafs are not "nailed" by the accordingly configured flags and
arbitrarily cabled. arbitrarily cabled.
A node MAY analyze the outstanding level offers on its interfaces and
generate warnings when its internal ruleset flags a possible
miscabling. As an example, when a node's sees ZTP level offers that
differ by more than one level from its chosen level (with proper
accounting for leaf's being at level 0) this can indicate miscabling.
. +---+ . +---+
. | As| . | As|
. | 24| . | 24|
. ++-++ . ++-++
. | | . | |
. +--+ +--+ . +--+ +--+
. | | . | |
. +--++ ++--+ . +--++ ++--+
. | E | | F | . | E | | F |
. | 23+-+ | 23+-------+ . | 23+-+ | 23+-------+
skipping to change at page 43, line 35 skipping to change at page 59, line 35
. | | | | | . | | | | |
. | +-----------------+ | . | +-----------------+ |
. | | | . | | |
. +---------+ | | . +---------+ | |
. | | | . | | |
. ++-++ | . ++-++ |
. | X +--------+ . | X +--------+
. | 0 | . | 0 |
. +---+ . +---+
Figure 9: Generic ZTP Topology Autoconfigured Figure 21: Generic ZTP Topology Autoconfigured
4.2.10. Stability Considerations 5.2.8. Stability Considerations
The autoconfiguration mechanism computes a global maximum of levels The autoconfiguration mechanism computes a global maximum of levels
by diffusion. The achieved equilibrium can be disturbed massively by by diffusion. The achieved equilibrium can be disturbed massively by
all nodes with highest level either leaving or entering the domain all nodes with highest level either leaving or entering the domain
(with some finer distinctions not explained further). It is (with some finer distinctions not explained further). It is
therefore recommended that each node is multi-homed towards nodes therefore recommended that each node is multi-homed towards nodes
with respective HAL offerings. Fortuntately, this is the natural with respective HAL offerings. Fortuntately, this is the natural
state of things for the topology variants considered in RIFT. state of things for the topology variants considered in RIFT.
4.3. Further Mechanisms 5.3. Further Mechanisms
4.3.1. Overload Bit 5.3.1. Overload Bit
Overload Bit MUST be respected in all according reachability Overload Bit MUST be respected in all according reachability
computations. A node with overload bit set SHOULD NOT advertise any computations. A node with overload bit set SHOULD NOT advertise any
reachability prefixes southbound except locally hosted ones. reachability prefixes southbound except locally hosted ones. A node
in overload SHOULD advertise all its locally hosted prefixes north
and southbound.
The leaf node SHOULD set the 'overload' bit on its node TIEs, since The leaf node SHOULD set the 'overload' bit on its node TIEs, since
if the spine nodes were to forward traffic not meant for the local if the spine nodes were to forward traffic not meant for the local
node, the leaf node does not have the topology information to prevent node, the leaf node does not have the topology information to prevent
a routing/forwarding loop. a routing/forwarding loop.
4.3.2. Optimized Route Computation on Leafs 5.3.2. Optimized Route Computation on Leafs
Since the leafs do see only "one hop away" they do not need to run a Since the leafs do see only "one hop away" they do not need to run a
full SPF but can simply gather prefix candidates from their neighbors full SPF but can simply gather prefix candidates from their neighbors
and build the according routing table. and build the according routing table.
A leaf will have no N-TIEs except its own and optionally from its A leaf will have no N-TIEs except its own and optionally from its
East-West neighbors. A leaf will have S-TIEs from its neighbors. East-West neighbors. A leaf will have S-TIEs from its neighbors.
Instead of creating a network graph from its N-TIEs and neighbor's Instead of creating a network graph from its N-TIEs and neighbor's
S-TIEs and then running an SPF, a leaf node can simply compute the S-TIEs and then running an SPF, a leaf node can simply compute the
minimum cost and next_hop_set to each leaf neighbor by examining its minimum cost and next_hop_set to each leaf neighbor by examining its
local interfaces, determining bi-directionality from the associated local adjacencies, determining bi-directionality from the associated
N-TIE, and specifying the neighbor's next_hop_set set and cost from N-TIE, and specifying the neighbor's next_hop_set set and cost from
the minimum cost local interfaces to that neighbor. the minimum cost local adjacency to that neighbor.
Then a leaf attaches prefixes as in Section 4.2.6 as well as the Then a leaf attaches prefixes as in Section 5.2.6.
policy-guided prefixes as in Section 4.2.7.
4.3.3. Mobility 5.3.3. Mobility
It is a requirement for RIFT to maintain at the control plane a real It is a requirement for RIFT to maintain at the control plane a real
time status of which prefix is attached to which port of which leaf, time status of which prefix is attached to which port of which leaf,
even in a context of mobility where the point of attachement may even in a context of mobility where the point of attachement may
change several times in a subsecond period of time. change several times in a subsecond period of time.
There are two classical approaches to maintain such knowledge in an There are two classical approaches to maintain such knowledge in an
unambiguous fashion: unambiguous fashion:
time stamp: With this method, the infrastructure memorizes the time stamp: With this method, the infrastructure memorizes the
skipping to change at page 45, line 35 skipping to change at page 61, line 36
knowledge of the source of the sequence counter is required to knowledge of the source of the sequence counter is required to
operate it, and the comparison between sequence counters from operate it, and the comparison between sequence counters from
heterogeneous sources can be hard to impossible. heterogeneous sources can be hard to impossible.
RIFT supports a hybrid approach contained in an optional RIFT supports a hybrid approach contained in an optional
`PrefixSequenceType` prefix attribute that we call a `monotonic `PrefixSequenceType` prefix attribute that we call a `monotonic
clock` consisting of a timestamp and optional sequence number. In clock` consisting of a timestamp and optional sequence number. In
case of presence of the attribute: case of presence of the attribute:
o The leaf node MUST advertise a time stamp of the latest sighting o The leaf node MUST advertise a time stamp of the latest sighting
of a prefix, e.g., by snooping IP protocols or the switch using of a prefix, e.g., by snooping IP protocols or the node using the
the time at which it advertised the prefix. RIFT transports the time at which it advertised the prefix. RIFT transports the time
time stamp within the desired prefix N-TIEs as 802.1AS timestamp. stamp within the desired prefix N-TIEs as 802.1AS timestamp.
o RIFT may interoperate with the "update to 6LoWPAN Neighbor o RIFT may interoperate with the "update to 6LoWPAN Neighbor
Discovery" [I-D.ietf-6lo-rfc6775-update], which provides a method Discovery" [I-D.ietf-6lo-rfc6775-update], which provides a method
for registering a prefix with a sequence counter called a for registering a prefix with a sequence counter called a
Transaction ID (TID). RIFT transports in such case the TID in its Transaction ID (TID). RIFT transports in such case the TID in its
native form. native form.
o RIFT also defines an abstract negative clock (ANSC) that compares o RIFT also defines an abstract negative clock (ANSC) that compares
as less than any other clock. By default, the lack of a as less than any other clock. By default, the lack of a
`PrefixSequenceType` in a Prefix N-TIE is interpreted as ANSC. We `PrefixSequenceType` in a Prefix N-TIE is interpreted as ANSC. We
skipping to change at page 46, line 15 skipping to change at page 62, line 15
o Any prefix present on the fabric in multiple nodes that has the o Any prefix present on the fabric in multiple nodes that has the
`same` clock is considered as anycast. ASNC is always considered `same` clock is considered as anycast. ASNC is always considered
smaller than any defined clock. smaller than any defined clock.
o RIFT implementation assumes by default that all nodes are being o RIFT implementation assumes by default that all nodes are being
synchronized to 200 milliseconds precision which is easily synchronized to 200 milliseconds precision which is easily
achievable even in very large fabrics using [RFC5905]. An achievable even in very large fabrics using [RFC5905]. An
implementation MAY provide a way to reconfigure a domain to a implementation MAY provide a way to reconfigure a domain to a
different value. We call this variable MAXIMUM_CLOCK_DELTA. different value. We call this variable MAXIMUM_CLOCK_DELTA.
4.3.3.1. Clock Comparison 5.3.3.1. Clock Comparison
All monotonic clock values are comparable to each other using the All monotonic clock values are comparable to each other using the
following rules: following rules:
1. ASNC is older than any other value except ASNC AND 1. ASNC is older than any other value except ASNC AND
2. Clock with timestamp differing by more than MAXIMUM_CLOCK_DELTA 2. Clock with timestamp differing by more than MAXIMUM_CLOCK_DELTA
are comparable by using the timestamps only AND are comparable by using the timestamps only AND
3. Clocks with timestamps differing by less than MAXIMUM_CLOCK_DELTA 3. Clocks with timestamps differing by less than MAXIMUM_CLOCK_DELTA
are comparable by using their TIDs only AND are comparable by using their TIDs only AND
4. An undefined TID is always older than any other TID AND 4. An undefined TID is always older than any other TID AND
5. TIDs are compared using rules of [I-D.ietf-6lo-rfc6775-update]. 5. TIDs are compared using rules of [I-D.ietf-6lo-rfc6775-update].
4.3.3.2. Interaction between Time Stamps and Sequence Counters 5.3.3.2. Interaction between Time Stamps and Sequence Counters
For slow movements that occur less frequently than e.g. once per For slow movements that occur less frequently than e.g. once per
second, the time stamp that the RIFT infrastruture captures is enough second, the time stamp that the RIFT infrastruture captures is enough
to determine the freshest discovery. If the point of attachement to determine the freshest discovery. If the point of attachement
changes faster than the maximum drift of the time stamping mechanism changes faster than the maximum drift of the time stamping mechanism
(i.e. MAXIMUM_CLOCK_DELTA), then a sequence counter is required to (i.e. MAXIMUM_CLOCK_DELTA), then a sequence counter is required to
add resolution to the freshness evaluation, and it must be sized so add resolution to the freshness evaluation, and it must be sized so
that the counters stay comparable within the resolution of the time that the counters stay comparable within the resolution of the time
stampling mechanism. stampling mechanism.
skipping to change at page 47, line 9 skipping to change at page 63, line 9
still preserve the capability to discover an error situation where still preserve the capability to discover an error situation where
counters are not comparable. counters are not comparable.
Within the resolution of MAXIMUM_CLOCK_DELTA the sequence counters Within the resolution of MAXIMUM_CLOCK_DELTA the sequence counters
captured during 2 sequential values of the time stamp must be captured during 2 sequential values of the time stamp must be
comparable. This means with default values that a node may move up comparable. This means with default values that a node may move up
to 16 times during a 200 milliseconds period and the clocks remain to 16 times during a 200 milliseconds period and the clocks remain
still comparable thus allowing the infrastructure to assert the still comparable thus allowing the infrastructure to assert the
freshest advertisement with no ambiguity. freshest advertisement with no ambiguity.
4.3.3.3. Anycast vs. Unicast 5.3.3.3. Anycast vs. Unicast
A unicast prefix can be attached to at most one leaf, whereas an A unicast prefix can be attached to at most one leaf, whereas an
anycast prefix may be reachable via more than one leaf. anycast prefix may be reachable via more than one leaf.
If a monotonic clock attribute is provided on the prefix, then the If a monotonic clock attribute is provided on the prefix, then the
prefix with the `newest` clock value is strictly prefered. An prefix with the `newest` clock value is strictly prefered. An
anycast prefix does not carry a clock or all clock attributes MUST be anycast prefix does not carry a clock or all clock attributes MUST be
the same under the rules of Section 4.3.3.1. the same under the rules of Section 5.3.3.1.
Observe that it is important that in mobility events the leaf is re- Observe that it is important that in mobility events the leaf is re-
flooding as quickly as possible the absence of the prefix that moved flooding as quickly as possible the absence of the prefix that moved
away. away.
Observe further that without support for Observe further that without support for
[I-D.ietf-6lo-rfc6775-update] movements on the fabric within [I-D.ietf-6lo-rfc6775-update] movements on the fabric within
intervals smaller than 100msec will be seen as anycast. intervals smaller than 100msec will be seen as anycast.
4.3.3.4. Overlays and Signaling 5.3.3.4. Overlays and Signaling
RIFT is agnostic whichever the overlay technology [MIP, LISP, VxLAN, RIFT is agnostic whichever the overlay technology [MIP, LISP, VxLAN,
NVO3] and the associated signaling is deployed over it. But it is NVO3] and the associated signaling is deployed over it. But it is
expected that leaf nodes, and possibly superspine nodes can perform expected that leaf nodes, and possibly Top-of-Fabric nodes can
the according encapsulation. perform the according encapsulation.
In the context of mobility, overlays provide a classical solution to In the context of mobility, overlays provide a classical solution to
avoid injecting mobile prefixes in the fabric and improve the avoid injecting mobile prefixes in the fabric and improve the
scalability of the solution. It makes sense on a data center that scalability of the solution. It makes sense on a data center that
already uses overlays to consider their applicability to the mobility already uses overlays to consider their applicability to the mobility
solution; as an example, a mobility protocol such as LISP may inform solution; as an example, a mobility protocol such as LISP may inform
the ingress leaf of the location of the egress leaf in real time. the ingress leaf of the location of the egress leaf in real time.
Another possibility is to consider that mobility as an underlay Another possibility is to consider that mobility as an underlay
service and support it in RIFT to an extent. The load on the fabric service and support it in RIFT to an extent. The load on the fabric
augments with the amount of mobility obviously since a move forces augments with the amount of mobility obviously since a move forces
flooding and computation on all nodes in the scope of the move so flooding and computation on all nodes in the scope of the move so
tunneling from leaf to the superspines may be desired. Future tunneling from leaf to the Top-of-Fabric may be desired. Future
versions of this document may describe support for such tunneling in versions of this document may describe support for such tunneling in
RIFT. RIFT.
4.3.4. Key/Value Store 5.3.4. Key/Value Store
4.3.4.1. Southbound 5.3.4.1. Southbound
The protocol supports a southbound distribution of key-value pairs The protocol supports a southbound distribution of key-value pairs
that can be used to e.g. distribute configuration information during that can be used to e.g. distribute configuration information during
topology bring-up. The KV S-TIEs can arrive from multiple nodes and topology bring-up. The KV S-TIEs can arrive from multiple nodes and
hence need tie-breaking per key. We use the following rules hence need tie-breaking per key. We use the following rules
1. Only KV TIEs originated by a node to which the receiver has an 1. Only KV TIEs originated by a node to which the receiver has an
adjacency are considered. adjacency are considered.
2. Within all valid KV S-TIEs containing the key, the value of the 2. Within all valid KV S-TIEs containing the key, the value of the
KV S-TIE for which the according node S-TIE is present, has the KV S-TIE for which the according node S-TIE is present, has the
highest level and within the same level has highest originator ID highest level and within the same level has highest originating
is preferred. If keys in the most preferred TIEs are system ID is preferred. If keys in the most preferred TIEs are
overlapping, the behavior is undefined. overlapping, the behavior is undefined.
Observe that if a node goes down, the node south of it looses Observe that if a node goes down, the node south of it looses
adjacencies to it and with that the KVs will be disregarded and on adjacencies to it and with that the KVs will be disregarded and on
tie-break changes new KV re-advertised to prevent stale information tie-break changes new KV re-advertised to prevent stale information
being used by nodes further south. KV information in southbound being used by nodes further south. KV information in southbound
direction is not result of independent computation of every node but direction is not result of independent computation of every node but
a diffused computation. a diffused computation.
4.3.4.2. Northbound 5.3.4.2. Northbound
Certain use cases seem to necessitate distribution of essentialy KV Certain use cases seem to necessitate distribution of essentialy KV
information that is generated in the leafs in the northbound information that is generated in the leafs in the northbound
direction. Such information is flooded in KV N-TIEs. Since the direction. Such information is flooded in KV N-TIEs. Since the
originator of northbound KV is preserved during northbound flooding, originator of northbound KV is preserved during northbound flooding,
overlapping keys could be used. However, to omit further protocol overlapping keys could be used. However, to omit further protocol
complexity, only the value of the key in TIE tie-broken in same complexity, only the value of the key in TIE tie-broken in same
fashion as southbound KV TIEs is used. fashion as southbound KV TIEs is used.
4.3.5. Interactions with BFD 5.3.5. Interactions with BFD
RIFT MAY incorporate BFD [RFC5881] to react quickly to link failures. RIFT MAY incorporate BFD [RFC5881] to react quickly to link failures.
In such case following procedures are introduced: In such case following procedures are introduced:
After RIFT three way hello adjacency convergence a BFD session MAY After RIFT three way hello adjacency convergence a BFD session MAY
be formed automatically between the RIFT endpoints without further be formed automatically between the RIFT endpoints without further
configuration using the exchanged discriminators. configuration using the exchanged discriminators.
In case established BFD session goes Down after it was Up, RIFT In case established BFD session goes Down after it was Up, RIFT
adjacency should be re-initialized started from Init. adjacency should be re-initialized started from Init.
skipping to change at page 49, line 18 skipping to change at page 65, line 18
In case RIFT changes link identifiers both the hello as well as In case RIFT changes link identifiers both the hello as well as
the BFD sessions SHOULD be brought down and back up again. the BFD sessions SHOULD be brought down and back up again.
Multiple RIFT instances MAY choose to share a single BFD session Multiple RIFT instances MAY choose to share a single BFD session
(in such case it is undefined what discriminators are used albeit (in such case it is undefined what discriminators are used albeit
RIFT CAN advertise the same link ID for the same interface in RIFT CAN advertise the same link ID for the same interface in
multiple instances and with that "share" the discriminators). multiple instances and with that "share" the discriminators).
BFD TTL follows [RFC5082]. BFD TTL follows [RFC5082].
4.3.6. Fabric Bandwidth Balancing 5.3.6. Fabric Bandwidth Balancing
A well understood problem in fabrics is that in case of link losses A well understood problem in fabrics is that in case of link losses
it would be ideal to rebalance how much traffic is offered to it would be ideal to rebalance how much traffic is offered to
switches in the next level based on the ingress and egress bandwidth switches in the next level based on the ingress and egress bandwidth
they have. Current attempts rely mostly on specialized traffic they have. Current attempts rely mostly on specialized traffic
engineering via controller or leafs being aware of complete topology engineering via controller or leafs being aware of complete topology
with according cost and complexity. with according cost and complexity.
RIFT can support a very light weight mechanism that can deal with the RIFT can support a very light weight mechanism that can deal with the
problem in an approximative way based on the fact that RIFT is loop- problem in an approximative way based on the fact that RIFT is loop-
free. free.
4.3.6.1. Northbound Direction 5.3.6.1. Northbound Direction
Every RIFT node SHOULD compute the amount of northbound bandwith Every RIFT node SHOULD compute the amount of northbound bandwith
available through neighbors at higher level and modify distance available through neighbors at higher level and modify distance
received on default route from this neighbor. Those different received on default route from this neighbor. Those different
distances SHOULD be used to support weighted ECMP forwarding towards distances SHOULD be used to support weighted ECMP forwarding towards
higher level when using default route. We call such a distance higher level when using default route. We call such a distance
Bandwidth Adjusted Distance or BAD. This is best illustrated by a Bandwidth Adjusted Distance or BAD. This is best illustrated by a
simple example. simple example.
. 100 x 100 100 MBits . 100 x 100 100 MBits
. | x | | . | x | |
. +-+---+-+ +-+---+-+ . +-+---+-+ +-+---+-+
. | | | | . | | | |
. |Node111| |Node112| . |Spin111| |Spin112|
. +-+---+++ ++----+++ . +-+---+++ ++----+++
. |x || || || . |x || || ||
. || |+---------------+ || . || |+---------------+ ||
. || +---------------+| || . || +---------------+| ||
. || || || || . || || || ||
. || || || || . || || || ||
. -----All Links 10 MBit------- . -----All Links 10 MBit-------
. || || || || . || || || ||
. || || || || . || || || ||
. || +------------+| || || . || +------------+| || ||
. || |+------------+ || || . || |+------------+ || ||
. |x || || || . |x || || ||
. +-+---+++ +--++-+++ . +-+---+++ +--++-+++
. | | | | . | | | |
. |Leaf111| |Leaf112| . |Leaf111| |Leaf112|
. +-------+ +-------+ . +-------+ +-------+
Figure 10: Balancing Bandwidth Figure 22: Balancing Bandwidth
All links from Leafs in Figure 10 are assumed to 10 MBit/s bandwidth All links from Leafs in Figure 22 are assumed to 10 MBit/s bandwidth
while the uplinks one level further up are assumed to be 100 MBit/s. while the uplinks one level further up are assumed to be 100 MBit/s.
Further, in Figure 10 we assume that Leaf111 lost one of the parallel Further, in Figure 22 we assume that Leaf111 lost one of the parallel
links to Node 111 and with that wants to possibly push more traffic links to Spine 111 and with that wants to possibly push more traffic
onto Node 112. Leaf 112 has equal bandwidth to Node 111 and Node 112 onto Spine 112. Leaf 112 has equal bandwidth to Spine 111 and Spine
but Node 111 lost one of its uplinks. 112 but Spine 111 lost one of its uplinks.
The local modification of the received default route distance from The local modification of the received default route distance from
upper level is achieved by running a relatively simple algorithm upper level is achieved by running a relatively simple algorithm
where the bandwidth is weighted exponentially while the distance on where the bandwidth is weighted exponentially while the distance on
the default route represents a multiplier for the bandwidth weight the default route represents a multiplier for the bandwidth weight
for easy operational adjustements. for easy operational adjustements.
On a node L use Node TIEs to compute for each non-overloaded On a node L use Node TIEs to compute for each non-overloaded
northbound neighbor N three values: northbound neighbor N three values:
skipping to change at page 51, line 17 skipping to change at page 67, line 17
of all M_N_u. of all M_N_u.
For each advertised default route from a node N modify the advertised For each advertised default route from a node N modify the advertised
distance D to BAD = D * (1 + MAX_M_N_u - M_N_u) and use BAD instead distance D to BAD = D * (1 + MAX_M_N_u - M_N_u) and use BAD instead
of distance D to weight balance default forwarding towards N. of distance D to weight balance default forwarding towards N.
For the example above a simple table of values will help the For the example above a simple table of values will help the
understanding. We assume the default route distance is advertised understanding. We assume the default route distance is advertised
with D=1 everywhere and OVERSUBSCRIPTION_CONSTANT = 1. with D=1 everywhere and OVERSUBSCRIPTION_CONSTANT = 1.
+---------+---------+-------+-------+-----+ +---------+-----------+-------+-------+-----+
| Node | N | T_N_u | M_N_u | BAD | | Node | N | T_N_u | M_N_u | BAD |
+---------+---------+-------+-------+-----+ +---------+-----------+-------+-------+-----+
| Leaf111 | Node111 | 110 | 7 | 2 | | Leaf111 | Spine 111 | 110 | 7 | 2 |
+---------+---------+-------+-------+-----+ +---------+-----------+-------+-------+-----+
| Leaf111 | Node112 | 220 | 8 | 1 | | Leaf111 | Spine 112 | 220 | 8 | 1 |
+---------+---------+-------+-------+-----+ +---------+-----------+-------+-------+-----+
| Leaf112 | Node111 | 120 | 7 | 2 | | Leaf112 | Spine 111 | 120 | 7 | 2 |
+---------+---------+-------+-------+-----+ +---------+-----------+-------+-------+-----+
| Leaf112 | Node112 | 220 | 8 | 1 | | Leaf112 | Spine 112 | 220 | 8 | 1 |
+---------+---------+-------+-------+-----+ +---------+-----------+-------+-------+-----+
Table 4: BAD Computation Table 5: BAD Computation
All the multiplications and additions are saturating, i.e. when All the multiplications and additions are saturating, i.e. when
exceeding range of the bandwidth type are set to highest possible exceeding range of the bandwidth type are set to highest possible
value of the type. value of the type.
Observe that since BAD is only computed for default routes any BAD is only computed for default routes. A node MAY compute and use
disaggregated prefixes so PGP or disaggregated routes are not BAD for any disaggregated prefixes or other RIFT routes.
affected, however, a node MAY choose to compute and use BAD for other
routes.
Observe further that a change in available bandwidth will only affect Observe further that a change in available bandwidth will only affect
at maximum two levels down in the fabric, i.e. blast radius of at maximum two levels down in the fabric, i.e. blast radius of
bandwidth changes is contained. bandwidth changes is contained no matter its height.
4.3.6.2. Southbound Direction 5.3.6.2. Southbound Direction
Due to its loop free properties a node could take during S-SPF into Due to its loop free properties a node could take during S-SPF into
account the available bandwidth on the nodes in lower levels and account the available bandwidth on the nodes in lower levels and
modify the amount of traffic offered to next level's "southbound" modify the amount of traffic offered to next level's "southbound"
nodes based as what it sees is the total achievable maximum flow nodes based as what it sees is the total achievable maximum flow
through those nodes. It is worth observing that such computations through those nodes. It is worth observing that such computations
will work better if standardized but does not have to be necessarily. will work better if standardized but does not have to be necessarily.
As long the packet keeps on heading south it will take one of the As long the packet keeps on heading south it will take one of the
available paths and arrive at the intended destination. available paths and arrive at the intended destination.
Future versions of this document will fill in more details. Future versions of this document will fill in more details.
4.3.7. Label Binding 5.3.7. Label Binding
A node MAY advertise on its TIEs a locally significant, downstream A node MAY advertise on its TIEs a locally significant, downstream
assigned label for the according interface. One use of such label is assigned label for the according interface. One use of such label is
a hop-by-hop encapsulation allowing to easily distinguish forwarding a hop-by-hop encapsulation allowing to easily distinguish forwarding
planes served by a multiplicity of RIFT instances. planes served by a multiplicity of RIFT instances.
4.3.8. Segment Routing Support with RIFT 5.3.8. Segment Routing Support with RIFT
Recently, alternative architecture to reuse labels as segment Recently, alternative architecture to reuse labels as segment
identifiers [I-D.ietf-spring-segment-routing] has gained traction and identifiers [I-D.ietf-spring-segment-routing] has gained traction and
may present use cases in DC fabric that would justify its deployment. may present use cases in IP fabric that would justify its deployment.
Such use cases will either precondition an assignment of a label per Such use cases will either precondition an assignment of a label per
node (or other entities where the mechanisms are equivalent) or a node (or other entities where the mechanisms are equivalent) or a
global assignment and a knowledge of topology everywhere to compute global assignment and a knowledge of topology everywhere to compute
segment stacks of interest. We deal with the two issues separately. segment stacks of interest. We deal with the two issues separately.
4.3.8.1. Global Segment Identifiers Assignment 5.3.8.1. Global Segment Identifiers Assignment
Global segment identifiers are normally assumed to be provided by Global segment identifiers are normally assumed to be provided by
some kind of a centralized "controller" instance and distributed to some kind of a centralized "controller" instance and distributed to
other entities. This can be performed in RIFT by attaching a other entities. This can be performed in RIFT by attaching a
controller to the superspine nodes at the top of the fabric where the controller to the Top-of-Fabric nodes at the top of the fabric where
whole topology is always visible, assign such identifiers and then the whole topology is always visible, assign such identifiers and
distribute those via the KV mechanism towards all nodes so they can then distribute those via the KV mechanism towards all nodes so they
perform things like probing the fabric for failures using a stack of can perform things like probing the fabric for failures using a stack
segments. of segments.
4.3.8.2. Distribution of Topology Information 5.3.8.2. Distribution of Topology Information
Some segment routing use cases seem to precondition full knowledge of Some segment routing use cases seem to precondition full knowledge of
fabric topology in all nodes which can be performed albeit at the fabric topology in all nodes which can be performed albeit at the
loss of one of highly desirable properties of RIFT, namely minimal loss of one of highly desirable properties of RIFT, namely minimal
blast radius. Basically, RIFT can function as a flat IGP by blast radius. Basically, RIFT can function as a flat IGP by
switching off its flooding scopes. All nodes will end up with full switching off its flooding scopes. All nodes will end up with full
topology view and albeit the N-SPF and S-SPF are still performed topology view and albeit the N-SPF and S-SPF are still performed
based on RIFT rules, any computation with segment identifiers that based on RIFT rules, any computation with segment identifiers that
needs full topology can use it. needs full topology can use it.
Beside blast radius problem, excessive flooding may present Beside blast radius problem, excessive flooding may present
significant load on implementations. significant load on implementations.
4.3.9. Leaf to Leaf Procedures 5.3.9. Leaf to Leaf Procedures
RIFT can optionally allow special leaf East-West adjacencies under RIFT can optionally allow special leaf East-West adjacencies under
additional set of rules. The leaf supporting those procedures MUST: additional set of rules. The leaf supporting those procedures MUST:
advertise the LEAF_2_LEAF flag in node capabilities AND advertise the LEAF_2_LEAF flag in node capabilities AND
set the overload bit on all leaf's node TIEs AND set the overload bit on all leaf's node TIEs AND
flood only node's own north and south TIEs over E-W leaf flood only node's own north and south TIEs over E-W leaf
adjacencies AND adjacencies AND
skipping to change at page 53, line 30 skipping to change at page 69, line 30
install a discard route for any advertised aggregate in leaf's install a discard route for any advertised aggregate in leaf's
TIEs AND TIEs AND
never form southbound adjacencies. never form southbound adjacencies.
This will allow the E-W leaf nodes to exchange traffic strictly for This will allow the E-W leaf nodes to exchange traffic strictly for
the prefixes advertised in each other's north prefix TIEs (since the the prefixes advertised in each other's north prefix TIEs (since the
southbound computation will find the reverse direction in the other southbound computation will find the reverse direction in the other
node's TIE and install its north prefixes). node's TIE and install its north prefixes).
4.3.10. Other End-to-End Services 5.3.10. Address Family and Multi Topology Considerations
Losing full, flat topology information at every node will have an
impact on some of the end-to-end network services. This is the price
paid for minimal disturbance in case of failures and reduced flooding
and memory requirements on nodes lower south in the level hierarchy.
4.3.11. Address Family and Multi Topology Considerations
Multi-Topology (MT)[RFC5120] and Multi-Instance (MI)[RFC6822] is used Multi-Topology (MT)[RFC5120] and Multi-Instance (MI)[RFC6822] is used
today in link-state routing protocols to support several domains on today in link-state routing protocols to support several domains on
the same physical topology. RIFT supports this capability by the same physical topology. RIFT supports this capability by
carrying transport ports in the LIE protocol exchanges. Multiplexing carrying transport ports in the LIE protocol exchanges. Multiplexing
of LIEs can be achieved by either choosing varying multicast of LIEs can be achieved by either choosing varying multicast
addresses or ports on the same address. addresses or ports on the same address.
BFD interactions in Section 4.3.5 are implementation dependent when BFD interactions in Section 5.3.5 are implementation dependent when
multiple RIFT instances run on the same link. multiple RIFT instances run on the same link.
4.3.12. Reachability of Internal Nodes in the Fabric 5.3.11. Reachability of Internal Nodes in the Fabric
RIFT does not precondition that its nodes have reachable addresses RIFT does not precondition that its nodes have reachable addresses
albeit for operational purposes this is clearly desirable. Under albeit for operational purposes this is clearly desirable. Under
normal operating conditions this can be easily achieved by e.g. normal operating conditions this can be easily achieved by e.g.
injecting the node's loopback address into North Prefix TIEs. injecting the node's loopback address into North Prefix TIEs.
Things get more interesting in case a node looses all its northbound Things get more interesting in case a node looses all its northbound
adjacencies but is not at the top of the fabric. In such a case a adjacencies but is not at the top of the fabric. That is outside the
node that detects that some other members at its level are scope of this document and may be covered in a separate document
advertising northbound adjacencies MAY inject its loopback address about policy guided prefixes [PGP reference].
into southbound PGP TIE and become reachable "from the south" that
way. Further, a solution may be implemented where based on e.g. a
"well known" community such a southbound PGP is reflected at level 0
and advertised as northbound PGP again to allow for "reachability
from the north" at the cost of additional flooding.
4.3.13. One-Hop Healing of Levels with East-West Links 5.3.12. One-Hop Healing of Levels with East-West Links
Based on the rules defined in Section 4.2.5, Section 4.2.3.7 and Based on the rules defined in Section 5.2.4, Section 5.2.3.7 and
given presence of E-W links, RIFT can provide a one-hop protection of given presence of E-W links, RIFT can provide a one-hop protection of
nodes that lost all their northbound links or in other complex link nodes that lost all their northbound links or in other complex link
set failure scenarios. Section 5.4 explains the resulting behavior set failure scenarios except at Top-of-Fabric where the links are
based on one such example. used exclusively to flood topology information in multi-plane
designs. Section 6.4 explains the resulting behavior based on one
such example.
5. Examples 6. Examples
5.1. Normal Operation 6.1. Normal Operation
This section describes RIFT deployment in the example topology This section describes RIFT deployment in the example topology
without any node or link failures. We disregard flooding reduction without any node or link failures. We disregard flooding reduction
for simplicity's sake. for simplicity's sake.
As first step, the following bi-directional adjacencies will be As first step, the following bi-directional adjacencies will be
created (and any other links that do not fulfill LIE rules in created (and any other links that do not fulfill LIE rules in
Section 4.2.2 disregarded): Section 5.2.2 disregarded):
1. Spine 21 (PoD 0) to Node 111, Node 112, Node 121, and Node 122 1. Spine 21 (PoD 0) to Spine 111, Spine 112, Spine 121, and Spine
122
2. Spine 22 (PoD 0) to Node 111, Node 112, Node 121, and Node 122 2. Spine 22 (PoD 0) to Spine 111, Spine 112, Spine 121, and Spine
122
3. Node 111 to Leaf 111, Leaf 112 3. Spine 111 to Leaf 111, Leaf 112
4. Node 112 to Leaf 111, Leaf 112 4. Spine 112 to Leaf 111, Leaf 112
5. Node 121 to Leaf 121, Leaf 122 5. Spine 121 to Leaf 121, Leaf 122
6. Node 122 to Leaf 121, Leaf 122 6. Spine 122 to Leaf 121, Leaf 122
Consequently, N-TIEs would be originated by Node 111 and Node 112 and
each set would be sent to both Spine 21 and Spine 22. N-TIEs also Consequently, N-TIEs would be originated by Spine 111 and Spine 112
would be originated by Leaf 111 (w/ Prefix 111) and Leaf 112 (w/ and each set would be sent to both Spine 21 and Spine 22. N-TIEs
also would be originated by Leaf 111 (w/ Prefix 111) and Leaf 112 (w/
Prefix 112 and the multi-homed prefix) and each set would be sent to Prefix 112 and the multi-homed prefix) and each set would be sent to
Node 111 and Node 112. Node 111 and Node 112 would then flood these Spine 111 and Spine 112. Spine 111 and Spine 112 would then flood
N-TIEs to Spine 21 and Spine 22. these N-TIEs to Spine 21 and Spine 22.
Similarly, N-TIEs would be originated by Node 121 and Node 122 and Similarly, N-TIEs would be originated by Spine 121 and Spine 122 and
each set would be sent to both Spine 21 and Spine 22. N-TIEs also each set would be sent to both Spine 21 and Spine 22. N-TIEs also
would be originated by Leaf 121 (w/ Prefix 121 and the multi-homed would be originated by Leaf 121 (w/ Prefix 121 and the multi-homed
prefix) and Leaf 122 (w/ Prefix 122) and each set would be sent to prefix) and Leaf 122 (w/ Prefix 122) and each set would be sent to
Node 121 and Node 122. Node 121 and Node 122 would then flood these Spine 121 and Spine 122. Spine 121 and Spine 122 would then flood
N-TIEs to Spine 21 and Spine 22. these N-TIEs to Spine 21 and Spine 22.
At this point both Spine 21 and Spine 22, as well as any controller At this point both Spine 21 and Spine 22, as well as any controller
to which they are connected, would have the complete network to which they are connected, would have the complete network
topology. At the same time, Node 111/112/121/122 hold only the topology. At the same time, Spine 111/112/121/122 hold only the
N-ties of level 0 of their respective PoD. Leafs hold only their own N-ties of level 0 of their respective PoD. Leafs hold only their own
N-TIEs. N-TIEs.
S-TIEs with adjacencies and a default IP prefix would then be S-TIEs with adjacencies and a default IP prefix would then be
originated by Spine 21 and Spine 22 and each would be flooded to Node originated by Spine 21 and Spine 22 and each would be flooded to
111, Node 112, Node 121, and Node 122. Node 111, Node 112, Node 121, Spine 111, Spine 112, Spine 121, and Spine 122. Spine 111, Spine
and Node 122 would each send the S-TIE from Spine 21 to Spine 22 and 112, Spine 121, and Spine 122 would each send the S-TIE from Spine 21
the S-TIE from Spine 22 to Spine 21. (S-TIEs are reflected up to to Spine 22 and the S-TIE from Spine 22 to Spine 21. (S-TIEs are
level from which they are received but they are NOT propagated reflected up to level from which they are received but they are NOT
southbound.) propagated southbound.)
An S Tie with a default IP prefix would be originated by Node 111 and A S-TIE with a default IP prefix would be originated by Node 111 and
Node 112 and each would be sent to Leaf 111 and Leaf 112. Leaf 111 Spine 112 and each would be sent to Leaf 111 and Leaf 112.
and Leaf 112 would each send the S-TIE from Node 111 to Node 112 and
the S-TIE from Node 112 to Node 111.
Similarly, an S Tie with a default IP prefix would be originated by Similarly, an S-TIE with a default IP prefix would be originated by
Node 121 and Node 122 and each would be sent to Leaf 121 and Leaf Node 121 and Spine 122 and each would be sent to Leaf 121 and Leaf
122. Leaf 121 and Leaf 122 would each send the S-TIE from Node 121 122. At this point IP connectivity with maximum possible ECMP has
to Node 122 and the S-TIE from Node 122 to Node 121. At this point been established between the leafs while constraining the amount of
IP connectivity with maximum possible ECMP has been established information held by each node to the minimum necessary for normal
between the leafs while constraining the amount of information held operation and dealing with failures.
by each node to the minimum necessary for normal operation and
dealing with failures. 6.2. Leaf Link Failure
5.2. Leaf Link Failure
. | | | | . | | | |
.+-+---+-+ +-+---+-+ .+-+---+-+ +-+---+-+
.| | | | .| | | |
.|Node111| |Node112| .|Spin111| |Spin112|
.+-+---+-+ ++----+-+ .+-+---+-+ ++----+-+
. | | | | . | | | |
. | +---------------+ X . | +---------------+ X
. | | | X Failure . | | | X Failure
. | +-------------+ | X . | +-------------+ | X
. | | | | . | | | |
.+-+---+-+ +--+--+-+ .+-+---+-+ +--+--+-+
.| | | | .| | | |
.|Leaf111| |Leaf112| .|Leaf111| |Leaf112|
.+-------+ +-------+ .+-------+ +-------+
. + + . + +
. Prefix111 Prefix112 . Prefix111 Prefix112
Figure 11: Single Leaf link failure Figure 23: Single Leaf link failure
In case of a failing leaf link between node 112 and leaf 112 the In case of a failing leaf link between spine 112 and leaf 112 the
link-state information will cause re-computation of the necessary SPF link-state information will cause re-computation of the necessary SPF
and the higher levels will stop forwarding towards prefix 112 through and the higher levels will stop forwarding towards prefix 112 through
node 112. Only nodes 111 and 112, as well as both spines will see spine 112. Only spines 111 and 112, as well as both spines will see
control traffic. Leaf 111 will receive a new S-TIE from node 112 and control traffic. Leaf 111 will receive a new S-TIE from spine 112
reflect back to node 111. Node 111 will de-aggregate prefix 111 and and reflect back to spine 111. Spine 111 will de-aggregate prefix
prefix 112 but we will not describe it further here since de- 111 and prefix 112 but we will not describe it further here since de-
aggregation is emphasized in the next example. It is worth observing aggregation is emphasized in the next example. It is worth observing
however in this example that if leaf 111 would keep on forwarding however in this example that if leaf 111 would keep on forwarding
traffic towards prefix 112 using the advertised south-bound default traffic towards prefix 112 using the advertised south-bound default
of node 112 the traffic would end up on spine 21 and spine 22 and of spine 112 the traffic would end up on Top-of-Fabric 21 and ToF 22
cross back into pod 1 using node 111. This is arguably not as bad as and cross back into pod 1 using spine 111. This is arguably not as
black-holing present in the next example but clearly undesirable. bad as black-holing present in the next example but clearly
Fortunately, de-aggregation prevents this type of behavior except for undesirable. Fortunately, de-aggregation prevents this type of
a transitory period of time. behavior except for a transitory period of time.
5.3. Partitioned Fabric 6.3. Partitioned Fabric
. +--------+ +--------+ S-TIE of Spine21 . +--------+ +--------+ S-TIE of Spine21
. | | | | received by . | | | | received by
. |Spine 21| |Spine 22| reflection of . |ToF 21| |ToF 22| south reflection of
. ++-+--+-++ ++-+--+-++ Nodes 112 and 111 . ++-+--+-++ ++-+--+-++ spines 112 and 111
. | | | | | | | | . | | | | | | | |
. | | | | | | | 0/0 . | | | | | | | 0/0
. | | | | | | | | . | | | | | | | |
. | | | | | | | | . | | | | | | | |
. +--------------+ | +--- XXXXXX + | | | +---------------+ . +--------------+ | +--- XXXXXX + | | | +---------------+
. | | | | | | | | . | | | | | | | |
. | +-----------------------------+ | | | . | +-----------------------------+ | | |
. 0/0 | | | | | | | . 0/0 | | | | | | |
. | 0/0 0/0 +- XXXXXXXXXXXXXXXXXXXXXXXXX -+ | . | 0/0 0/0 +- XXXXXXXXXXXXXXXXXXXXXXXXX -+ |
. | 1.1/16 | | | | | | . | 1.1/16 | | | | | |
. | | +-+ +-0/0-----------+ | | . | | +-+ +-0/0-----------+ | |
. | | | 1.1./16 | | | | . | | | 1.1./16 | | | |
.+-+----++ +-+-----+ ++-----0/0 ++----0/0 .+-+----++ +-+-----+ ++-----0/0 ++----0/0
.| | | | | 1.1/16 | 1.1/16 .| | | | | 1.1/16 | 1.1/16
.|Node111| |Node112| |Node121| |Node122| .|Spin111| |Spin112| |Spin121| |Spin122|
.+-+---+-+ ++----+-+ +-+---+-+ ++---+--+ .+-+---+-+ ++----+-+ +-+---+-+ ++---+--+
. | | | | | | | | . | | | | | | | |
. | +---------------+ | | +----------------+ | . | +---------------+ | | +----------------+ |
. | | | | | | | | . | | | | | | | |
. | +-------------+ | | | +--------------+ | | . | +-------------+ | | | +--------------+ | |
. | | | | | | | | . | | | | | | | |
.+-+---+-+ +--+--+-+ +-+---+-+ +---+-+-+ .+-+---+-+ +--+--+-+ +-+---+-+ +---+-+-+
.| | | | | | | | .| | | | | | | |
.|Leaf111| |Leaf112| |Leaf121| |Leaf122| .|Leaf111| |Leaf112| |Leaf121| |Leaf122|
.+-+-----+ ++------+ +-----+-+ +-+-----+ .+-+-----+ ++------+ +-----+-+ +-+-----+
. + + + + . + + + +
. Prefix111 Prefix112 Prefix121 Prefix122 . Prefix111 Prefix112 Prefix121 Prefix122
. 1.1/16 . 1.1/16
Figure 12: Fabric partition Figure 24: Fabric partition
Figure 12 shows the arguably most catastrophic but also the most Figure 24 shows the arguably most catastrophic but also the most
interesting case. Spine 21 is completely severed from access to interesting case. Spine 21 is completely severed from access to
Prefix 121 (we use in the figure 1.1/16 as example) by double link Prefix 121 (we use in the figure 1.1/16 as example) by double link
failure. However unlikely, if left unresolved, forwarding from leaf failure. However unlikely, if left unresolved, forwarding from leaf
111 and leaf 112 to prefix 121 would suffer 50% black-holing based on 111 and leaf 112 to prefix 121 would suffer 50% black-holing based on
pure default route advertisements by spine 21 and spine 22. pure default route advertisements by Top-of-Fabric 21 and ToF 22.
The mechanism used to resolve this scenario is hinging on the The mechanism used to resolve this scenario is hinging on the
distribution of southbound representation by spine 21 that is distribution of southbound representation by Top-of-Fabric 21 that is
reflected by node 111 and node 112 to spine 22. Spine 22, having reflected by spine 111 and spine 112 to ToF 22. Spine 22, having
computed reachability to all prefixes in the network, advertises with computed reachability to all prefixes in the network, advertises with
the default route the ones that are reachable only via lower level the default route the ones that are reachable only via lower level
neighbors that spine 21 does not show an adjacency to. That results neighbors that ToF 21 does not show an adjacency to. That results in
in node 111 and node 112 obtaining a longest-prefix match to prefix spine 111 and spine 112 obtaining a longest-prefix match to prefix
121 which leads through spine 22 and prevents black-holing through 121 which leads through ToF 22 and prevents black-holing through ToF
spine 21 still advertising the 0/0 aggregate only. 21 still advertising the 0/0 aggregate only.
The prefix 121 advertised by spine 22 does not have to be propagated The prefix 121 advertised by Top-of-Fabric 22 does not have to be
further towards leafs since they do no benefit from this information. propagated further towards leafs since they do no benefit from this
Hence the amount of flooding is restricted to spine 21 reissuing its information. Hence the amount of flooding is restricted to ToF 21
S-TIEs and reflection of those by node 111 and node 112. The reissuing its S-TIEs and south reflection of those by spine 111 and
resulting SPF in spine 22 issues a new prefix S-TIEs containing spine 112. The resulting SPF in ToF 22 issues a new prefix S-TIEs
1.1/16. None of the leafs become aware of the changes and the containing 1.1/16. None of the leafs become aware of the changes and
failure is constrained strictly to the level that became partitioned. the failure is constrained strictly to the level that became
partitioned.
To finish with an example of the resulting sets computed using To finish with an example of the resulting sets computed using
notation introduced in Section 4.2.8, spine 22 constructs the notation introduced in Section 5.2.5, Top-of-Fabric 22 constructs the
following sets: following sets:
|R = Prefix 111, Prefix 112, Prefix 121, Prefix 122 |R = Prefix 111, Prefix 112, Prefix 121, Prefix 122
|H (for r=Prefix 111) = Node 111, Node 112 |H (for r=Prefix 111) = Spine 111, Spine 112
|H (for r=Prefix 112) = Node 111, Node 112 |H (for r=Prefix 112) = Spine 111, Spine 112
|H (for r=Prefix 121) = Node 121, Node 122 |H (for r=Prefix 121) = Spine 121, Spine 122
|H (for r=Prefix 122) = Node 121, Node 122 |H (for r=Prefix 122) = Spine 121, Spine 122
|A (for Spine 21) = Node 111, Node 112 |A (for Spine 21) = Spine 111, Spine 112
With that and |H (for r=prefix 121) and |H (for r=prefix 122) being With that and |H (for r=prefix 121) and |H (for r=prefix 122) being
disjoint from |A (for spine 21), spine 22 will originate an S-TIE disjoint from |A (for Top-of-Fabric 21), ToF 22 will originate an
with prefix 121 and prefix 122, that is flooded to nodes 112, 112, S-TIE with prefix 121 and prefix 122, that is flooded to spines 112,
121 and 122. 112, 121 and 122.
5.4. Northbound Partitioned Router and Optional East-West Links 6.4. Northbound Partitioned Router and Optional East-West Links
. + + + . + + +
. X N1 | N2 | N3 . X N1 | N2 | N3
. X | | . X | |
.+--+----+ +--+----+ +--+-----+ .+--+----+ +--+----+ +--+-----+
.| |0/0> <0/0| |0/0> <0/0| | .| |0/0> <0/0| |0/0> <0/0| |
.| A01 +----------+ A02 +----------+ A03 | Level 1 .| A01 +----------+ A02 +----------+ A03 | Level 1
.++-+-+--+ ++--+--++ +---+-+-++ .++-+-+--+ ++--+--++ +---+-+-++
. | | | | | | | | | . | | | | | | | | |
. | | +----------------------------------+ | | | . | | +----------------------------------+ | | |
. | | | | | | | | | . | | | | | | | | |
skipping to change at page 59, line 25 skipping to change at page 75, line 25
. | | | | | | | | | . | | | | | | | | |
. | +----------------+ | +-----------------+ | . | +----------------+ | +-----------------+ |
. | | | | | | | | | . | | | | | | | | |
. | | +------------------------------------+ | | . | | +------------------------------------+ | |
. | | | | | | | | | . | | | | | | | | |
.++-+-+--+ | +---+---+ | +-+---+-++ .++-+-+--+ | +---+---+ | +-+---+-++
.| | +-+ +-+ | | .| | +-+ +-+ | |
.| L01 | | L02 | | L03 | Level 0 .| L01 | | L02 | | L03 | Level 0
.+-------+ +-------+ +--------+ .+-------+ +-------+ +--------+
Figure 13: North Partitioned Router Figure 25: North Partitioned Router
Figure 13 shows a part of a fabric where level 1 is horizontally Figure 25 shows a part of a fabric where level 1 is horizontally
connected and A01 lost its only northbound adjacency. Based on N-SPF connected and A01 lost its only northbound adjacency. Based on N-SPF
rules in Section 4.2.5.1 A01 will compute northbound reachability by rules in Section 5.2.4.1 A01 will compute northbound reachability by
using the link A01 to A02 (whereas A02 will NOT use this link during using the link A01 to A02 (whereas A02 will NOT use this link during
N-SPF). Hence A01 will still advertise the default towards level 0 N-SPF). Hence A01 will still advertise the default towards level 0
and route unidirectionally using the horizontal link. Moreover, and route unidirectionally using the horizontal link.
based on Section 4.3.12 it may advertise its loopback address as
south PGP to remain reachable "from the south" for operational
purposes. This is necessary since A02 will NOT route towards A01
using the E-W link (doing otherwise may form routing loops).
As further consideration, the moment A02 looses link N2 the situation As further consideration, the moment A02 looses link N2 the situation
evolves again. A01 will have no more northbound reachability while evolves again. A01 will have no more northbound reachability while
still seeing A03 advertising northbound adjacencies in its south node still seeing A03 advertising northbound adjacencies in its south node
tie. With that it will stop advertising a default route due to tie. With that it will stop advertising a default route due to
Section 4.2.3.7. Moreover, A02 may now inject its loopback address Section 5.2.3.7.
as south PGP.
6. Implementation and Operation: Further Details 6.5. Multi-Plane Fabric and Negative Disaggregation
6.1. Considerations for Leaf-Only Implementation
TODO
7. Implementation and Operation: Further Details
7.1. Considerations for Leaf-Only Implementation
Ideally RIFT can be stretched out to the loWest level in the IP Ideally RIFT can be stretched out to the loWest level in the IP
fabric to integrate ToRs or even servers. Since those entities would fabric to integrate ToRs or even servers. Since those entities would
run as leafs only, it is worth to observe that a leaf only version is run as leafs only, it is worth to observe that a leaf only version is
significantly simpler to implement and requires much less resources: significantly simpler to implement and requires much less resources:
1. Under normal conditions, the leaf needs to support a multipath 1. Under normal conditions, the leaf needs to support a multipath
default route only. In worst partitioning case it has to be default route only. In worst partitioning case it has to be
capable of accommodating all the leaf routes in its own POD to capable of accommodating all the leaf routes in its own PoD to
prevent black-holing. prevent black-holing.
2. Leaf nodes hold only their own N-TIEs and S-TIEs of Level 1 nodes 2. Leaf nodes hold only their own N-TIEs and S-TIEs of Level 1 nodes
they are connected to; so overall few in numbers. they are connected to; so overall few in numbers.
3. Leaf node does not have to support flooding reduction and de- 3. Leaf node does not have to support flooding reduction or any type
aggregation. of de-aggregation computation or propagation.
4. Unless optional leaf-2-leaf procedures are desired default route 4. Unless optional leaf-2-leaf procedures are desired default route
origination, S-TIE origination is unnecessary. origination and S-TIE origination is unnecessary.
6.2. Adaptations to Other Proposed Data Center Topologies 7.2. Adaptations to Other Proposed Data Center Topologies
. +-----+ +-----+ . +-----+ +-----+
. | | | | . | | | |
.+-+ S0 | | S1 | .+-+ S0 | | S1 |
.| ++---++ ++---++ .| ++---++ ++---++
.| | | | | .| | | | |
.| | +------------+ | .| | +------------+ |
.| | | +------------+ | .| | | +------------+ |
.| | | | | .| | | | |
.| ++-+--+ +--+-++ .| ++-+--+ +--+-++
skipping to change at page 60, line 48 skipping to change at page 76, line 44
.| +-+--++ ++---++ .| +-+--++ ++---++
.| | | | | .| | | | |
.| | +------------+ | .| | +------------+ |
.| | +-----------+ | | .| | +-----------+ | |
.| | | | | .| | | | |
.| +-+-+-+ +--+-++ .| +-+-+-+ +--+-++
.+-+ | | | .+-+ | | |
. | L0 | | L1 | . | L0 | | L1 |
. +-----+ +-----+ . +-----+ +-----+
Figure 14: Level Shortcut Figure 26: Level Shortcut
Strictly speaking, RIFT is not limited to Clos variations only. The Strictly speaking, RIFT is not limited to Clos variations only. The
protocol preconditions only a sense of 'compass rose direction' protocol preconditions only a sense of 'compass rose direction'
achieved by configuration (or derivation) of levels and other achieved by configuration (or derivation) of levels and other
topologies are possible within this framework. So, conceptually, one topologies are possible within this framework. So, conceptually, one
could include leaf to leaf links and even shortcut between levels but could include leaf to leaf links and even shortcut between levels but
certain requirements in Section 3 will not be met anymore. As an certain requirements in Section 4 will not be met anymore. As an
example, shortcutting levels illustrated in Figure 14 will lead example, shortcutting levels illustrated in Figure 26 will lead
either to suboptimal routing when L0 sends traffic to L1 (since using either to suboptimal routing when L0 sends traffic to L1 (since using
S0's default route will lead to the traffic being sent back to A0 or S0's default route will lead to the traffic being sent back to A0 or
A1) or the leafs need each other's routes installed to understand A1) or the leafs need each other's routes installed to understand
that only A0 and A1 should be used to talk to each other. that only A0 and A1 should be used to talk to each other.
Whether such modifications of topology constraints make sense is Whether such modifications of topology constraints make sense is
dependent on many technology variables and the exhausting treatment dependent on many technology variables and the exhausting treatment
of the topic is definitely outside the scope of this document. of the topic is definitely outside the scope of this document.
6.3. Originating Non-Default Route Southbound 7.3. Originating Non-Default Route Southbound
Obviously, an implementation may choose to originate southbound Obviously, an implementation may choose to originate southbound
instead of a strict default route (as described in Section 4.2.3.7) a instead of a strict default route (as described in Section 5.2.3.7) a
shorter prefix P' but in such a scenario all addresses carried within shorter prefix P' but in such a scenario all addresses carried within
the RIFT domain must be contained within P'. the RIFT domain must be contained within P'.
7. Security Considerations 8. Security Considerations
8.1. General
The protocol has provisions for nonces and can include authentication The protocol has provisions for nonces and can include authentication
mechanisms in the future comparable to [RFC5709] and [RFC7987]. mechanisms in the future comparable to [RFC5709] and [RFC7987].
One can consider additionally attack vectors where a router may One can consider additionally attack vectors where a router may
reboot many times while changing its system ID and pollute the reboot many times while changing its system ID and pollute the
network with many stale TIEs or TIEs are sent with very long network with many stale TIEs or TIEs are sent with very long
lifetimes and not cleaned up when the routes vanishes. Those attack lifetimes and not cleaned up when the routes vanishes. Those attack
vectors are not unique to RIFT. Given large memory footprints vectors are not unique to RIFT. Given large memory footprints
available today those attacks should be relatively benign. Otherwise available today those attacks should be relatively benign. Otherwise
a node can implement a strategy of e.g. discarding contents of all a node SHOULD implement a strategy of discarding contents of all TIEs
TIEs of nodes that were not present in the SPF tree over a certain that were not present in the SPF tree over a certain, configurable
period of time. Since the protocol, like all modern link-state period of time. Since the protocol, like all modern link-state
protocols, is self-stabilizing and will advertise the presence of protocols, is self-stabilizing and will advertise the presence of
such TIEs to its neighbors, they can be re-requested again if a such TIEs to its neighbors, they can be re-requested again if a
computation finds that it sees an adjacency formed towards the system computation finds that it sees an adjacency formed towards the system
ID of the discarded TIEs. ID of the discarded TIEs.
Section 4.2.9 presents many attack vectors in untrusted environments, 8.2. ZTP
Section 5.2.7 presents many attack vectors in untrusted environments,
starting with nodes that oscillate their level offers to the starting with nodes that oscillate their level offers to the
possiblity of a node offering a three way adjacency with the highest possiblity of a node offering a three way adjacency with the highest
possible level value with a very long holdtime trying to put itself possible level value with a very long holdtime trying to put itself
"on top of the lattice" and with that gaining access to the whole "on top of the lattice" and with that gaining access to the whole
southbound topology. Session authentication mechanisms are necessary southbound topology. Session authentication mechanisms are necessary
in environments where this is possible. in environments where this is possible.
8. IANA Considerations 8.3. Lifetime
The protocol uses in TIE flooding the traditional lifetime approach
that is vulnerable to sophisticated attack vectors under normal
circumstances. However, on IP fabrics with some kind of, even
coarse, clock synchronization, RIFT allows to recognize such attacks
by including optional, protected information at origin.
9. IANA Considerations
This specification will request at an opportune time multiple This specification will request at an opportune time multiple
registry points to exchange protocol packets in a standardized way, registry points to exchange protocol packets in a standardized way,
amongst them multicast address assignments and standard port numbers. amongst them multicast address assignments and standard port numbers.
The schema itself defines many values and codepoints which can be The schema itself defines many values and codepoints which can be
considered registries themselves. considered registries themselves.
9. Acknowledgments 10. Acknowledgments
Many thanks to Naiming Shen for some of the early discussions around Many thanks to Naiming Shen for some of the early discussions around
the topic of using IGPs for routing in topologies related to Clos. the topic of using IGPs for routing in topologies related to Clos.
Russ White to be especially acknowledged for the key conversation on Russ White to be especially acknowledged for the key conversation on
epistomology that allowed to tie current asynchronous distributed epistomology that allowed to tie current asynchronous distributed
systems theory results to a modern protocol design presented here. systems theory results to a modern protocol design presented here.
Adrian Farrel, Joel Halpern, Jeffrey Zhang and Krzysztof Szarkowicz Adrian Farrel, Joel Halpern, Jeffrey Zhang, Krzysztof Szarkowicz,
provided thoughtful comments that improved the readability of the Nagendra Kumar provided thoughtful comments that improved the
document and found good amount of corners where the light failed to readability of the document and found good amount of corners where
shine. Kris Price was first to mention single router, single arm the light failed to shine. Kris Price was first to mention single
default considerations. Jeff Tantsura helped out with some initial router, single arm default considerations. Jeff Tantsura helped out
thoughts on BFD interactions while Jeff Haas corrected several with some initial thoughts on BFD interactions while Jeff Haas
misconceptions about BFD's finer points. Artur Makutunowicz pointed corrected several misconceptions about BFD's finer points. Artur
out many possible improvements and acted as sounding board in regard Makutunowicz pointed out many possible improvements and acted as
to modern protocol implementation techniques RIFT is exploring. sounding board in regard to modern protocol implementation techniques
Barak Gafni formalized first time clearly the problem of partitioned RIFT is exploring. Barak Gafni formalized first time clearly the
spine on a (clean) napkin in Singapore. problem of partitioned spine and fallen leafs on a (clean) napkin in
Singapore that led to the very important part of the specification
centered around multiple Top-of-Fabric planes and negative
disaggregation.
10. References 11. References
10.1. Normative References 11.1. Normative References
[I-D.ietf-6lo-rfc6775-update] [I-D.ietf-6lo-rfc6775-update]
Thubert, P., Nordmark, E., Chakrabarti, S., and C. Thubert, P., Nordmark, E., Chakrabarti, S., and C.
Perkins, "Registration Extensions for 6LoWPAN Neighbor Perkins, "Registration Extensions for 6LoWPAN Neighbor
Discovery", draft-ietf-6lo-rfc6775-update-21 (work in Discovery", draft-ietf-6lo-rfc6775-update-21 (work in
progress), June 2018. progress), June 2018.
[ISO10589] [ISO10589]
ISO "International Organization for Standardization", ISO "International Organization for Standardization",
"Intermediate system to Intermediate system intra-domain "Intermediate system to Intermediate system intra-domain
skipping to change at page 64, line 30 skipping to change at page 80, line 41
[RFC6234] Eastlake 3rd, D. and T. Hansen, "US Secure Hash Algorithms [RFC6234] Eastlake 3rd, D. and T. Hansen, "US Secure Hash Algorithms
(SHA and SHA-based HMAC and HKDF)", RFC 6234, (SHA and SHA-based HMAC and HKDF)", RFC 6234,
DOI 10.17487/RFC6234, May 2011, DOI 10.17487/RFC6234, May 2011,
<https://www.rfc-editor.org/info/rfc6234>. <https://www.rfc-editor.org/info/rfc6234>.
[RFC6822] Previdi, S., Ed., Ginsberg, L., Shand, M., Roy, A., and D. [RFC6822] Previdi, S., Ed., Ginsberg, L., Shand, M., Roy, A., and D.
Ward, "IS-IS Multi-Instance", RFC 6822, Ward, "IS-IS Multi-Instance", RFC 6822,
DOI 10.17487/RFC6822, December 2012, DOI 10.17487/RFC6822, December 2012,
<https://www.rfc-editor.org/info/rfc6822>. <https://www.rfc-editor.org/info/rfc6822>.
[RFC7752] Gredler, H., Ed., Medved, J., Previdi, S., Farrel, A., and
S. Ray, "North-Bound Distribution of Link-State and
Traffic Engineering (TE) Information Using BGP", RFC 7752,
DOI 10.17487/RFC7752, March 2016,
<https://www.rfc-editor.org/info/rfc7752>.
[RFC7855] Previdi, S., Ed., Filsfils, C., Ed., Decraene, B., [RFC7855] Previdi, S., Ed., Filsfils, C., Ed., Decraene, B.,
Litkowski, S., Horneffer, M., and R. Shakir, "Source Litkowski, S., Horneffer, M., and R. Shakir, "Source
Packet Routing in Networking (SPRING) Problem Statement Packet Routing in Networking (SPRING) Problem Statement
and Requirements", RFC 7855, DOI 10.17487/RFC7855, May and Requirements", RFC 7855, DOI 10.17487/RFC7855, May
2016, <https://www.rfc-editor.org/info/rfc7855>. 2016, <https://www.rfc-editor.org/info/rfc7855>.
[RFC7938] Lapukhov, P., Premji, A., and J. Mitchell, Ed., "Use of [RFC7938] Lapukhov, P., Premji, A., and J. Mitchell, Ed., "Use of
BGP for Routing in Large-Scale Data Centers", RFC 7938, BGP for Routing in Large-Scale Data Centers", RFC 7938,
DOI 10.17487/RFC7938, August 2016, DOI 10.17487/RFC7938, August 2016,
<https://www.rfc-editor.org/info/rfc7938>. <https://www.rfc-editor.org/info/rfc7938>.
skipping to change at page 65, line 5 skipping to change at page 81, line 20
[RFC7987] Ginsberg, L., Wells, P., Decraene, B., Przygienda, T., and [RFC7987] Ginsberg, L., Wells, P., Decraene, B., Przygienda, T., and
H. Gredler, "IS-IS Minimum Remaining Lifetime", RFC 7987, H. Gredler, "IS-IS Minimum Remaining Lifetime", RFC 7987,
DOI 10.17487/RFC7987, October 2016, DOI 10.17487/RFC7987, October 2016,
<https://www.rfc-editor.org/info/rfc7987>. <https://www.rfc-editor.org/info/rfc7987>.
[RFC8200] Deering, S. and R. Hinden, "Internet Protocol, Version 6 [RFC8200] Deering, S. and R. Hinden, "Internet Protocol, Version 6
(IPv6) Specification", STD 86, RFC 8200, (IPv6) Specification", STD 86, RFC 8200,
DOI 10.17487/RFC8200, July 2017, DOI 10.17487/RFC8200, July 2017,
<https://www.rfc-editor.org/info/rfc8200>. <https://www.rfc-editor.org/info/rfc8200>.
10.2. Informative References 11.2. Informative References
[CLOS] Yuan, X., "On Nonblocking Folded-Clos Networks in Computer [CLOS] Yuan, X., "On Nonblocking Folded-Clos Networks in Computer
Communication Environments", IEEE International Parallel & Communication Environments", IEEE International Parallel &
Distributed Processing Symposium, 2011. Distributed Processing Symposium, 2011.
[DIJKSTRA] [DIJKSTRA]
Dijkstra, E., "A Note on Two Problems in Connexion with Dijkstra, E., "A Note on Two Problems in Connexion with
Graphs", Journal Numer. Math. , 1959. Graphs", Journal Numer. Math. , 1959.
[DOT] Ellson, J. and L. Koutsofios, "Graphviz: open source graph [DOT] Ellson, J. and L. Koutsofios, "Graphviz: open source graph
drawing tools", Springer-Verlag , 2001. drawing tools", Springer-Verlag , 2001.
[DYNAMO] De Candia et al., G., "Dynamo: amazon's highly available [DYNAMO] De Candia et al., G., "Dynamo: amazon's highly available
key-value store", ACM SIGOPS symposium on Operating key-value store", ACM SIGOPS symposium on Operating
systems principles (SOSP '07), 2007. systems principles (SOSP '07), 2007.
[EPPSTEIN] [EPPSTEIN]
Eppstein, D., "Finding the k-Shortest Paths", 1997. Eppstein, D., "Finding the k-Shortest Paths", 1997.
[EUI64] IEEE, "Guidelines for Use of Extended Unique Identifier
(EUI), Organizationally Unique Identifier (OUI), and
Company ID (CID)", IEEE EUI,
<http://standards.ieee.org/develop/regauth/tut/eui.pdf>.
[FATTREE] Leiserson, C., "Fat-Trees: Universal Networks for [FATTREE] Leiserson, C., "Fat-Trees: Universal Networks for
Hardware-Efficient Supercomputing", 1985. Hardware-Efficient Supercomputing", 1985.
[I-D.ietf-spring-segment-routing] [I-D.ietf-spring-segment-routing]
Filsfils, C., Previdi, S., Ginsberg, L., Decraene, B., Filsfils, C., Previdi, S., Ginsberg, L., Decraene, B.,
Litkowski, S., and R. Shakir, "Segment Routing Litkowski, S., and R. Shakir, "Segment Routing
Architecture", draft-ietf-spring-segment-routing-15 (work Architecture", draft-ietf-spring-segment-routing-15 (work
in progress), January 2018. in progress), January 2018.
[IEEEstd1588] [IEEEstd1588]
skipping to change at page 66, line 9 skipping to change at page 82, line 29
International Organization for Standardization, International Organization for Standardization,
"Intermediate system to Intermediate system intra-domain "Intermediate system to Intermediate system intra-domain
routeing information exchange protocol for use in routeing information exchange protocol for use in
conjunction with the protocol for providing the conjunction with the protocol for providing the
connectionless-mode Network Service (ISO 8473)", Nov 2002. connectionless-mode Network Service (ISO 8473)", Nov 2002.
[MAKSIC2013] [MAKSIC2013]
Maksic et al., N., "Improving Utilization of Data Center Maksic et al., N., "Improving Utilization of Data Center
Networks", IEEE Communications Magazine, Nov 2013. Networks", IEEE Communications Magazine, Nov 2013.
[PROTOBUF]
Google, Inc., "Protocol Buffers,
https://developers.google.com/protocol-buffers".
[QUIC] Iyengar et al., J., "QUIC: A UDP-Based Multiplexed and
Secure Transport", 2016.
[RFC0826] Plummer, D., "An Ethernet Address Resolution Protocol: Or [RFC0826] Plummer, D., "An Ethernet Address Resolution Protocol: Or
Converting Network Protocol Addresses to 48.bit Ethernet Converting Network Protocol Addresses to 48.bit Ethernet
Address for Transmission on Ethernet Hardware", STD 37, Address for Transmission on Ethernet Hardware", STD 37,
RFC 826, DOI 10.17487/RFC0826, November 1982, RFC 826, DOI 10.17487/RFC0826, November 1982,
<https://www.rfc-editor.org/info/rfc826>. <https://www.rfc-editor.org/info/rfc826>.
[RFC2131] Droms, R., "Dynamic Host Configuration Protocol", [RFC2131] Droms, R., "Dynamic Host Configuration Protocol",
RFC 2131, DOI 10.17487/RFC2131, March 1997, RFC 2131, DOI 10.17487/RFC2131, March 1997,
<https://www.rfc-editor.org/info/rfc2131>. <https://www.rfc-editor.org/info/rfc2131>.
skipping to change at page 68, line 26 skipping to change at page 84, line 38
/** @note MUST be interpreted in implementation as unsigned 16 bits */ /** @note MUST be interpreted in implementation as unsigned 16 bits */
typedef i16 LevelType typedef i16 LevelType
/** @note MUST be interpreted in implementation as unsigned 32 bits */ /** @note MUST be interpreted in implementation as unsigned 32 bits */
typedef i32 PodType typedef i32 PodType
/** @note MUST be interpreted in implementation as unsigned 16 bits */ /** @note MUST be interpreted in implementation as unsigned 16 bits */
typedef i16 VersionType typedef i16 VersionType
/** @note MUST be interpreted in implementation as unsigned 32 bits */ /** @note MUST be interpreted in implementation as unsigned 32 bits */
typedef i32 MetricType typedef i32 MetricType
/** @note MUST be interpreted in implementation as unstructured 64 bits */ /** @note MUST be interpreted in implementation as unstructured 64 bits */
typedef i64 RouteTagType typedef i64 RouteTagType
/** @note MUST be interpreted in implementation as unstructured 32 bits /** @note MUST be interpreted in implementation as unstructured 32 bits label value */
label value */
typedef i32 LabelType typedef i32 LabelType
/** @note MUST be interpreted in implementation as unsigned 32 bits */ /** @note MUST be interpreted in implementation as unsigned 32 bits */
typedef i32 BandwithInMegaBitsType typedef i32 BandwithInMegaBitsType
typedef string KeyIDType typedef string KeyIDType
/** node local, unique identification for a link (interface/tunnel /** node local, unique identification for a link (interface/tunnel
* etc. Basically anything RIFT runs on). This is kept * etc. Basically anything RIFT runs on). This is kept
* at 32 bits so it aligns with BFD [RFC5880] discriminator size. * at 32 bits so it aligns with BFD [RFC5880] discriminator size.
*/ */
typedef i32 LinkIDType typedef i32 LinkIDType
typedef string KeyNameType typedef string KeyNameType
typedef i8 PrefixLenType typedef i8 PrefixLenType
/** timestamp in seconds since the epoch */ /** timestamp in seconds since the epoch */
typedef i64 TimestampInSecsType typedef i64 TimestampInSecsType
/** security nonce */ /** security nonce */
typedef i64 NonceType typedef i64 NonceType
/** adjacency holdtime */ /** LIE FSM holdtime type */
typedef i16 HoldTimeInSecType typedef i16 TimeIntervalInSecType
/** Transaction ID type for prefix mobility as specified by RFC6550, value /** Transaction ID type for prefix mobility as specified by RFC6550, value
MUST be interpreted in implementation as unsigned */ MUST be interpreted in implementation as unsigned */
typedef i8 PrefixTransactionIDType typedef i8 PrefixTransactionIDType
/** timestamp per IEEE 802.1AS, values MUST be interpreted in /** timestamp per IEEE 802.1AS, values MUST be interpreted in implementation as unsigned */
implementation as unsigned */
struct IEEE802_1ASTimeStampType { struct IEEE802_1ASTimeStampType {
1: required i64 AS_sec; 1: required i64 AS_sec;
2: optional i32 AS_nsec; 2: optional i32 AS_nsec;
} }
/** Flags indicating nodes behavior in case of ZTP and support /** Flags indicating nodes behavior in case of ZTP and support
for special optimization procedures. It will force level to `leaf_level` for special optimization procedures. It will force level to `leaf_level` or
`top-of-fabric` level accordingly and enable according procedures
*/ */
enum LeafIndications { enum HierarchyIndications {
leaf_only =0, leaf_only = 0,
leaf_only_and_leaf_2_leaf_procedures =1, leaf_only_and_leaf_2_leaf_procedures = 1,
top_of_fabric = 2,
} }
/** This MUST be used when node is configured as top of fabric in ZTP.
This is kept reasonably low to alow for fast ZTP convergence on
failures. */
const LevelType top_of_fabric_level = 24
/** default bandwidth on a link */ /** default bandwidth on a link */
const BandwithInMegaBitsType default_bandwidth = 100 const BandwithInMegaBitsType default_bandwidth = 100
/** fixed leaf level when ZTP is not used */ /** fixed leaf level when ZTP is not used */
const LevelType leaf_level = 0 const LevelType leaf_level = 0
const LevelType default_level = leaf_level const LevelType default_level = leaf_level
/** This MUST be used when node is configured as superspine in ZTP. const PodType default_pod = 0
This is kept reasonably low to alow for fast ZTP convergence on const LinkIDType undefined_linkid = 0
failures. */
const LevelType default_superspine_level = 24
const PodType default_pod = 0
const LinkIDType undefined_linkid = 0
/** default distance used */ /** default distance used */
const MetricType default_distance = 1 const MetricType default_distance = 1
/** any distance larger than this will be considered infinity */ /** any distance larger than this will be considered infinity */
const MetricType infinite_distance = 0x7FFFFFFF const MetricType infinite_distance = 0x7FFFFFFF
/** any element with 0 distance will be ignored, /** represents invalid distance */
* missing metrics will be replaced with default_distance
*/
const MetricType invalid_distance = 0 const MetricType invalid_distance = 0
const bool overload_default = false const bool overload_default = false
const bool flood_reduction_default = true const bool flood_reduction_default = true
const HoldTimeInSecType default_holdtime = 3 /** default LIE FSM holddown time */
const TimeIntervalInSecType default_lie_holdtime = 3
/** default ZTP FSM holddown time */
const TimeIntervalInSecType default_ztp_holdtime = 1
/** by default LIE levels are ZTP offers */ /** by default LIE levels are ZTP offers */
const bool default_not_a_ztp_offer = false const bool default_not_a_ztp_offer = false
/** by default e'one is repeating flooding */ /** by default e'one is repeating flooding */
const bool default_you_are_not_flood_repeater = false const bool default_you_are_flood_repeater = true
/** 0 is illegal for SystemID */ /** 0 is illegal for SystemID */
const SystemIDType IllegalSystemID = 0 const SystemIDType IllegalSystemID = 0
/** empty set of nodes */ /** empty set of nodes */
const set<SystemIDType> empty_set_of_nodeids = {} const set<SystemIDType> empty_set_of_nodeids = {}
/** default lifetime is one week */
const LifeTimeInSecType default_lifetime = 604800
/** any `TieHeader` that has a smaller lifetime difference
than this constant is equal (if other fields equal) */
const LifeTimeInSecType lifetime_diff2ignore = 300
/** default UDP port to run LIEs on */ /** default UDP port to run LIEs on */
const UDPPortType default_lie_udp_port = 911 const UDPPortType default_lie_udp_port = 911
/** default UDP port to receive TIEs on, that can be peer specific */ /** default UDP port to receive TIEs on, that can be peer specific */
const UDPPortType default_tie_udp_flood_port = 912 const UDPPortType default_tie_udp_flood_port = 912
/** default MTU link size to use */ /** default MTU link size to use */
const MTUSizeType default_mtu_size = 1400 const MTUSizeType default_mtu_size = 1400
/** indicates whether the direction is northbound/east-west /** indicates whether the direction is northbound/east-west
skipping to change at page 71, line 4 skipping to change at page 87, line 22
} }
/** @note: Sequence of a prefix. Comparison function: /** @note: Sequence of a prefix. Comparison function:
if diff(timestamps) < 200msecs better transactionid wins if diff(timestamps) < 200msecs better transactionid wins
else better time wins else better time wins
*/ */
struct PrefixSequenceType { struct PrefixSequenceType {
1: required IEEE802_1ASTimeStampType timestamp; 1: required IEEE802_1ASTimeStampType timestamp;
2: optional PrefixTransactionIDType transactionid; 2: optional PrefixTransactionIDType transactionid;
} }
enum TIETypeType { enum TIETypeType {
Illegal = 0, Illegal = 0,
TIETypeMinValue = 1, TIETypeMinValue = 1,
/** first legal value */ /** first legal value */
NodeTIEType = 2, NodeTIEType = 2,
PrefixTIEType = 3, PrefixTIEType = 3,
TransitivePrefixTIEType = 4, PositiveDisaggregationPrefixTIEType = 4,
PGPrefixTIEType = 5, NegativeDisaggregationPrefixTIEType = 5,
KeyValueTIEType = 6, PGPrefixTIEType = 6,
TIETypeMaxValue = 7, KeyValueTIEType = 7,
ExternalPrefixTIEType = 8,
TIETypeMaxValue = 9,
} }
/** @note: route types which MUST be ordered on their preference /** @note: route types which MUST be ordered on their preference
* PGP prefixes are most preferred attracting * PGP prefixes are most preferred attracting
* traffic north (towards spine) and then south * traffic north (towards spine) and then south
* normal prefixes are attracting traffic south (towards leafs), * normal prefixes are attracting traffic south (towards leafs),
* i.e. prefix in NORTH PREFIX TIE is preferred over SOUTH PREFIX TIE * i.e. prefix in NORTH PREFIX TIE is preferred over SOUTH PREFIX TIE
* *
* @todo: external routes * @todo: external routes
* @note: The only purpose of those values is to introduce an
* ordering whereas an implementation can choose internally
* any other values as long the ordering is preserved
*/ */
enum RouteType { enum RouteType {
Illegal = 0, Illegal = 0,
RouteTypeMinValue = 1, RouteTypeMinValue = 1,
/** First legal value. */ /** First legal value. */
/** Discard routes are most prefered */ /** Discard routes are most prefered */
Discard = 2, Discard = 2,
/** Local prefixes are directly attached prefixes on the /** Local prefixes are directly attached prefixes on the
* system such as e.g. interface routes. * system such as e.g. interface routes.
*/ */
LocalPrefix = 3, LocalPrefix = 3,
/** advertised in S-TIEs */ /** advertised in S-TIEs */
SouthPGPPrefix = 4, SouthPGPPrefix = 4,
/** advertised in N-TIEs */ /** advertised in N-TIEs */
NorthPGPPrefix = 5, NorthPGPPrefix = 5,
/** advertised in N-TIEs */ /** advertised in N-TIEs */
NorthPrefix = 6, NorthPrefix = 6,
/** advertised in S-TIEs */ /** advertised in S-TIEs, either normal prefix or positive disaggregation */
SouthPrefix = 7, SouthPrefix = 7,
/** transitive southbound are least preferred */ /** externally imported north */
TransitiveSouthPrefix = 8, NorthExternalPrefix = 8,
RouteTypeMaxValue = 9 /** externally imported south */
SouthExternalPrefix = 9,
/** negative, transitive are least preferred of
local variety */
NegativeSouthPrefix = 10,
RouteTypeMaxValue = 11,
} }
A.2. encoding.thrift A.2. encoding.thrift
/** /**
Thrift file for packet encodings for RIFT Thrift file for packet encodings for RIFT
*/ */
include "common.thrift" include "common.thrift"
/** represents protocol encoding schema major version */ /** represents protocol encoding schema major version */
const i32 protocol_major_version = 11 const i32 protocol_major_version = 19
/** represents protocol encoding schema minor version */ /** represents protocol encoding schema minor version */
const i32 protocol_minor_version = 0 const i32 protocol_minor_version = 0
/** common RIFT packet header */ /** common RIFT packet header */
struct PacketHeader { struct PacketHeader {
1: required common.VersionType major_version = protocol_major_version; 1: required common.VersionType major_version = protocol_major_version;
2: required common.VersionType minor_version = protocol_minor_version; 2: required common.VersionType minor_version = protocol_minor_version;
/** this is the node sending the packet, in case of LIE/TIRE/TIDE /** this is the node sending the packet, in case of LIE/TIRE/TIDE
also the originator of it */ also the originator of it */
3: required common.SystemIDType sender; 3: required common.SystemIDType sender;
skipping to change at page 72, line 46 skipping to change at page 89, line 24
/** Neighbor structure */ /** Neighbor structure */
struct Neighbor { struct Neighbor {
1: required common.SystemIDType originator; 1: required common.SystemIDType originator;
2: required common.LinkIDType remote_id; 2: required common.LinkIDType remote_id;
} }
/** Capabilities the node supports */ /** Capabilities the node supports */
struct NodeCapabilities { struct NodeCapabilities {
/** can this node participate in flood reduction */ /** can this node participate in flood reduction */
1: optional bool flood_reduction = 1: optional bool flood_reduction =
common.flood_reduction_default; common.flood_reduction_default;
/** does this node restrict itself to be leaf only (in ZTP) and /** does this node restrict itself to be top-of-fabric or
does it support leaf-2-leaf procedures */ leaf only (in ZTP) and does it support leaf-2-leaf procedures */
2: optional common.HierarchyIndications hierarchy_indications;
2: optional common.LeafIndications leaf_indications;
} }
/** RIFT LIE packet /** RIFT LIE packet
@note this node's level is already included on the packet header */ @note this node's level is already included on the packet header */
struct LIEPacket { struct LIEPacket {
/** optional node or adjacency name */ /** optional node or adjacency name */
1: optional string name; 1: optional string name;
/** local link ID */ /** local link ID */
2: required common.LinkIDType local_id; 2: required common.LinkIDType local_id;
/** UDP port to which we can receive flooded TIEs */ /** UDP port to which we can receive flooded TIEs */
3: required common.UDPPortType flood_port = 3: required common.UDPPortType flood_port =
common.default_tie_udp_flood_port; common.default_tie_udp_flood_port;
/** layer 3 MTU, used to discover to mismatch */ /** layer 3 MTU, used to discover to mismatch */
4: optional common.MTUSizeType link_mtu_size = 4: optional common.MTUSizeType link_mtu_size =
common.default_mtu_size; common.default_mtu_size;
/** local link bandwidth on the interface */
5: optional common.BandwithInMegaBitsType link_bandwidth =
common.default_bandwidth;
/** this will reflect the neighbor once received to provid /** this will reflect the neighbor once received to provid
3-way connectivity */ 3-way connectivity */
5: optional Neighbor neighbor; 6: optional Neighbor neighbor;
6: optional common.PodType pod = common.default_pod; 7: optional common.PodType pod = common.default_pod;
/** optional nonce used for security computations */ /** optional local nonce used for security computations */
7: optional common.NonceType nonce; 8: optional common.NonceType nonce;
/** optional neighbor's reflected nonce for security purposes. Significant delta
in nonces seen compared to current local nonce can be used to prevent replays */
9: optional common.NonceType last_neighbor_nonce;
/** optional node capabilities shown in the LIE. The capabilies /** optional node capabilities shown in the LIE. The capabilies
MUST match the capabilities shown in the Node TIEs, otherwise MUST match the capabilities shown in the Node TIEs, otherwise
the behavior is unspecified. A node detecting the mismatch the behavior is unspecified. A node detecting the mismatch
SHOULD generate according error. SHOULD generate according error.
*/ */
8: optional NodeCapabilities capabilities; 10: optional NodeCapabilities capabilities;
/** required holdtime of the adjacency, i.e. how much time /** required holdtime of the adjacency, i.e. how much time
MUST expire without LIE for the adjacency to drop MUST expire without LIE for the adjacency to drop
*/ */
9: required common.HoldTimeInSecType holdtime = 11: required common.TimeIntervalInSecType holdtime =
common.default_holdtime; common.default_lie_holdtime;
/** indicates that the level on the LIE MUST NOT be used /** indicates that the level on the LIE MUST NOT be used
to derive a ZTP level by the receiving node. */ to derive a ZTP level by the receiving node. */
10: optional bool not_a_ztp_offer = 12: optional bool not_a_ztp_offer =
common.default_not_a_ztp_offer; common.default_not_a_ztp_offer;
/** indicates to northbound neighbor that it should not /** indicates to northbound neighbor that it should
be reflooding this node's N-TIEs to flood reduce and be reflooding this node's N-TIEs to achieve flood reducuction and
balance northbound flooding. To be ignored if received from a balancing for northbound flooding. To be ignored if received from a
northbound adjacency. */ northbound adjacency. */
11: optional bool you_are_not_flood_repeater= 13: optional bool you_are_flood_repeater =
common.default_you_are_not_flood_repeater; common.default_you_are_flood_repeater;
/** optional downstream assigned locally significant label /** optional downstream assigned locally significant label
value for the adjacency. */ value for the adjacency. */
12: optional common.LabelType label; 14: optional common.LabelType label;
} }
/** LinkID pair describes one of parallel links between two nodes */ /** LinkID pair describes one of parallel links between two nodes */
struct LinkIDPair { struct LinkIDPair {
/** node-wide unique value for the local link */ /** node-wide unique value for the local link */
1: required common.LinkIDType local_id; 1: required common.LinkIDType local_id;
/** received remote link ID for this link */ /** received remote link ID for this link */
2: required common.LinkIDType remote_id; 2: required common.LinkIDType remote_id;
/** more properties of the link can go in here */ /** more properties of the link can go in here */
} }
/** ID of a TIE /** ID of a TIE
skipping to change at page 74, line 17 skipping to change at page 90, line 48
1: required common.LinkIDType local_id; 1: required common.LinkIDType local_id;
/** received remote link ID for this link */ /** received remote link ID for this link */
2: required common.LinkIDType remote_id; 2: required common.LinkIDType remote_id;
/** more properties of the link can go in here */ /** more properties of the link can go in here */
} }
/** ID of a TIE /** ID of a TIE
@note: TIEID space is a total order achieved by comparing the elements @note: TIEID space is a total order achieved by comparing the elements
in sequence defined and comparing each value as an in sequence defined and comparing each value as an
unsigned integer of according length unsigned integer of according length.
*/ */
struct TIEID { struct TIEID {
/** indicates direction of the TIE */ /** indicates direction of the TIE */
1: required common.TieDirectionType direction; 1: required common.TieDirectionType direction;
/** indicates originator of the TIE */ /** indicates originator of the TIE */
2: required common.SystemIDType originator; 2: required common.SystemIDType originator;
3: required common.TIETypeType tietype; 3: required common.TIETypeType tietype;
4: required common.TIENrType tie_nr; 4: required common.TIENrType tie_nr;
} }
/** Header of a TIE */ /** Header of a TIE.
@note: TIEID space is a total order achieved by comparing the elements
in sequence defined and comparing each value as an
unsigned integer of according length. `origination_time` is
disregarded for comparison purposes.
*/
struct TIEHeader { struct TIEHeader {
2: required TIEID tieid; 2: required TIEID tieid;
3: required common.SeqNrType seq_nr; 3: required common.SeqNrType seq_nr;
/** lifetime expires down to 0 just like in ISIS */ /** remaining lifetime that expires down to 0 just like in ISIS.
4: required common.LifeTimeInSecType lifetime; TIEs with lifetimes differing by less than `lifetime_diff2ignore` MUST
be considered EQUAL. */
4: required common.LifeTimeInSecType remaining_lifetime;
/** optional absolute timestamp when the TIE
was generated. This can be used on fabrics with
synchronized clock to prevent lifetime modification attacks. */
10: optional common.IEEE802_1ASTimeStampType origination_time;
/** optional original lifetime when the TIE
was generated. This can be used on fabrics with
synchronized clock to prevent lifetime modification attacks. */
12: optional common.LifeTimeInSecType origination_lifetime;
} }
/** A sorted TIDE packet, if unsorted, behavior is undefined */ /** A TIDE with sorted TIE headers, if headers unsorted, behavior is undefined */
struct TIDEPacket { struct TIDEPacket {
/** all 00s marks starts */ /** all 00s marks starts */
1: required TIEID start_range; 1: required TIEID start_range;
/** all FFs mark end */ /** all FFs mark end */
2: required TIEID end_range; 2: required TIEID end_range;
/** _sorted_ list of headers */ /** _sorted_ list of headers */
3: required list<TIEHeader> headers; 3: required list<TIEHeader> headers;
} }
/** A TIRE packet */ /** A TIRE packet */
skipping to change at page 76, line 6 skipping to change at page 93, line 5
1: required common.LevelType level; 1: required common.LevelType level;
/** if neighbor systemID repeats in other node TIEs of same node /** if neighbor systemID repeats in other node TIEs of same node
the behavior is undefined. Equivalent to |A_(n,s)(N) in spec. */ the behavior is undefined. Equivalent to |A_(n,s)(N) in spec. */
2: required map<common.SystemIDType, 2: required map<common.SystemIDType,
NodeNeighborsTIEElement> neighbors; NodeNeighborsTIEElement> neighbors;
3: optional NodeCapabilities capabilities; 3: optional NodeCapabilities capabilities;
4: optional NodeFlags flags; 4: optional NodeFlags flags;
/** optional node name for easier operations */ /** optional node name for easier operations */
5: optional string name; 5: optional string name;
/** Nodes seen an the same level through reflection through nodes
having backlink to both nodes. They are equivalent to |V(N) in
future specifications. Ignored in Node S-TIEs if present.
*/
6: optional set<common.SystemIDType> visible_in_same_level
= common.empty_set_of_nodeids;
/** Non-overloaded nodes in |V seen as attached to another north
* level partition due to the fact that some nodes in its |V have
* adjacencies to higher level nodes that this node doesn't see.
* This may be used in the computation at higher levels to prevent
* blackholing. Ignored in Node S-TIEs if present.
* Equivalent to |PUL(N) in spec. */
7: optional set<common.SystemIDType> same_level_unknown_north_partitions
= common.empty_set_of_nodeids;
} }
struct PrefixAttributes { struct PrefixAttributes {
2: required common.MetricType metric = common.default_distance; 2: required common.MetricType metric = common.default_distance;
/** generic unordered set of route tags, can be redistributed to /** generic unordered set of route tags, can be redistributed to other protocols or use
other protocols or use
within the context of real time analytics */ within the context of real time analytics */
3: optional set<common.RouteTagType> tags; 3: optional set<common.RouteTagType> tags;
/** optional monotonic clock for mobile addresses */ /** optional monotonic clock for mobile addresses */
4: optional common.PrefixSequenceType monotonic_clock; 4: optional common.PrefixSequenceType monotonic_clock;
} }
/** multiple prefixes */ /** multiple prefixes */
struct PrefixTIEElement { struct PrefixTIEElement {
/** prefixes with the associated attributes. /** prefixes with the associated attributes.
if the same prefix repeats in multiple TIEs of same node if the same prefix repeats in multiple TIEs of same node
skipping to change at page 76, line 50 skipping to change at page 93, line 34
/** keys with their values */ /** keys with their values */
struct KeyValueTIEElement { struct KeyValueTIEElement {
/** if the same key repeats in multiple TIEs of same node /** if the same key repeats in multiple TIEs of same node
or with different values, behavior is unspecified */ or with different values, behavior is unspecified */
1: required map<common.KeyIDType,string> keyvalues; 1: required map<common.KeyIDType,string> keyvalues;
} }
/** single element in a TIE. enum common.TIETypeType /** single element in a TIE. enum common.TIETypeType
in TIEID indicates which elements MUST be present in TIEID indicates which elements MUST be present
in the TIEElement. In case of mismatch the unexpected in the TIEElement. In case of mismatch the unexpected
elements MUST be ignored. elements MUST be ignored. In case of lack of expected
element the TIE an error MUST be reported and the TIE
MUST be ignored.
*/ */
union TIEElement { union TIEElement {
/** in case of enum common.TIETypeType.NodeTIEType */ /** in case of enum common.TIETypeType.NodeTIEType */
1: optional NodeTIEElement node; 1: optional NodeTIEElement node;
/** in case of enum common.TIETypeType.PrefixTIEType */ /** in case of enum common.TIETypeType.PrefixTIEType */
2: optional PrefixTIEElement prefixes; 2: optional PrefixTIEElement prefixes;
/** transitive prefixes (always southbound) which SHOULD be propagated /** positive prefixes (always southbound)
* southwards towards lower levels to heal It MUST NOT be advertised within a North TIE.
* pathological upper level partitioning, otherwise */
* blackholes may occur. MUST NOT be advertised within a North TIE. 3: optional PrefixTIEElement positive_disaggregation_prefixes;
*/ /** transitive, negative prefixes (always southbound) which
3: optional PrefixTIEElement transitive_prefixes; MUST be aggregated and propagated
4: optional KeyValueTIEElement keyvalues; according to the specification
southwards towards lower levels to heal
pathological upper level partitioning, otherwise
blackholes may occur in multiplane fabrics.
It MUST NOT be advertised within a North TIE.
*/
4: optional PrefixTIEElement negative_disaggregation_prefixes;
/** externally reimported prefixes */
5: optional PrefixTIEElement external_prefixes;
/** Key-Value store elements */
6: optional KeyValueTIEElement keyvalues;
/** @todo: policy guided prefixes */ /** @todo: policy guided prefixes */
} }
/** @todo: flood header separately in UDP to allow changing lifetime and SHA /** @todo: flood header separately in UDP to allow changing lifetime and SHA without reserialization
without reserialization
*/ */
struct TIEPacket { struct TIEPacket {
1: required TIEHeader header; 1: required TIEHeader header;
2: required TIEElement element; 2: required TIEElement element;
} }
union PacketContent { union PacketContent {
1: optional LIEPacket lie; 1: optional LIEPacket lie;
2: optional TIDEPacket tide; 2: optional TIDEPacket tide;
3: optional TIREPacket tire; 3: optional TIREPacket tire;
4: optional TIEPacket tie; 4: optional TIEPacket tie;
} }
/** protocol packet structure */ /** protocol packet structure */
struct ProtocolPacket { struct ProtocolPacket {
1: required PacketHeader header; 1: required PacketHeader header;
2: required PacketContent content; 2: required PacketContent content;
} }
Appendix B. Finite State Machines Appendix B. Finite State Machines and Precise Operational
Specifications
All FSM figures are provided as [DOT] description due to limiations Some FSM figures are provided as [DOT] description due to limitations
of ASCII art. of ASCII art.
B.1. LIE On Entry action is performed every time and right before the
according state is entered, i.e. after any transitions from previous
state.
On Exit action is performed every time and immediately when a state
is exited, i.e. before any transitions towards target state are
performed.
Any attempt to transition from a state towards another on reception
of an event where no action is specified MUST be considered an
unrecoverable error.
The FSMs and procedures are NOT normative in the sense that an
implementation MUST implement them literally (which would be
overspecification) but an implementation MUST exhibit externally
observable behavior that is identical to the execution of the
specified FSMs.
Where a FSM representation is inconvenient, i.e. the amount of
procedures and kept state exceeds the amount of transitions, we defer
to a more procedural description on data structures.
B.1. LIE FSM
Initial state is `OneWay`.
Event `MultipleNeighbors` occurs normally when more than two nodes
see each other on the same link or a remote node is quickly
reconfigured or rebooted without regressing to `OneWay` first. Each
occurence of the event SHOULD generate a clear, according
notification to help operational deployments.
The machine sends LIEs on several transitions to accelerate adjacency
bring-up without waiting for the timer tic.
digraph Ga556dde74c30450aae125eaebc33bd57 {
Nd16ab5092c6b421c88da482eb4ae36b6[label="ThreeWay"][shape="oval"];
N54edd2b9de7641688608f44fca346303[label="OneWay"][shape="oval"];
Nfeef2e6859ae4567bd7613a32cc28c0e[label="TwoWay"][shape="oval"];
N7f2bb2e04270458cb5c9bb56c4b96e23[label="Enter"][style="invis"][shape="plain"];
N292744a4097f492f8605c926b924616b[label="Enter"][style="dashed"][shape="plain"];
Nc48847ba98e348efb45f5b78f4a5c987[label="Exit"][style="invis"][shape="plain"];
Nd16ab5092c6b421c88da482eb4ae36b6 -> N54edd2b9de7641688608f44fca346303
[label="|NeighborChangedLevel|\n|NeighborChangedAddress|\n|UnacceptableHeader|\n|MTUMismatch|\n|PODMismatch|\n|HoldtimeExpired|\n|MultipleNeighbors|"]
[color="black"][arrowhead="normal" dir="both" arrowtail="none"];
Nd16ab5092c6b421c88da482eb4ae36b6 -> Nd16ab5092c6b421c88da482eb4ae36b6
[label="|TimerTick|\n|LieRcvd|\n|SendLie|"][color="black"]
[arrowhead="normal" dir="both" arrowtail="none"];
Nfeef2e6859ae4567bd7613a32cc28c0e -> Nfeef2e6859ae4567bd7613a32cc28c0e
[label="|TimerTick|\n|LieRcvd|\n|SendLie|"][color="black"]
[arrowhead="normal" dir="both" arrowtail="none"];
N54edd2b9de7641688608f44fca346303 -> Nd16ab5092c6b421c88da482eb4ae36b6
[label="|ValidReflection|"][color="red"][arrowhead="normal" dir="both" arrowtail="none"];
Nd16ab5092c6b421c88da482eb4ae36b6 -> Nd16ab5092c6b421c88da482eb4ae36b6
[label="|HALChanged|\n|HATChanged|\n|HALSChanged|\n|UpdateZTPOffer|"][color="blue"]
[arrowhead="normal" dir="both" arrowtail="none"];
Nd16ab5092c6b421c88da482eb4ae36b6 -> Nd16ab5092c6b421c88da482eb4ae36b6
[label="|ValidReflection|"][color="red"][arrowhead="normal" dir="both" arrowtail="none"];
Nfeef2e6859ae4567bd7613a32cc28c0e -> N54edd2b9de7641688608f44fca346303
[label="|LevelChanged|"][color="blue"][arrowhead="normal" dir="both" arrowtail="none"];
Nfeef2e6859ae4567bd7613a32cc28c0e -> N54edd2b9de7641688608f44fca346303
[label="|NeighborChangedLevel|\n|NeighborChangedAddress|\n|UnacceptableHeader|\n|MTUMismatch|\n|PODMismatch|\n|HoldtimeExpired|\n|MultipleNeighbors|"]
[color="black"][arrowhead="normal" dir="both" arrowtail="none"];
Nfeef2e6859ae4567bd7613a32cc28c0e -> Nd16ab5092c6b421c88da482eb4ae36b6
[label="|ValidReflection|"][color="red"][arrowhead="normal" dir="both" arrowtail="none"];
N54edd2b9de7641688608f44fca346303 -> N54edd2b9de7641688608f44fca346303
[label="|TimerTick|\n|LieRcvd|\n|NeighborChangedLevel|\n|NeighborChangedAddress|\n|UnacceptableHeader|\n|MTUMismatch|\n|PODMismatch|\n|HoldtimeExpired|\n|SendLie|"]
[color="black"][arrowhead="normal" dir="both" arrowtail="none"];
N292744a4097f492f8605c926b924616b -> N54edd2b9de7641688608f44fca346303
[label=""][color="black"][arrowhead="normal" dir="both" arrowtail="none"];
Nd16ab5092c6b421c88da482eb4ae36b6 -> N54edd2b9de7641688608f44fca346303
[label="|LevelChanged|"][color="blue"][arrowhead="normal" dir="both" arrowtail="none"];
N54edd2b9de7641688608f44fca346303 -> Nfeef2e6859ae4567bd7613a32cc28c0e
[label="|NewNeighbor|"][color="black"][arrowhead="normal" dir="both" arrowtail="none"];
N54edd2b9de7641688608f44fca346303 -> N54edd2b9de7641688608f44fca346303
[label="|LevelChanged|\n|HALChanged|\n|HATChanged|\n|HALSChanged|\n|UpdateZTPOffer|"]
[color="blue"][arrowhead="normal" dir="both" arrowtail="none"];
Nfeef2e6859ae4567bd7613a32cc28c0e -> Nfeef2e6859ae4567bd7613a32cc28c0e
[label="|HALChanged|\n|HATChanged|\n|HALSChanged|\n|UpdateZTPOffer|"]
[color="blue"][arrowhead="normal" dir="both" arrowtail="none"];
Nd16ab5092c6b421c88da482eb4ae36b6 -> Nfeef2e6859ae4567bd7613a32cc28c0e
[label="|NeighborDroppedReflection|"]
[color="red"][arrowhead="normal" dir="both" arrowtail="none"];
N54edd2b9de7641688608f44fca346303 -> N54edd2b9de7641688608f44fca346303
[label="|NeighborDroppedReflection|"][color="red"]
[arrowhead="normal" dir="both" arrowtail="none"];
}
digraph G791bb566f5cf48b09e26193a727dadfd {
N91ea7c47496746d880c10a5def7874c2[label="TwoWay"][shape="oval"];
Nc5d62000e5dc45a9ac1379c28cfda9b3[label="OneWay"][shape="oval"];
Nd7b87acca28f4613a68bbc4ef79a3c50[label="Enter"][style="dashed"]
[shape="plain"];
Ne0fb2564cd334a44ad080f73b07cca86[label="ThreeWay"][shape="oval"];
N51443826b9c84d8b83cc252b471047c9[label="Enter"][style="invis"]
[shape="plain"];
N19343f3f3a9b41c29f3ac23c8dccc179[label="Exit"][style="invis"]
[shape="plain"];
N91ea7c47496746d880c10a5def7874c2 -> Nc5d62000e5dc45a9ac1379c28cfda9b3
[label="|LevelChanged|"][color="blue"]
[arrowhead="normal" dir="both" arrowtail="none"];
Ne0fb2564cd334a44ad080f73b07cca86 -> Ne0fb2564cd334a44ad080f73b07cca86
[label="|HALChanged|\n|HATChanged|\n|HALSChanged|\n|UpdateZTPOffer|"]
[color="blue"][arrowhead="normal" dir="both" arrowtail="none"];
Nc5d62000e5dc45a9ac1379c28cfda9b3 -> N91ea7c47496746d880c10a5def7874c2
[label="|NewNeighbor|"][color="black"]
[arrowhead="normal" dir="both" arrowtail="none"];
Ne0fb2564cd334a44ad080f73b07cca86 -> N91ea7c47496746d880c10a5def7874c2
[label="|NeighborDroppedReflection|"][color="red"]
[arrowhead="normal" dir="both" arrowtail="none"];
Nc5d62000e5dc45a9ac1379c28cfda9b3 -> Nc5d62000e5dc45a9ac1379c28cfda9b3
[label="|TimerTick|\n|LieRcvd|\n|UnacceptableHeader|\n|HoldtimeExpired|\n|SendLie|"]
[color="black"][arrowhead="normal" dir="both" arrowtail="none"];
N91ea7c47496746d880c10a5def7874c2 -> Ne0fb2564cd334a44ad080f73b07cca86
[label="|ValidReflection|"][color="red"]
[arrowhead="normal" dir="both" arrowtail="none"];
Nd7b87acca28f4613a68bbc4ef79a3c50 -> Nc5d62000e5dc45a9ac1379c28cfda9b3
[label=""]
[color="black"][arrowhead="normal" dir="both" arrowtail="none"];
N91ea7c47496746d880c10a5def7874c2 -> N91ea7c47496746d880c10a5def7874c2
[label="|HALChanged|\n|HATChanged|\n|HALSChanged|\n|UpdateZTPOffer|"]
[color="blue"][arrowhead="normal" dir="both" arrowtail="none"];
Nc5d62000e5dc45a9ac1379c28cfda9b3 -> Nc5d62000e5dc45a9ac1379c28cfda9b3
[label="|LevelChanged|\n|HALChanged|\n|HATChanged|\n|HALSChanged|\n|UpdateZTPOffer|"]
[color="blue"][arrowhead="normal" dir="both" arrowtail="none"];
N91ea7c47496746d880c10a5def7874c2 -> N91ea7c47496746d880c10a5def7874c2
[label="|TimerTick|\n|LieRcvd|\n|SendLie|"][color="black"]
[arrowhead="normal" dir="both" arrowtail="none"];
Ne0fb2564cd334a44ad080f73b07cca86 -> Nc5d62000e5dc45a9ac1379c28cfda9b3
[label="|NeighborChangedLevel|\n|NeighborChangedAddress|\n|UnacceptableHeader|\n|HoldtimeExpired|\n|MultipleNeighbors|"]
[color="black"][arrowhead="normal" dir="both" arrowtail="none"];
Ne0fb2564cd334a44ad080f73b07cca86 -> Nc5d62000e5dc45a9ac1379c28cfda9b3
[label="|LevelChanged|"][color="blue"]
[arrowhead="normal" dir="both" arrowtail="none"];
Ne0fb2564cd334a44ad080f73b07cca86 -> Ne0fb2564cd334a44ad080f73b07cca86
[label="|TimerTick|\n|LieRcvd|\n|SendLie|"][color="black"]
[arrowhead="normal" dir="both" arrowtail="none"];
N91ea7c47496746d880c10a5def7874c2 -> Nc5d62000e5dc45a9ac1379c28cfda9b3
[label="|NeighborChangedLevel|\n|NeighborChangedAddress|\n|UnacceptableHeader|\n|HoldtimeExpired|\n|MultipleNeighbors|"]
[color="black"][arrowhead="normal" dir="both" arrowtail="none"];
}
LIE FSM DOT LIE FSM DOT
.. To be updated ..
LIE FSM Figure
Events Events
o TimerTick: one second timer tic o TimerTick: one second timer tic
o LevelChanged: node's level has been changed by ZTP or o LevelChanged: node's level has been changed by ZTP or
configuration configuration
o HALChanged: best HAL computed by ZTP has changed o HALChanged: best HAL computed by ZTP has changed
o HATChanged: HAT computed by ZTP has changed o HATChanged: HAT computed by ZTP has changed
skipping to change at page 79, line 35 skipping to change at page 97, line 22
o NeighborDroppedReflection: lost previous own reflection from o NeighborDroppedReflection: lost previous own reflection from
neighbor neighbor
o NeighborChangedLevel: neighbor changed advertised level o NeighborChangedLevel: neighbor changed advertised level
o NeighborChangedAddress: neighbor changed IP address o NeighborChangedAddress: neighbor changed IP address
o UnacceptableHeader: unacceptable header seen o UnacceptableHeader: unacceptable header seen
o MTUMismatch: MTU mismatched
o PODMismatch: Unacceptable PoD seen
o HoldtimeExpired: adjacency hold down expired o HoldtimeExpired: adjacency hold down expired
o MultipleNeighbors: more than one neighbor seen on interface o MultipleNeighbors: more than one neighbor seen on interface
o LIECorrupt: corrupted LIE seen
o SendLie: send a LIE out o SendLie: send a LIE out
o UpdateZTPOffer: update this node's ZTP offer o UpdateZTPOffer: update this node's ZTP offer
Actions Actions
on UpdateZTPOffer in TwoWay finishes in TwoWay: send offer to ZTP on TimerTick in TwoWay finishes in TwoWay: PUSH SendLie event, if
FSM holdtime expired PUSH HoldtimeExpired event
on HALChanged in TwoWay finishes in TwoWay: store new HAL
on MTUMismatch in ThreeWay finishes in OneWay: no action
on HALChanged in OneWay finishes in OneWay: store new HAL
on HALChanged in ThreeWay finishes in ThreeWay: store new HAL on HALChanged in ThreeWay finishes in ThreeWay: store new HAL
on HoldtimeExpired in OneWay finishes in OneWay: no action on ValidReflection in TwoWay finishes in ThreeWay: no action
on UpdateZTPOffer in OneWay finishes in OneWay: send offer to ZTP on ValidReflection in OneWay finishes in ThreeWay: no action
FSM
on LevelChanged in ThreeWay finishes in OneWay: update level with on NeighborDroppedReflection in ThreeWay finishes in TwoWay: no
event value action
on LieRcvd in ThreeWay finishes in ThreeWay: PROCESS_LIE
on MultipleNeighbors in TwoWay finishes in OneWay: no action on MultipleNeighbors in TwoWay finishes in OneWay: no action
on NeighborChangedLevel in TwoWay finishes in OneWay: no action on UnacceptableHeader in ThreeWay finishes in OneWay: no action
on HATChanged in OneWay finishes in OneWay: store HAT on MTUMismatch in TwoWay finishes in OneWay: no action
on HATChanged in ThreeWay finishes in ThreeWay: store HAT on LevelChanged in OneWay finishes in OneWay: update level with
event value, PUSH SendLie event
on MultipleNeighbors in ThreeWay finishes in OneWay: no action on UnacceptableHeader in TwoWay finishes in OneWay: no action
on SendLie in ThreeWay finishes in ThreeWay: SENDLIE on HALSChanged in TwoWay finishes in TwoWay: store HALS
on TimerTick in TwoWay finishes in TwoWay: PUSH SendLie event, if on UpdateZTPOffer in TwoWay finishes in TwoWay: send offer to ZTP
holdtime expired PUSH HoldtimeExpired event FSM
on HALSChanged in OneWay finishes in OneWay: store HALS on NeighborChangedLevel in TwoWay finishes in OneWay: no action
on SendLie in OneWay finishes in OneWay: SENDLIE on NewNeighbor in OneWay finishes in TwoWay: PUSH SendLie event
on NeighborChangedAddress in ThreeWay finishes in OneWay: no
action
on HALChanged in OneWay finishes in OneWay: store new HAL
on NeighborChangedLevel in OneWay finishes in OneWay: no action
on HoldtimeExpired in TwoWay finishes in OneWay: no action
on SendLie in TwoWay finishes in TwoWay: SEND_LIE
on LevelChanged in TwoWay finishes in OneWay: update level with on LevelChanged in TwoWay finishes in OneWay: update level with
event value event value
on LieRcvd in TwoWay finishes in TwoWay: PROCESS_LIE on NeighborChangedAddress in OneWay finishes in OneWay: no action
on HALSChanged in ThreeWay finishes in ThreeWay: store HALS on HATChanged in TwoWay finishes in TwoWay: store HAT
on UpdateZTPOffer in ThreeWay finishes in ThreeWay: send offer to on LieRcvd in TwoWay finishes in TwoWay: PROCESS_LIE
ZTP FSM
on HALSChanged in TwoWay finishes in TwoWay: store HALS on MultipleNeighbors in ThreeWay finishes in OneWay: no action
on LieRcvd in OneWay finishes in OneWay: PROCESS_LIE on MTUMismatch in OneWay finishes in OneWay: no action
on NeighborChangedLevel in ThreeWay finishes in OneWay: no action on SendLie in OneWay finishes in OneWay: SEND_LIE
on LieRcvd in OneWay finishes in OneWay: PROCESS_LIE
on HoldtimeExpired in TwoWay finishes in OneWay: no action
on TimerTick in ThreeWay finishes in ThreeWay: PUSH SendLie event, on TimerTick in ThreeWay finishes in ThreeWay: PUSH SendLie event,
if holdtime expired PUSH HoldtimeExpired event if holdtime expired PUSH HoldtimeExpired event
on UnacceptableHeader in ThreeWay finishes in OneWay: no action on TimerTick in OneWay finishes in OneWay: PUSH SendLie event
on SendLie in TwoWay finishes in TwoWay: SENDLIE on PODMismatch in ThreeWay finishes in OneWay: no action
on LevelChanged in OneWay finishes in OneWay: update level with on LevelChanged in ThreeWay finishes in OneWay: update level with
event value, PUSH SendLie event event value
on NeighborChangedAddress in ThreeWay finishes in OneWay: no on NeighborChangedLevel in ThreeWay finishes in OneWay: no action
action
on HALChanged in TwoWay finishes in TwoWay: store new HAL on UpdateZTPOffer in OneWay finishes in OneWay: send offer to ZTP
FSM
on NewNeighbor in OneWay finishes in TwoWay: PUSH SendLie event on UpdateZTPOffer in ThreeWay finishes in ThreeWay: send offer to
ZTP FSM
on ValidReflection in TwoWay finishes in ThreeWay: no action on HATChanged in OneWay finishes in OneWay: store HAT
on UnacceptableHeader in TwoWay finishes in OneWay: no action on HATChanged in ThreeWay finishes in ThreeWay: store HAT
on LieRcvd in ThreeWay finishes in ThreeWay: PROCESS_LIE on HoldtimeExpired in OneWay finishes in OneWay: no action
on NeighborDroppedReflection in ThreeWay finishes in TwoWay: no on UnacceptableHeader in OneWay finishes in OneWay: no action
action
on PODMismatch in OneWay finishes in OneWay: no action
on SendLie in ThreeWay finishes in ThreeWay: SEND_LIE
on NeighborChangedAddress in TwoWay finishes in OneWay: no action on NeighborChangedAddress in TwoWay finishes in OneWay: no action
on ValidReflection in ThreeWay finishes in ThreeWay: no action
on HALSChanged in OneWay finishes in OneWay: store HALS
on HoldtimeExpired in ThreeWay finishes in OneWay: no action on HoldtimeExpired in ThreeWay finishes in OneWay: no action
on HATChanged in TwoWay finishes in TwoWay: store HAT on HALSChanged in ThreeWay finishes in ThreeWay: store HALS
on UnacceptableHeader in OneWay finishes in OneWay: no action on NeighborDroppedReflection in OneWay finishes in OneWay: no
action
on TimerTick in OneWay finishes in OneWay: PUSH SendLie event on PODMismatch in TwoWay finishes in OneWay: no action
on Entry into OneWay: CLEANUP and then process event SendLie on Entry into OneWay: CLEANUP
Following words are used for well known procedures: Following words are used for well known procedures:
1. PUSH Event: pushes an event to be executed by the FSM upon exit 1. PUSH Event: pushes an event to be executed by the FSM upon exit
of this action of this action
2. CLEANUP: neighbor MUST be reset to unknown 2. CLEANUP: neighbor MUST be reset to unknown
3. SENDLIE: create a new LIE packet 3. SEND_LIE: create a new LIE packet
1. reflecting the neighbor if known and valid and 1. reflecting the neighbor if known and valid and
2. setting the necessary `not_a_ztp_offer` variable if level was 2. setting the necessary `not_a_ztp_offer` variable if level was
derived from last known neighbor on this interface and derived from last known neighbor on this interface and
3. setting `you_are_not_flood_repeater` to computed value 3. setting `you_are_not_flood_repeater` to computed value
4. PROCESS_LIE: 4. PROCESS_LIE:
1. if lie has wrong major version OR our own system ID or 1. if lie has wrong major version OR our own system ID or
invalid system ID then CLEANUP else invalid system ID then CLEANUP else
skipping to change at page 82, line 14 skipping to change at page 100, line 26
2. setting the necessary `not_a_ztp_offer` variable if level was 2. setting the necessary `not_a_ztp_offer` variable if level was
derived from last known neighbor on this interface and derived from last known neighbor on this interface and
3. setting `you_are_not_flood_repeater` to computed value 3. setting `you_are_not_flood_repeater` to computed value
4. PROCESS_LIE: 4. PROCESS_LIE:
1. if lie has wrong major version OR our own system ID or 1. if lie has wrong major version OR our own system ID or
invalid system ID then CLEANUP else invalid system ID then CLEANUP else
2. if lie has undefined level OR my level is undefined OR this 2. if lie has non matching MTUs then CLEANUP, PUSH
UpdateZTPOffer, PUSH MTUMismatch else
3. if PoD rules do not allow adjacency forming then CLEANUP,
PUSH PODMismatch, PUSH MTUMismatch else
4. if lie has undefined level OR my level is undefined OR this
node is leaf and remote level lower than HAT OR (lie's level node is leaf and remote level lower than HAT OR (lie's level
is not leaf AND its difference is more than one from my is not leaf AND its difference is more than one from my
level) then CLEANUP, PUSH UpdateZTPOffer, PUSH level) then CLEANUP, PUSH UpdateZTPOffer, PUSH
UnacceptableHeader else UnacceptableHeader else
3. push UpdateZTPOffer, construct temporary new neighbor 5. PUSH UpdateZTPOffer, construct temporary new neighbor
structure with values from lie, if no current neighbor exists structure with values from lie, if no current neighbor exists
then set neighbor to new neighbor, PUSH NewNeighbor event, then set neighbor to new neighbor, PUSH NewNeighbor event,
CHECK_THREE_WAY else CHECK_THREE_WAY else
1. if current neighbor system ID differs from lie's system 1. if current neighbor system ID differs from lie's system
ID then PUSH MultipleNeighbors else ID then PUSH MultipleNeighbors else
2. if current neighbor stored level differs from lie's level 2. if current neighbor stored level differs from lie's level
then PUSH NeighborChangedLevel else then PUSH NeighborChangedLevel else
skipping to change at page 83, line 5 skipping to change at page 101, line 19
5. CHECK_THREE_WAY: if current state is one-way do nothing else 5. CHECK_THREE_WAY: if current state is one-way do nothing else
1. if lie packet does not contain neighbor then if current state 1. if lie packet does not contain neighbor then if current state
is three-way then PUSH NeighborDroppedReflection else is three-way then PUSH NeighborDroppedReflection else
2. if packet reflects this system's ID and local port and state 2. if packet reflects this system's ID and local port and state
is three-way then PUSH event ValidReflection else PUSH event is three-way then PUSH event ValidReflection else PUSH event
MultipleNeighbors MultipleNeighbors
B.2. ZTP B.2. ZTP FSM
Initial state is ComputeBestOffer.
digraph G04743cd825cc40c5b93de0616ffb851b { digraph G04743cd825cc40c5b93de0616ffb851b {
N29e7db3976644f62b6f3b2801bccb854[label="Enter"] N29e7db3976644f62b6f3b2801bccb854[label="Enter"]
[style="dashed"][shape="plain"]; [style="dashed"][shape="plain"];
N33df4993a1664be18a2196001c27a64c[label="HoldingDown"][shape="oval"]; N33df4993a1664be18a2196001c27a64c[label="HoldingDown"][shape="oval"];
N839f77189e324c82b21b8a709b4b021d[label="ComputeBestOffer"][shape="oval"]; N839f77189e324c82b21b8a709b4b021d[label="ComputeBestOffer"][shape="oval"];
Nc97f2b02808d4751afcc630687bf7421[label="UpdatingClients"][shape="oval"]; Nc97f2b02808d4751afcc630687bf7421[label="UpdatingClients"][shape="oval"];
N7ad21867360c44709be20a99f33dd1f7[label="Enter"] N7ad21867360c44709be20a99f33dd1f7[label="Enter"]
[style="dashed"][shape="plain"]; [style="dashed"][shape="plain"];
N33df4993a1664be18a2196001c27a64c -> N33df4993a1664be18a2196001c27a64c N33df4993a1664be18a2196001c27a64c -> N33df4993a1664be18a2196001c27a64c
skipping to change at page 84, line 21 skipping to change at page 102, line 37
[color="red"][arrowhead="normal" dir="both" arrowtail="none"]; [color="red"][arrowhead="normal" dir="both" arrowtail="none"];
N33df4993a1664be18a2196001c27a64c -> N839f77189e324c82b21b8a709b4b021d N33df4993a1664be18a2196001c27a64c -> N839f77189e324c82b21b8a709b4b021d
[label="|HoldDownExpired|"][color="green"][arrowhead="normal" dir="both" [label="|HoldDownExpired|"][color="green"][arrowhead="normal" dir="both"
arrowtail="none"]; arrowtail="none"];
Nc97f2b02808d4751afcc630687bf7421 -> N839f77189e324c82b21b8a709b4b021d Nc97f2b02808d4751afcc630687bf7421 -> N839f77189e324c82b21b8a709b4b021d
[label="|ChangeLocalLeafIndications|\n|ChangeLocalConfiguredLevel|"] [label="|ChangeLocalLeafIndications|\n|ChangeLocalConfiguredLevel|"]
[color="gold"] [color="gold"]
[arrowhead="normal" dir="both" arrowtail="none"]; [arrowhead="normal" dir="both" arrowtail="none"];
} }
LIE FSM DOT ZTP FSM DOT
Events Events
o ChangeLocalLeafIndications: node configured with new leaf flags o ChangeLocalLeafIndications: node configured with new leaf flags
o ChangeLocalConfiguredLevel: node locally configured with a defined o ChangeLocalConfiguredLevel: node locally configured with a defined
level level
o NeighborOffer: a new neighbor offer with optional level and o NeighborOffer: a new neighbor offer with optional level and
neighbor state neighbor state
skipping to change at page 85, line 26 skipping to change at page 103, line 43
on NeighborOffer in ComputeBestOffer finishes in ComputeBestOffer: on NeighborOffer in ComputeBestOffer finishes in ComputeBestOffer:
if no level offered REMOVE_OFFER else if no level offered REMOVE_OFFER else
if level > leaf then UPDATE_OFFER else REMOVE_OFFER if level > leaf then UPDATE_OFFER else REMOVE_OFFER
on BetterHAT in UpdatingClients finishes in ComputeBestOffer: no on BetterHAT in UpdatingClients finishes in ComputeBestOffer: no
action action
on ChangeLocalConfiguredLevel in HoldingDown finishes in on ChangeLocalConfiguredLevel in HoldingDown finishes in
ComputeBestOffer: store level ComputeBestOffer: store configured level
on BetterHAL in ComputeBestOffer finishes in ComputeBestOffer: on BetterHAL in ComputeBestOffer finishes in ComputeBestOffer:
LEVEL_COMPUTE LEVEL_COMPUTE
on HoldDownExpired in HoldingDown finishes in ComputeBestOffer: on HoldDownExpired in HoldingDown finishes in ComputeBestOffer:
PURGE_OFFERS PURGE_OFFERS
on ShortTic in HoldingDown finishes in HoldingDown: if holddown on ShortTic in HoldingDown finishes in HoldingDown: if holddown
timer expired PUSH_EVENT HoldDownExpired timer expired PUSH_EVENT HoldDownExpired
on ComputationDone in ComputeBestOffer finishes in on ComputationDone in ComputeBestOffer finishes in
UpdatingClients: no action UpdatingClients: no action
on LostHAL in UpdatingClients finishes in HoldingDown: if any on LostHAL in UpdatingClients finishes in HoldingDown: if any
southbound adjacencies present update holddown timer to normal southbound adjacencies present update holddown timer to normal
duration else fire holddown timer immediately duration else fire holddown timer immediately