draft-ietf-rtgwg-mrt-frr-architecture-02.txt   draft-ietf-rtgwg-mrt-frr-architecture-03.txt 
Routing Area Working Group A. Atlas, Ed. Routing Area Working Group A. Atlas, Ed.
Internet-Draft R. Kebler Internet-Draft R. Kebler
Intended status: Standards Track Juniper Networks Intended status: Standards Track Juniper Networks
Expires: August 28, 2013 G. Enyedi Expires: January 13, 2014 G. Enyedi
A. Csaszar A. Csaszar
J. Tantsura J. Tantsura
Ericsson Ericsson
M. Konstantynowicz M. Konstantynowicz
Cisco Systems Cisco Systems
R. White R. White
Verisign VCE
M. Shand July 12, 2013
February 24, 2013
An Architecture for IP/LDP Fast-Reroute Using Maximally Redundant Trees An Architecture for IP/LDP Fast-Reroute Using Maximally Redundant Trees
draft-ietf-rtgwg-mrt-frr-architecture-02 draft-ietf-rtgwg-mrt-frr-architecture-03
Abstract Abstract
As IP and LDP Fast-Reroute are increasingly deployed, the coverage With increasing deployment of Loop-Free Alternates (LFA) [RFC5286],
limitations of Loop-Free Alternates are seen as a problem that it is clear that a complete solution for IP and LDP Fast-Reroute is
requires a straightforward and consistent solution for IP and LDP, required. This specification provides that solution. IP/LDP Fast-
for unicast and multicast. This draft describes an architecture Reroute with Maximally Redundant Trees (MRT-FRR) is a technology that
based on redundant backup trees where a single failure can cut a gives link-protection and node-protection with 100% coverage in any
point-of-local-repair from the destination only on one of the pair of network topology that is still connected after the failure.
redundant trees.
One innovative algorithm to compute such topologies is maximally
disjoint backup trees. Each router can compute its next-hops for
each pair of maximally disjoint trees rooted at each node in the IGP
area with computational complexity similar to that required by
Dijkstra.
The additional state, address and computation requirements are MRT removes all need to engineer for coverage. MRT is also extremely
believed to be significantly less than the Not-Via architecture computationally efficient. For any router in the network, the MRT
requires. computation is less than the LFA computation for a node with three or
more neighbors.
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 http://datatracker.ietf.org/drafts/current/. Drafts is at http://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 August 28, 2013. This Internet-Draft will expire on January 13, 2014.
Copyright Notice Copyright Notice
Copyright (c) 2013 IETF Trust and the persons identified as the Copyright (c) 2013 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
(http://trustee.ietf.org/license-info) in effect on the date of (http://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. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1. Goals for Extending IP Fast-Reroute coverage beyond LFA . 4 1.1. Importance of 100% Coverage . . . . . . . . . . . . . . . 4
2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.2. Partial Deployment and Backwards Compatibility . . . . . 5
3. Maximally Redundant Trees (MRT) . . . . . . . . . . . . . . . 6 2. Requirements Language . . . . . . . . . . . . . . . . . . . . 6
4. Maximally Redundant Trees (MRT) and Fast-Reroute . . . . . . . 8 3. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 6
5. Unicast Forwarding with MRT Fast-Reroute . . . . . . . . . . . 9 4. Maximally Redundant Trees (MRT) . . . . . . . . . . . . . . . 7
5.1. LDP Unicast Forwarding - Avoid Tunneling . . . . . . . . . 10 5. Maximally Redundant Trees (MRT) and Fast-Reroute . . . . . . 9
5.2. IP Unicast Traffic . . . . . . . . . . . . . . . . . . . . 10 6. Unicast Forwarding with MRT Fast-Reroute . . . . . . . . . . 10
6. Protocol Extensions and Considerations: OSPF and ISIS . . . . 12 6.1. LDP Unicast Forwarding - Avoid Tunneling . . . . . . . . 10
7. Protocol Extensions and considerations: LDP . . . . . . . . . 14 6.2. IP Unicast Traffic . . . . . . . . . . . . . . . . . . . 11
8. Multi-homed Prefixes . . . . . . . . . . . . . . . . . . . . . 15 7. Protocol Extensions and Considerations: OSPF and ISIS . . . . 12
9. Inter-Area and ABR Forwarding Behavior . . . . . . . . . . . . 16 8. Protocol Extensions and considerations: LDP . . . . . . . . . 14
10. Issues with Area Abstraction . . . . . . . . . . . . . . . . . 19 9. Inter-Area and ABR Forwarding Behavior . . . . . . . . . . . 15
11. Partial Deployment and Islands of Compatible MRT FRR 10. Prefixes Multiply Attached to the MRT Island . . . . . . . . 18
routers . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 10.1. Endpoint Selection . . . . . . . . . . . . . . . . . . . 19
12. Network Convergence and Preparing for the Next Failure . . . . 22 10.2. Named Proxy-Nodes . . . . . . . . . . . . . . . . . . . 21
12.1. Micro-forwarding loop prevention and MRTs . . . . . . . . 22 10.2.1. Computing if an Island Neighbor (IN) is loop-free . 22
12.2. MRT Recalculation . . . . . . . . . . . . . . . . . . . . 23 10.3. MRT Alternates for Destinations Outside the MRT Island . 23
13. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 23 11. Network Convergence and Preparing for the Next Failure . . . 24
14. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 23 11.1. Micro-forwarding loop prevention and MRTs . . . . . . . 24
15. Security Considerations . . . . . . . . . . . . . . . . . . . 24 11.2. MRT Recalculation . . . . . . . . . . . . . . . . . . . 24
16. References . . . . . . . . . . . . . . . . . . . . . . . . . . 24 12. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 25
16.1. Normative References . . . . . . . . . . . . . . . . . . . 24 13. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 25
16.2. Informative References . . . . . . . . . . . . . . . . . . 24 14. Security Considerations . . . . . . . . . . . . . . . . . . . 25
Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 25 15. References . . . . . . . . . . . . . . . . . . . . . . . . . 25
15.1. Normative References . . . . . . . . . . . . . . . . . . 25
15.2. Informative References . . . . . . . . . . . . . . . . . 26
Appendix A. General Issues with Area Abstraction . . . . . . . . 27
Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 28
1. Introduction 1. Introduction
There is still work required to completely provide IP and LDP Fast- This document gives a complete solution for IP/LDP fast-reroute
Reroute[RFC5714] for unicast and multicast traffic. This draft [RFC5714]. MRT-FRR creates two alternate trees separate from the
proposes an architecture to provide 100% coverage for unicast primary next-hop forwarding used during stable operation. These two
traffic. The associated multicast architecture is described in trees are maximally diverse from each other, providing link and node
[I-D.atlas-rtgwg-mrt-mc-arch]. protection for 100% of paths and failures as long as the failure does
not cut the network into multiple pieces. This document defines the
Loop-free alternates (LFAs)[RFC5286] provide a useful mechanism for architecture for IP/LDP fast-reroute with MRT. The associated
link and node protection but getting complete coverage is quite hard. protocol extensions are defined in [I-D.atlas-ospf-mrt] and
[LFARevisited] defines sufficient conditions to determine if a [I-D.atlas-mpls-ldp-mrt]. The exact MRT algorithm is defined in
network provides link-protecting LFAs and also proves that augmenting [I-D.enyedi-rtgwg-mrt-frr-algorithm].
a network to provide better coverage is NP-hard.
[I-D.ietf-rtgwg-lfa-applicability] discusses the applicability of LFA
to different topologies with a focus on common PoP architectures.
While Not-Via [I-D.ietf-rtgwg-ipfrr-notvia-addresses] is defined as IP/LDP Fast-Reroute with MRT (MRT-FRR) uses two maximally diverse
an architecture, in practice, it has proved too complicated and forwarding topologies to provide alternates. A primary next-hop
stateful to spark substantial interest in implementation or should be on only one of the diverse forwarding topologies; thus, the
deployment. Academic implementations [LightweightNotVia] exist and other can be used to provide an alternate. Once traffic has been
have found the address management complexity high (but no moved to one of MRTs, it is not subject to further repair actions.
standardization has been done to reduce this). Thus, the traffic will not loop even if a worse failure (e.g. node)
occurs when protection was only available for a simpler failure (e.g.
link).
A different approach is needed and that is what is described here. In addition to supporting IP and LDP unicast fast-reroute, the
It is based on the idea of using disjoint backup topologies as diverse forwarding topologies and guarantee of 100% coverage permit
realized by Maximally Redundant Trees (described in fast-reroute technology to be applied to multicast traffic as
[LightweightNotVia]); the general architecture can also apply to described in [I-D.atlas-rtgwg-mrt-mc-arch].
future improved redundant tree algorithms.
1.1. Goals for Extending IP Fast-Reroute coverage beyond LFA Other existing or proposed solutions are partial solutions or have
significant issues, as described below.
Any scheme proposed for extending IPFRR network topology coverage Summary Comparison of IP/LDP FRR Methods
beyond LFA, apart from attaining basic IPFRR properties, should also
aim to achieve the following usability goals:
o ensure maximum physically feasible link and node disjointness +-----------+---------------+---------------+-----------------------+
regardless of topology, | Method | Coverage | Alternate | Computation (in SPFs) |
| | | Looping? | |
+-----------+---------------+---------------+-----------------------+
| MRT-FRR | 100% | None | less than 3 |
| | Link/Node | | |
| | | | |
| LFA | Partial | Possible | per neighbor |
| | Link/Node | | |
| | | | |
| Remote | Partial | Possible | per neighbor (link) |
| LFA | Link/Node | | or neighbor's |
| | | | neighbor (node) |
| | | | |
| Not-Via | 100% | None | per link and node |
| | Link/Node | | |
+-----------+---------------+---------------+-----------------------+
o automatically compute backup next-hops based on the topology Table 1
information distributed by link-state IGP,
o do not require any signaling in the case of failure and use pre- Loop-Free Alternates (LFA): LFAs [RFC5286] provide limited
programmed backup next-hops for forwarding, topology-dependent coverage for link and node protection.
Restrictions on choice of alternates can be relaxed to improve
coverage, but this can cause forwarding loops if a worse failure
is experienced than protected against. Augmenting a network to
provide better coverage is NP-hard [LFARevisited]. [RFC6571]
discusses the applicability of LFA to different topologies with a
focus on common PoP architectures.
o introduce minimal amount of additional addressing and state on Remote LFA: Remote LFAs [I-D.ietf-rtgwg-remote-lfa] improve
routers, coverage over LFAs for link protection but still cannot guarantee
complete coverage. The trade-off of looping traffic to improve
coverage is still made. Remote LFAs can provide node-protection
[I-D.litkowski-rtgwg-node-protect-remote-lfa] but not guaranteed
coverage and the computation required is quite high (an SPF per
neighbor's neighbor). [I-D.bryant-ipfrr-tunnels] describes
additional mechanisms to further improve coverage, at the cost of
added complexity.
o enable gradual introduction of the new scheme and backward Not-Via: Not-Via [I-D.ietf-rtgwg-ipfrr-notvia-addresses] is the
compatibility, only other solution that provides 100% coverage for link and node
failures and does not have potential looping. However, the
computation is very high (an SPF per failure point) and academic
implementations [LightweightNotVia] have found the address
management complexity to be high.
o and do not impose requirements for external computation. 1.1. Importance of 100% Coverage
Fast-reroute is based upon the single failure assumption - that the
time between single failures is long enough for a network to
reconverge and start forwarding on the new shortest paths. That does
not imply that the network will only experience one failure or
change.
2. Terminology It is straightforward to analyze a particular network topology for
coverage. However, a real network does not always have the same
topology. For instance, maintenance events will take links or nodes
out of use. Simply costing out a link can have a significant effect
on what LFAs are available. Similarly, after a single failure has
happened, the topology is changed and its associated coverage.
Finally, many networks have new routers or links added and removed;
each of those changes can have an effect on the coverage for
topology-sensitive methods such as LFA and Remote LFA. If fast-
reroute is important for the network services provided, then a method
that guarantees 100% coverage is important to accomodate natural
network topology changes.
2-connected: A graph that has no cut-vertices. This is a graph Asymmetric link costs are also a common aspect of networks. There
that requires two nodes to be removed before the network is are at least three common causes for them. First, any broadcast
partitioned. interface is represented by a pseudo-node and has asymmetric link
costs to and from that pseudo-node. Second, when routers come up or
a link with LDP comes up, it is recommended in [RFC5443] and
[RFC3137] that the link metric be raised to the maximum cost; this
may not be symmetric and for [RFC3137] is not expected to be. Third,
techniques such as IGP metric tuning for traffic-engineering can
result in asymmetric link costs. A fast-reroute solution needs to
handle network topologies with asymmetric link costs.
2-connected cluster: A maximal set of nodes that are 2-connected. When a network needs to use a micro-loop prevention mechanism
[RFC5715] such as Ordered FIB[I-D.ietf-rtgwg-ordered-fib] or Farside
Tunneling[RFC5715], then the whole IGP area needs to have alternates
available so that the micro-loop prevention mechanism, which requires
slower network convergence, can take the necessary time without
impacting traffic badly. Without complete coverage, traffic to the
unprotected destinations will be dropped for significantly longer
than with current convergence - where routers individually converge
as fast as possible.
2-edge-connected: A network graph where at least two links must be 1.2. Partial Deployment and Backwards Compatibility
removed to partition the network.
ADAG: Almost Directed Acyclic Graph - a graph that, if all links MRT-FRR supports partial deployment. As with many new features, the
incoming to the root were removed, would be a DAG. protocols (OSPF, LDP, ISIS) indicate their capability to support MRT.
Inside the MRT-capable connected group of routers (referred to as an
MRT Island), the MRTs are computed. Alternates to destinations
outside the MRT Island are computed and depend upon the existence of
a loop-free neighbor of the MRT Island for that destination.
block: Either a 2-connected cluster, a cut-edge, or an isolated 2. Requirements Language
vertex.
cut-link: A link whose removal partitions the network. A cut-link The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
by definition must be connected between two cut-vertices. If "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
there are multiple parallel links, then they are referred to as document are to be interpreted as described in [RFC2119]
cut-links in this document if removing the set of parallel links
would partition the network.
cut-vertex: A vertex whose removal partitions the network. 3. Terminology
DAG: Directed Acyclic Graph - a graph where all links are directed network graph: A graph that reflects the network topology where all
and there are no cycles in it. links connect exactly two nodes and broadcast links have been
transformed into the standard pseudo-node representation.
GADAG: Generalized ADAG - a graph that is the combination of the Redundant Trees (RT): A pair of trees where the path from any node
ADAGs of all blocks. X to the root R along the first tree is node-disjoint with the
path from the same node X to the root along the second tree.
These can be computed in 2-connected graphs.
Maximally Redundant Trees (MRT): A pair of trees where the path Maximally Redundant Trees (MRT): A pair of trees where the path
from any node X to the root R along the first tree and the path from any node X to the root R along the first tree and the path
from the same node X to the root along the second tree share the from the same node X to the root along the second tree share the
minimum number of nodes and the minimum number of links. Each minimum number of nodes and the minimum number of links. Each
such shared node is a cut-vertex. Any shared links are cut-links. such shared node is a cut-vertex. Any shared links are cut-links.
Any RT is an MRT but many MRTs are not RTs. Any RT is an MRT but many MRTs are not RTs.
network graph: A graph that reflects the network topology where all MRT-Red: MRT-Red is used to describe one of the two MRTs; it is
links connect exactly two nodes and broadcast links have been used to described the associated forwarding topology and MT-ID.
transformed into the standard pseudo-node representation. Specifically, MRT-Red is the decreasing MRT where links in the
GADAG are taken in the direction from a higher topologically
ordered node to a lower one.
Redundant Trees (RT): A pair of trees where the path from any node MRT-Blue: MRT-Blue is used to describe one of the two MRTs; it is
X to the root R along the first tree is node-disjoint with the used to described the associated forwarding topology and MT-ID.
path from the same node X to the root along the second tree. Specifically, MRT-Blue is the increasing MRT where links in the
These can be computed in 2-connected graphs. GADAG are taken in the direction from a lower topologically
ordered node to a higher one.
3. Maximally Redundant Trees (MRT) Rainbow MRT: It is useful to have an MT-ID that refers to the
multiple MRT topologies and to the default topology. This is
referred to as the Rainbow MRT MT-ID and is used by LDP to reduce
signaling and permit the same label to always be advertised to all
peers for the same (MT-ID, Prefix).
In the last few years, there's been substantial research on how to MRT Island: From the computing router, the set of routers that
compute and use redundant trees. Redundant trees are directed support a particular MRT profile and are connected.
spanning trees that provide disjoint paths towards their common root.
These redundant trees only exist and provide link protection if the
network is 2-edge-connected and node protection if the network is
2-connected. Such connectiveness may not be the case in real
networks, either due to architecture or due to a previous failure.
The work on maximally redundant trees has added two useful pieces
that make them ready for use in a real network.
o Computable regardless of network topology: The maximally redundant Island Border Router (IBR): A router in the MRT Island that is
trees are computed so that only the cut-edges or cut-vertices are connected to a router not in the MRT Island and both routers are
shared between the multiple trees. in a common area or level.
o Computationally practical algorithm is based on a common network Island Neighbor (IN): A router that is not in the MRT Island but is
topology database. Algorithm variants can compute in O( e) or O(e adjacent to an IBR and in the same area/level as the IBR.
+ n log n), as given in [I-D.enyedi-rtgwg-mrt-frr-algorithm].
There is, of course, significantly more in the literature related to cut-link: A link whose removal partitions the network. A cut-link
redundant trees and even fast-reroute, but the formulation of the by definition must be connected between two cut-vertices. If
Maximally Redundant Trees (MRT) algorithm makes it very well suited there are multiple parallel links, then they are referred to as
to use in routers. cut-links in this document if removing the set of parallel links
would partition the network graph.
A known disadvantage of MRT, and redundant trees in general, is that cut-vertex: A vertex whose removal partitions the network graph.
the trees do not necessarily provide shortest detour paths. The use
of the shortest-path-first algorithm in tree-building and including
all links in the network as possibilities for one path or another
should improve this. Modeling is underway to investigate and compare
the MRT alternates to the optimal
[I-D.enyedi-rtgwg-mrt-frr-algorithm]. Providing shortest detour
paths would require failure-specific detour paths to the
destinations, but the state-reduction advantage of MRT lies in the
detour being established per destination (root) instead of per
destination AND per failure.
The specific algorithms to compute MRTs as well as the logic behind 2-connected: A graph that has no cut-vertices. This is a graph
that algorithm and alternative computational approaches are given in that requires two nodes to be removed before the network is
detail in [I-D.enyedi-rtgwg-mrt-frr-algorithm]. Those interested are partitioned.
highly recommended to read that document. This document describes
how the MRTs can be used and not how to compute them. 2-connected cluster: A maximal set of nodes that are 2-connected.
2-edge-connected: A network graph where at least two links must be
removed to partition the network.
block: Either a 2-connected cluster, a cut-edge, or an isolated
vertex.
DAG: Directed Acyclic Graph - a graph where all links are directed
and there are no cycles in it.
ADAG: Almost Directed Acyclic Graph - a graph that, if all links
incoming to the root were removed, would be a DAG.
GADAG: Generalized ADAG - a graph that is the combination of the
ADAGs of all blocks.
named proxy-node: A proxy-node can represent a destination prefix
that can be attached to the MRT Island via at least two routers.
It is named if there is a way that traffic can be encapsulated to
reach specifically that proxy node; this could be because there is
an LDP FEC for the associated prefix or because MRT-Red and MRT-
Blue IP addresses are advertised in an undefined fashion for that
proxy-node.
4. Maximally Redundant Trees (MRT)
A pair of Maximally Redundant Trees are directed spanning trees that
provide maximally disjoint paths towards their common root. Only
links or nodes whose failure would partition the network (i.e. cut-
links and cut-vertices) are shared between the trees. The algorithm
to compute MRTs is given in [I-D.enyedi-rtgwg-mrt-frr-algorithm].
This algorithm can be computed in O(e + n log n); it is less than
three SPFs. Modeling results comparing MRT alternates to the optimal
are described in [I-D.enyedi-rtgwg-mrt-frr-algorithm]. This document
describes how the MRTs can be used and not how to compute them.
MRT provides destination-based trees for each destination. Each
router stores its normal primary next-hop(s) as well as MRT-Blue
next-hop(s) and MRT-Red next-hop(s) toward each destination. The
alternate will be selected between the MRT-Blue and MRT-Red.
The most important thing to understand about MRTs is that for each The most important thing to understand about MRTs is that for each
pair of destination-routed MRTs, there is a path from every node X to pair of destination-routed MRTs, there is a path from every node X to
the destination D on the Blue MRT that is as disjoint as possible the destination D on the Blue MRT that is as disjoint as possible
from the path on the Red MRT. The two paths along the two MRTs to a from the path on the Red MRT.
given destination-root of a 2-connected graph are node-disjoint and
link-disjoint, while in any non-2-connected graph, only the cut-
vertices and cut-edges can be contained by both of the paths.
For example, in Figure 1, there is a network graph that is For example, in Figure 1, there is a network graph that is
2-connected in (a) and associated MRTs in (b) and (c). One can 2-connected in (a) and associated MRTs in (b) and (c). One can
consider the paths from B to R; on the Blue MRT, the paths are consider the paths from B to R; on the Blue MRT, the paths are
B->F->D->E->R or B->C->D->E->R. On the Red MRT, the path is B->A->R. B->F->D->E->R or B->C->D->E->R. On the Red MRT, the path is B->A->R.
These are clearly link and node-disjoint. These MRTs are redundant These are clearly link and node-disjoint. These MRTs are redundant
trees because the paths are disjoint. trees because the paths are disjoint.
[E]---[D]---| [E]<--[D]<--| [E]-->[D]---| [E]---[D]---| [E]<--[D]<--| [E]-->[D]---|
| | | | ^ | | | | | | | ^ | | |
| | | V | | V V | | | V | | V V
[R] [F] [C] [R] [F] [C] [R] [F] [C] [R] [F] [C] [R] [F] [C] [R] [F] [C]
| | | ^ ^ ^ | | | | | ^ ^ ^ | |
| | | | | | V | | | | | | | V |
[A]---[B]---| [A]-->[B]---| [A]---[B]<--| [A]---[B]---| [A]-->[B]---| [A]<--[B]<--|
(a) (b) (c) (a) (b) (c)
a 2-connected graph Blue MRT towards R Red MRT towards R a 2-connected graph Blue MRT towards R Red MRT towards R
Figure 1: A 2-connected Network Figure 1: A 2-connected Network
By contrast, in Figure 2, the network in (a) is not 2-connected. If By contrast, in Figure 2, the network in (a) is not 2-connected. If
F, G or the link F<->G failed, then the network would be partitioned. F, G or the link F<->G failed, then the network would be partitioned.
It is clearly impossible to have two link-disjoint or node-disjoint It is clearly impossible to have two link-disjoint or node-disjoint
paths from G, I or J to R. The MRTs given in (b) and (c) offer paths paths from G, I or J to R. The MRTs given in (b) and (c) offer paths
that are as disjoint as possible. For instance, the paths from B to that are as disjoint as possible. For instance, the paths from B to
R are the same as in Figure 1 and the path from G to R on the Blue R are the same as in Figure 1 and the path from G to R on the Blue
MRT is G->F->D->E->R and on the Red MRT is G->F->B->A->R. MRT is G->F->D->E->R and on the Red MRT is G->F->B->A->R.
[E]---[D]---| [E]---[D]---|
| | | |----[I] | | | |----[I]
| | | | | | | | | |
[R]---[C] [F]---[G] | [R]---[C] [F]---[G] |
| | | | | | | | | |
| | | |----[J] | | | |----[J]
[A]---[B]---| [A]---[B]---|
(a) (a)
a non-2-connected graph a non-2-connected graph
[E]<--[D]<--| [E]-->[D]---| [E]<--[D]<--| [E]-->[D]
| ^ | [I] | | [I] | ^ | [I] | |----[I]
V | | ^ V V | V | | | V V ^
[R]<--[C] [F]<--[G] | [R]---[C] [F]<--[G] | [R] [C] [F]<--[G] | [R]<--[C] [F]<--[G] |
^ ^ | | ^ | | ^ V ^ ^ ^ V ^ | |
| | |--->[J] | V | |----[J] | | |----[J] | | [J]
[A]-->[B]---| [A]<--[B]<--| [A]-->[B]---| [A]<--[B]<--|
(b) (c) (b) (c)
Blue MRT towards R Red MRT towards R Blue MRT towards R Red MRT towards R
Figure 2: A non-2-connected network Figure 2: A non-2-connected network
4. Maximally Redundant Trees (MRT) and Fast-Reroute 5. Maximally Redundant Trees (MRT) and Fast-Reroute
In normal IGP routing, each router has its shortest-path-tree to all In normal IGP routing, each router has its shortest-path-tree to all
destinations. From the perspective of a particular destination, D, destinations. From the perspective of a particular destination, D,
this looks like a reverse SPT (rSPT). To use maximally redundant this looks like a reverse SPT (rSPT). To use maximally redundant
trees, in addition, each destination D has two MRTs associated with trees, in addition, each destination D has two MRTs associated with
it; by convention these will be called the blue and red MRTs. it; by convention these will be called the MRT-Blue and MRT-Red.
MRT-FRR is realized by using multi-topology forwarding. There is a
MRT-Blue forwarding topology and a MRT-Red forwarding topology.
Any IP/LDP fast-reroute technique beyond LFA requires an additional Any IP/LDP fast-reroute technique beyond LFA requires an additional
dataplane procedure, such as an additional forwarding mechanism. The dataplane procedure, such as an additional forwarding mechanism. The
well-known options are tunneling (e.g. well-known options are multi-topology forwarding (used by MRT-FRR),
[I-D.ietf-rtgwg-ipfrr-notvia-addresses] or tunneling (e.g. [I-D.ietf-rtgwg-ipfrr-notvia-addresses] or
[I-D.ietf-rtgwg-remote-lfa]), per-interface forwarding (e.g. Loop- [I-D.ietf-rtgwg-remote-lfa]), and per-interface forwarding (e.g.
Free Failure Insensitive Routing in [EnyediThesis]), and multi- Loop-Free Failure Insensitive Routing in [EnyediThesis]).
topology forwarding. MRT is realized by using multi-topology
forwarding. There is a Blue MRT forwarding topology and a Red MRT
forwarding topology.
MRTs are practical to maintain redundancy even after a single link or
node failure. If a pair of MRTs is computed rooted at each
destination, all the destinations remain reachable along one of the
MRTs in the case of a single link or node failure.
When there is a link or node failure affecting the rSPT, each node When there is a link or node failure affecting, but not partitioning,
will still have at least one path via one of the MRTs to reach the the network, each node will still have at least one path via one of
destination D. For example, in Figure 2, C would normally forward the MRTs to reach the destination D. For example, in Figure 2, C
traffic to R across the C<->R link. If that C<->R link fails, then C would normally forward traffic to R across the C<->R link. If that
could use either the Blue MRT path C->D->E->R or the Red MRT path C<->R link fails, then C could use the Blue MRT path C->D->E->R.
C->B->A->R.
As is always the case with fast-reroute technologies, forwarding does As is always the case with fast-reroute technologies, forwarding does
not change until a local failure is detected. Packets are forwarded not change until a local failure is detected. Packets are forwarded
along the shortest path. The appropriate alternate to use is pre- along the shortest path. The appropriate alternate to use is pre-
computed. [I-D.enyedi-rtgwg-mrt-frr-algorithm] describes exactly how computed. [I-D.enyedi-rtgwg-mrt-frr-algorithm] describes exactly how
to determine whether the Blue MRT next-hops or the Red MRT next-hops to determine whether the MRT-Blue next-hops or the MRT-Red next-hops
should be the MRT alternate next-hops for a particular primary next- should be the MRT alternate next-hops for a particular primary next-
hop N to a particular destination D. hop N to a particular destination D.
MRT alternates are always available to use, unless the network has MRT alternates are always available to use. It is a local decision
been partitioned. It is a local decision whether to use an MRT whether to use an MRT alternate, a Loop-Free Alternate or some other
alternate, a Loop-Free Alternate or some other type of alternate. type of alternate.
When a network needs to use a micro-loop prevention mechanism
[RFC5715] such as Ordered FIB[I-D.ietf-rtgwg-ordered-fib] or Farside
Tunneling[RFC5715], then the whole IGP area needs to have alternates
available so that the micro-loop prevention mechanism, which requires
slower network convergence, can take the necessary time without
impacting traffic badly.
As described in [RFC5286], when a worse failure than is anticipated As described in [RFC5286], when a worse failure than is anticipated
happens, using LFAs that are not downstream neighbors can cause happens, using LFAs that are not downstream neighbors can cause
micro-looping. An example is given of link-protecting alternates micro-looping. Section 1.1 of [RFC5286] gives an example of link-
causing a loop on node failure. Even if a worse failure than protecting alternates causing a loop on node failure. Even if a
anticipated happened, the use of MRT alternates will not cause worse failure than anticipated happens, the use of MRT alternates
looping. Therefore, while node-protecting LFAs may be prefered, an will not cause looping. Therefore, while node-protecting LFAs may be
the certainty that no alternate-induced looping will occur is an preferred, the certainty that no alternate-induced looping will occur
advantage of using MRT alternates when the available node-protecting is an advantage of using MRT alternates when the available node-
LFA is not a downstream path. protecting LFA is not a downstream path.
5. Unicast Forwarding with MRT Fast-Reroute 6. Unicast Forwarding with MRT Fast-Reroute
With LFA, there is no need to tunnel unicast traffic, whether IP or With LFA, there is no need to tunnel unicast traffic, whether IP or
LDP. The traffic is simply sent to an alternate. As mentioned LDP. The traffic is simply sent to an alternate. As mentioned
earlier in Section 4, MRT needs multi-topology forwarding. earlier in Section 5, MRT needs multi-topology forwarding.
Unfortunately, neither IP nor LDP provide extra bits for a packet to Unfortunately, neither IP nor LDP provides extra bits for a packet to
indicate its topology. indicate its topology.
Once the MRTs are computed, the two sets of MRTs are seen by the Once the MRTs are computed, the two sets of MRTs are seen by the
forwarding plane as essentially two additional topologies. The same forwarding plane as essentially two additional topologies. The same
considerations apply for forwarding along the MRTs as for handling considerations apply for forwarding along the MRTs as for handling
multiple topologies. multiple topologies.
5.1. LDP Unicast Forwarding - Avoid Tunneling 6.1. LDP Unicast Forwarding - Avoid Tunneling
For LDP, it is very desirable to avoid tunneling because, for at For LDP, it is very desirable to avoid tunneling because, for at
least node protection, tunneling requires knowledge of remote LDP least node protection, tunneling requires knowledge of remote LDP
label mappings and thus requires targeted LDP sessions and the label mappings and thus requires targeted LDP sessions and the
associated management complexity. There are two different mechanisms associated management complexity. There are two different mechanisms
that can be used. that can be used; Option A MUST be supported.
1. Option A - Encode MT-ID in Labels: In addition to sending a 1. Option A - Encode MT-ID in Labels: In addition to sending a
single label for a FEC, a router would provide two additional single label for a FEC, a router would provide two additional
labels with the MT-IDs associated with the Blue MRT or Red MRT labels with the MT-IDs associated with the Blue MRT or Red MRT
forwarding topologies. This is very simple for hardware support. forwarding topologies. This is very simple for hardware support.
It does reduce the label space for other uses. It also increases It does reduce the label space for other uses. It also increases
the memory to store the labels and the communication required by the memory to store the labels and the communication required by
LDP. LDP.
2. Option B - Create Topology-Identification Labels: Use the label- 2. Option B - Create Topology-Identification Labels: Use the label-
skipping to change at page 10, line 45 skipping to change at page 11, line 26
Note that with LDP unicast forwarding, regardless of whether Note that with LDP unicast forwarding, regardless of whether
topology-identification label or encoding topology in label is used, topology-identification label or encoding topology in label is used,
no additional loopbacks per router are required. This is because LDP no additional loopbacks per router are required. This is because LDP
labels are used on a hop-by-hop basis to identify MRT-blue and MRT- labels are used on a hop-by-hop basis to identify MRT-blue and MRT-
red forwading topologies. red forwading topologies.
For greatest hardware compatibility, routers implementing MRT LDP For greatest hardware compatibility, routers implementing MRT LDP
fast-reroute MUST support Option A of encoding the MT-ID in the fast-reroute MUST support Option A of encoding the MT-ID in the
labels. The extensions to indicate an MT-ID for a FEC are described labels. The extensions to indicate an MT-ID for a FEC are described
in Section 3.2.1 of [I-D.ietf-mpls-ldp-multi-topology] in Section 3.2.1 of [I-D.ietf-mpls-ldp-multi-topology].
5.2. IP Unicast Traffic 6.2. IP Unicast Traffic
For IP, there is no currently practical alternative except tunneling. For IP, there is no currently practical alternative except tunneling
The tunnel egress could be the original destination in the area, the to gain the bits needed to indicate the MRT-Blue or MRT-Red
next-next-hop, etc.. If the tunnel egress is the original forwarding topology. The choice of tunnel egress MAY be flexible
destination router, then the traffic remains on the redundant tree since any router closer to the destination than the next-hop can
with sub-optimal routing. If the tunnel egress is the next-next-hop, work. This architecture assumes that the original destination in the
then protection of multi-homed prefixes and node-failure for ABRs is area is selected (see Section 10 for handling of multi-homed
not available. Selection of the tunnel egress is a router-local prefixes); another possible choice is the next-next-hop towards the
decision. destination. For LDP traffic, using the original destination
simplifies MRT-FRR by avoiding the need for targeted LDP sessions to
the next-next-hop. For IP, that consideration doesn't apply but
consistency with LDP is RECOMMENDED. If the tunnel egress is the
original destination router, then the traffic remains on the
redundant tree with sub-optimal routing. Selection of the tunnel
egress is a router-local decision.
There are three options available for marking IP packets with which There are three options available for marking IP packets with which
MRT it should be forwarded in. MRT it should be forwarded in. For greatest hardware compatibility
and ease in removing the MRT-topology marking at area/level
boundaries, routers that support MPLS and implement IP MRT fast-
reroute MUST support Option A - using an LDP label that indicates the
destination and MT-ID.
1. Tunnel IP packets via an LDP LSP. This has the advantage that 1. Tunnel IP packets via an LDP LSP. This has the advantage that
more installed routers can do line-rate encapsulation and more installed routers can do line-rate encapsulation and
decapsulation. Also, no additional IP addresses would need to be decapsulation. Also, no additional IP addresses would need to be
allocated or signaled. allocated or signaled.
A. Option A - LDP Destination-Topology Label: Use a label that a. Option A - LDP Destination-Topology Label: Use a label that
indicates both destination and MRT. This method allows easy indicates both destination and MRT. This method allows easy
tunneling to the next-next-hop as well as to the IGP-area tunneling to the next-next-hop as well as to the IGP-area
destination. For a proxy-node, the destination to use is the destination. For a proxy-node, the destination to use is the
non-proxy-node immediately before the proxy-node on that non-proxy-node immediately before the proxy-node on that
particular color MRT. particular color MRT.
B. Option B - LDP Topology Label: Use a Topology-Identifier b. Option B - LDP Topology Label: Use a Topology-Identifier
label on top of the IP packet. This is very simple. If label on top of the IP packet. This is very simple. If
tunneling to a next-next-hop is desired, then a two-deep tunneling to a next-next-hop is desired, then a two-deep
label stack can be used with [ Topology-ID label, Next-Next- label stack can be used with [ Topology-ID label, Next-Next-
Hop Label ]. Hop Label ].
2. Tunnel IP packets in IP. Each router supporting this option 2. Tunnel IP packets in IP. Each router supporting this option
would announce two additional loopback addresses and their would announce two additional loopback addresses and their
associated MRT color. Those addresses are used as destination associated MRT color. Those addresses are used as destination
addresses for MRT-blue and MRT-red IP tunnels respectively. They addresses for MRT-blue and MRT-red IP tunnels respectively. They
allow the transit nodes to identify the traffic as being allow the transit nodes to identify the traffic as being
forwarded along either MRT-blue or MRT-red tree topology to reach forwarded along either MRT-blue or MRT-red tree topology to reach
the tunnel destination. Announcements of these two additional the tunnel destination. Announcements of these two additional
loopback addresses per router with their MRT color requires IGP loopback addresses per router with their MRT color requires IGP
extensions. extensions.
For greatest hardware compatibility and ease in removing the MRT- 7. Protocol Extensions and Considerations: OSPF and ISIS
topology marking at area/level boundaries, routers that support MPLS
and implement IP MRT fast-reroute SHOULD support Option A - using an
LDP label that indicates the destination and MT-ID.
For proxy-nodes associated with one or more multi-homed prefixes,
there is no router associated with the proxy-node, so its loopbacks
can't be known or used. Instead, the loopback addresses of the
routers that are attached to the proxy-node can be used. One of
those routers will be on the Red MRT and the other on the Blue MRT.
The MRT-red loopback of the first router would be used to reach the
router on the Red MRT and similarly the MRT-blue loopback of the
second router would be used. The routers connected to the proxy-node
are the end of the area/level and can decapsulate the traffic and
properly forward it into the next area.
6. Protocol Extensions and Considerations: OSPF and ISIS For simplicity, the approach of defining a well-known profile is
taken in [I-D.atlas-ospf-mrt]. The purpose of communicating support
for MRT in the IGP is to indicate thatqq the MRT-Blue and MRT-Red
forwarding topologies are created for transit traffic. This section
describes the various options to be selected. The default MRT
profile is described here and the signaling extensions for OSPF are
given in [I-D.atlas-ospf-mrt].
There are two possible approaches to what additional information to For any MRT profile, the MRT Island is created by starting from the
distribute in the IGP. The first is to allow full flexibility in all computing router. If the computing router supports the default MRT
information and distribute whichever values and combinations are profile, add it to the MRT Island. Add a router to the MRT Island if
desired. The second is to simply distribute flags indicating a the router supports the default MRT profile and is connected to the
particular well-known profile is supported. Thus the MRT Island MRT Island via bidirectional links eligible for MRT.
Creation process is trivial. The profile approach is recommended,
with the added flexibility of being able to specify more specific
information if necessary and supported.
For example, a simple profile "metric-insensitive MRT unicast fast- If a router advertises support for multiple MRT profiles, then it
reroute via LDP" could specify: MUST create the transit forwarding topologies for each of those,
unless the profile specifies No Forwarding Mechanism (e.g. as might
be done for a profile used only for multicast global protection). A
router MUST NOT advertise multiple MRT profiles that overlap in their
MRT-Red MT-ID or MRT-Blue MT-ID.
MRT Island Creation: Only include other routers advertising this The MRT Profile also defines different behaviors such as how MRT
profile. recomputation is handled and how area/level boundaries are dealt
with.
MRT Algorithm ID: The MRT Lowpoint algorithm defined in MRT Algorithm: MRT Lowpoint algorithm defined in
[I-D.enyedi-rtgwg-mrt-frr-algorithm]. [I-D.enyedi-rtgwg-mrt-frr-algorithm].
Red MRT MT-ID: The Red MRT MT-ID is the single well-known value MRT-Red MT-ID: experimental 3997, final value assigned by IANA
allocated by IANA from the OSPF, ISIS, LDP and PIM MT-ID spaces. allocated from the LDP MT-ID space
Blue MRT MT-ID: The Blue MRT MT-ID is the single well-known value MRT-Blue MT-ID: experimental 3998, final value assigned by IANA
allocated by IANA from the OSPF, ISIS, LDP and PIM MT-ID spaces. allocated from the LDP MT-ID space
GADAG Root Election Priority: Pick the router with the lowest GADAG Root Selection Priority: Among the routers in the MRT Island
router ID to be the GADAG root. and with the highest priority advertised, an implementation MUST
pick the router with the highest Router ID to be the GADAG root.
Forwarding Mechanisms for IP: Use IP-in-LDP. Forwarding Mechanisms: LDP
MRT Capabilities: Computes MRTs, IP Fast-Reroute, LDP Fast-Reroute Recalculation: Recalculation of MRTs SHOULD occur as described in
Section 11.2. This allows the MRT forwarding topologies to
support IP/LDP fast-reroute traffic.
The following captures an initial understanding of the aspects that Area/Level Border Behavior: As described in Section 9, ABRs/LBRs
must be considered to fully form a profile to advertise. For some SHOULD ensure that traffic leaving the area also exits the MRT-Red
profiles, associated information may need to be distributed, such as or MRT-Blue forwarding topology.
GADAG Root Election Priority, Red MRT Loopback Address, Blue MRT
Loopback Address, or MRT Algorithm ID.
MRT Island Creation ID: This identifies the process that the router The following describes the aspects to be considered to define a
uses to form an MRT Island. By advertising an ID for the process, profile to advertise. For some profiles, associated information may
it is possible to have different processes in the future. It may need to be distributed, such as GADAG Root Selection Priority, Red
be desirable to advertise a list ordered by preference to allow MRT Loopback Address, Blue MRT Loopback Address.
transitions.
MRT Algorithm ID: This identifies the particular MRT algorithm used MRT Algorithm: This identifies the particular MRT algorithm used by
by the router. By having an Algorithm ID, it is possible to the router for this profile. Algorithm transitions can be managed
change the algorithm used or use different ones in different by advertising multiple MRT profiles.
networks. It may be desirable to advertise a list ordered by
preference to allow transitions.
Red MRT MT-ID: This specifies the MT-ID to be associated with the MRT-Red MT-ID: This specifies the MT-ID to be associated with the
Red MRT forwarding topology. It is needed for use in signaling. MRT-Red forwarding topology. It is needed for use in LDP
All routers in the MRT Island MUST agree on a value. signaling. All routers in the MRT Island MUST agree on a value.
Blue MRT MT-ID: This specifies the MT-ID to be associated with the MRT-Blue MT-ID: This specifies the MT-ID to be associated with the
Blue MRT forwarding topology. It is needed for use in signaling. MRT-Blue forwarding topology. It is needed for use in LDP
All routers in the MRT Island MUST agree on a value. signaling. All routers in the MRT Island MUST agree on a value.
GADAG Root Election Priority: This specifies the priority of the GADAG Root Selection Priority: A MRT profile might specify this to
router for being used as the GADAG root of its island. A GADAG provide the network operator with a knob to force a particular
root is elected from the set of routers with the highest priority; GADAG root selection. If not specified in the MRT profile, the
ties are broken based upon highest Router ID. The sensitivity of highest Router ID in the profile's MRT Island will be elected the
the MRT Algorithms to GADAG root selection is still being GADAG Root. If a GADAG Root Selection Priority is specified, then
evaluated. This provides the network operator with a knob to the MRT profile must also specify how the GADAG Root is elected.
force particular GADAG root selection.
Forwarding Mechanism for IP: This specifies which forwarding Forwarding Mechanism: This specifies which forwarding mechanisms
mechanisms the router supports for IP traffic. An MRT island must the router supports for transit traffic. An MRT island must
support a common set of forwarding mechanisms, which may be less program appropriate next-hops into the forwarding plane. The
than the full set advertised. Multiple forwarding mechanisms may known options are IPv4, IPv6, LDP, and None. If IPv4 is
be specified, such as IP-in-IPv4, IP-in-IPv6 or IP-in-LDP Label. supported, then both MRT-Red and MRT-Blue IPv4 Loopback Addresses
None is also an option. SHOULD be specified. If IPv6 is supported, both MRT-Red and MRT-
Blue IPv6 Loopback Addresses SHOULD be specified. If LDP is
supported, then LDP support and signaling extensions MUST be
supported.
Red MRT Loopback Address: This provides the router's loopback MRT-Red Loopback Address: This provides the router's loopback
address to reach the router via the Red MRT forwarding topology. address to reach the router via the MRT-Red forwarding topology.
It can, of course, be specified for both IPv4 and IPv6. It can, of course, be specified for both IPv4 and IPv6.
Blue MRT Loopback Address: This provides the router's loopback MRT-Blue Loopback Address: This provides the router's loopback
address to reach the router via the Blue MRT forwarding topology. address to reach the router via the MRT-Blue forwarding topology.
It can, of course, be specified for both IPv4 and IPv6. It can, of course, be specified for both IPv4 and IPv6.
MRT Capabilities Available: This is the set of capabilities that Recalculation: As part of what process and timing should the new
the router is configured to support. MRTs be computed on a modified topology? Section 11.2 describes
the minimum behavior required to support fast-reroute.
MRT Capabilities Required: This is the set of capabilities that
other routers must have available to be added into the MRT island.
MRT Capability: Computes MRTs: The router can compute MRTs.
MRT Capability: IP Fast-Reroute: The router can use the computed
MRTs for IP fast-reroute.
MRT Capability: LDP Fast-Reroute: The router can use the computed
MRTs for LDP fast-reroute.
MRT Capability: PIM Fast-Reroute: The router can use the computed
MRTs for PIM fast-reroute.
MRT Capability: mLDP Fast-Reroute: The router can use the computed
MRTs for mLDP fast-reroute.
MRT Capability: PIM Global Protection: The router can use the
computed MRTs for PIM Global Protection 1+1.
MRT Capability: mLDP Global Protection: The router can use the Area/Level Border Behavior: Should inter-area traffic on the MRT-
computed MRTs for mLDP Global Protection 1+1. Blue or MRT-Red be put back onto the shortest path tree? Should
it be swapped from MRT-Blue or MRT-Red in one area/level to MRT-
Red or MRT-Blue in the next area/level to avoid the potential
failure of an ABR? (See [I-D.atlas-rtgwg-mrt-mc-arch] for use-
case details.
The assumption is that a router will form an MRT island, compute MRTs Other Profile-Specific Behavior: Depending upon the use-case for
within that island, and then use those MRTs for the purposes the profile, there may be additional profile-specific behavior.
specified in the profile. If multiple profiles are supported with
different purposes (e.g. mLDP Global Protection), then the router may
use a different profile and associated MRT island to be used for the
purposes in that different profile. If a router wanted to form
multiple MRT islands for different application purposes, that could
be done by specifying different Red MRT MT-ID and Blue MRT MT-IDs.
As with LFA, it is expected that OSPF Virtual Links will not be As with LFA, it is expected that OSPF Virtual Links will not be
supported. supported.
7. Protocol Extensions and considerations: LDP 8. Protocol Extensions and considerations: LDP
The protocol extensions for LDP are defined in
Capability negotiation in LDP is needed to indicate support for MRT; [I-D.atlas-mpls-ldp-mrt]. A router must indicate that it has the
having this explicit allows the use of MRT-specific signaling ability to support MRT; having this explicit allows the use of MRT-
extensions. A router also needs to indicate, via FEC advertisement, specific processing, such as special handling of FECs sent with the
whether it supports LDP Destination-Topology Labels, LDP Topology Rainbow MRT MT-ID.
Labels, or both. Since the label or labels are swapped at each LSR,
consistency across the network is not required.
If both mechanisms are supported, then if a Destination-Topology
label is provided for a FEC, that should be used so that an ABR/LBR
can indicate the appropriate labels, as discussed in Section
Section 9.
8. Multi-homed Prefixes
One advantage of LFAs that is necessary to preserve is the ability to
protect multi-homed prefixes against ABR failure. For instance, if a
prefix from the backbone is available via both ABR A and ABR B, if A
fails, then the traffic should be redirected to B. This can also be
done for backups via MRT.
This generalizes to any multi-homed prefix. A multi-homed prefix
could be:
o An out-of-area prefix announced by more than one ABR,
o An AS-External route announced by 2 or more ASBRs,
o A prefix with iBGP multipath to different ASBRs,
o etc.
For each prefix, the attached ABRs are selected and a proxy-node is
created connected to those ABRs. If there exist multiple multi-homed
prefixes that share the same connectivity and costs to each of those
ABRs, then a single proxy-node can be used to represent the set. An
example of this is shown in Figure 3.
2 2 2 2
A----B----C A----B----C
2 | | 2 2 | | 2
| | | |
[ABR1] [ABR2] [ABR1] [ABR2]
| | | |
p,10 p,15 10 |---[P]---| 15
(a) Initial topology (b)with proxy-node
A<---B<---C A--->B--->C
| ^ ^ |
V | | V
[ABR1] [ABR2] [ABR1] [ABR2]
| |
|-->[P] [P]<--|
(c) Blue MRT (d) Red MRT
Figure 3: Prefixes Advertised by Multiple ABRs
The proxy-nodes and associated links are added to the network
topology after all real links have been assigned to a direction and
before the actual MRTs are computed. Proxy-nodes cannot be transited
when computing the MRTs. In addition to computing the pair of MRTs
associated with each router destination D in the area, a pair of MRTs
can be computed for each such proxy-node to fully protect against ABR
failure.
Each ABR or attaching router must remove the MRT marking[see
Section 5] and then forward the traffic outside of the area (or
island of MRT-fast-reroute-supporting routers).
If ASBR protection is desired, this has additonal complexities if the A FEC sent with the Rainbow MRT MT-ID indicates that the FEC applies
ASBRs are in different areas. Similarly, protecting labeled BGP to all the MRT-Blue and MRT-Red MT-IDs in supported MRT profiles as
traffic in the event of an ASBR failure has additional complexities well as to the default shortest-path based MT-ID 0. The Rainbow MRT
due to the per-ASBR label spaces involved. MT-ID is defined to provide an easy way to handle the special
signaling that is needed at ABRs or LBRs. It avoids the problem of
needing to signal different MPLS labels for the same FEC. Because
the Rainbow MRT MT-ID is used only by ABRs/LBRs or the LDP egress, it
is not MRT profile specific. The proposed experimental value is 3999
and the final value will be assigned by IANA and allocated from the
LDP MT-ID space. The authoritative values are given in
[I-D.atlas-mpls-ldp-mrt].
9. Inter-Area and ABR Forwarding Behavior 9. Inter-Area and ABR Forwarding Behavior
In regular forwarding, packets destined outside the area arrive at An ABR/LBR has two forwarding roles. First, it forwards traffic
the ABR and the ABR forwards them into the other area because the inside its area. Second, it forwards traffic from one area into
next-hops from the area with the best route (according to tie- another. These same two roles apply for MRT transit traffic.
breaking rules) are used by the ABR. The question is then what to do Traffic on MRT-Red or MRT-Blue destined inside the area needs to stay
with packets marked with an MRT that are received by the ABR. on MRT-Red or MRT-Blue in that area. However, it is desirable for
traffic leaving the area to also exit MRT-Red or MRT-Blue back to the
shortest-path forwarding.
For unicast fast-reroute, the need to stay on an MRT forwarding For unicast MRT-FRR, the need to stay on an MRT forwarding topology
topology terminates at the ABR/LBR whose best route is via a terminates at the ABR/LBR whose best route is via a different area/
different area/level. It is highly desirable to go back to the level. It is highly desirable to go back to the default forwarding
default forwarding topology when leaving an area/level. There are topology when leaving an area/level. There are three basic reasons
three basic reasons for this. First, the default topology uses for this. First, the default topology uses shortest paths; the
shortest paths; the packet will thus take the shortest possible route packet will thus take the shortest possible route to the destination.
to the destination. Second, this allows failures that might appear Second, this allows failures that might appear in multiple areas
in multiple areas (e.g. ABR/LBR failures) to be separately (e.g. ABR/LBR failures) to be separately identified and repaired
identified and repaired around. Third, the packet can be fast- around. Third, the packet can be fast-rerouted again, if necessary,
rerouted again, if necessary, due to a failure in a different area. due to a failure in a different area.
An ABR/LBR that receives a packet marked with an MRT towards a An ABR/LBR that receives a packet on MRT-Red or MRT-Blue towards a
destination in another area/level should forward the MRT marked destination in another area/level should forward the packet in the
packet in the area/level with the best route along its associated area/level with the best route along MRT-Red or MRT-Blue. If the
MRT. If the packet came from that area/level, this correctly avoids packet came from that area/level, this correctly avoids the failure.
the failure. However, if the traffic came from a different area/level, the packet
should be removed from MRT-Red or MRT-Blue and forwarded on the
shortest-path default forwarding topology.
How does an ABR/LBR ensure that MRT-marked packets do not arrive at To avoid per-interface forwarding state for MRT-Red and MRT-Blue, the
the ABR/LBR? There are two different mechanisms depending upon the ABR/LBR needs to arrange that packets destined to a different area
forwarding mechanism being used. arrive at the ABR/LBR already not marked as MRT-Red or MRT-Blue.
If the LDP label encodes the MT-ID as well as the destination, then For LDP forwarding where the MPLS label specifies (MT-ID, FEC), the
the ABR/LBR is responsible for advertising a particular label to each ABR/LBR is responsible for advertising the proper label to each
neighbor. Additionally, an LDP label is associated with an MT-ID due neighbor. Assume that an ABR/LBR has allocated three labels for a
to the MT FEC that was used and not due to any intrisic particular particular destination; those labels are L_primary, L_blue, and
value for the label. Assume that an ABR/LBR has allocated three L_red. When the ABR/LBR advertises label bindings to routers in the
labels for a particular destination; those labels are L_primary, area with the best route to the destination, the ABR/LBR provides
L_blue, and L_red. When the ABR/LBR advertises label bindings to L_primary for the default topology, L_blue for the MRT-Blue MT-ID and
routers in the area with the best route to the destination, the ABR/ L_red for the MRT-Red MT-ID, exactly as expected. However, when the
LBR provides L_primary for the default topology, L_blue for the Blue ABR/LBR advertises label bindings to routers in other areas, the ABR/
MRT MT-ID and L_red for the Red MRT MT-ID, exactly as expected. LBR advertises L_primary for the Rainbow MRT MT-ID, which is then
However, when the ABR/LBR advertises label bindings to routers in used for the default topology, for the MRT-Blue MT-ID and for the
other areas, the ABR/LBR advertises L_primary for the default MRT-Red MT-ID.
topology, for the Blue MRT MT-ID, and for the Red MRT MT-ID. The
ABR/LBR installs next-hops from the best area for L_primary based on
the default topology, for L_blue based on the Blue MRT forwarding
topology, and for L_red based on the Red MRT forwarding topology.
Therefore, packets from the non-best area will arrive at the ABR/LBR
with a label L_primary and will be forwarded into the best area along
the default topology. By controlling what labels are advertised, the
ABR/LBR can thus enforce that packets exiting the area do so on the
shortest-path default topology.
If IP-in-IP forwarding is used, then the ABR/LBR behavior is The ABR/LBR installs all next-hops from the best area: primary next-
dependent upon the outermost IP address. If the outermost IP address hops for L_primary, MRT-Blue next-hops for L_blue, and MRT-Red next-
is an MRT loopback address of the ABR/LBR, then the packet is hops for L_red. Because the ABR/LBR advertised (Rainbow MRT MT-ID,
decapsulated and forwarded based upon the inner IP address, which FEC) with L_primary to neighbors not in the best area, packets from
should go on the default SPT topology. If the outermost IP address those neighbors will arrive at the ABR/LBR with a label L_primary and
is not an MRT loopback address of the ABR/LBR, then the packet is will be forwarded into the best area along the default topology. By
simply forwarded along the associated forwarding topology. A PLR controlling what labels are advertised, the ABR/LBR can thus enforce
sending traffic to a destination outside its local area/level will that packets exiting the area do so on the shortest-path default
pick the MRT and use the associated MRT loopback address of the ABR/ topology.
LBR immediately before the proxy-node on that MRT.
If IP forwarding is used, then the ABR/LBR behavior is dependent upon
the outermost IP address. If the outermost IP address is an MRT
loopback address of the ABR/LBR, then the packet is decapsulated and
forwarded based upon the inner IP address, which should go on the
default SPT topology. If the outermost IP address is not an MRT
loopback address of the ABR/LBR, then the packet is simply forwarded
along the associated forwarding topology. A PLR sending traffic to a
destination outside its local area/level will pick the MRT and use
the associated MRT loopback address of the selected ABR/LBR connected
to the external destination.
Thus, regardless of which of these two forwarding mechanisms are Thus, regardless of which of these two forwarding mechanisms are
used, there is no need for additional computation or per-area used, there is no need for additional computation or per-area
forwarding state. forwarding state.
+----[C]---- --[D]--[E] --[D]--[E] +----[C]---- --[D]--[E] --[D]--[E]
| \ / \ / \ | \ / \ / \
p--[A] Area 10 [ABR1] Area 0 [H]--p +-[ABR1] Area 0 [H]-+ p--[A] Area 10 [ABR1] Area 0 [H]--p +-[ABR1] Area 0 [H]-+
| / \ / | \ / | | / \ / | \ / |
+----[B]---- --[F]--[G] | --[F]--[G] | +----[B]---- --[F]--[G] | --[F]--[G] |
skipping to change at page 18, line 38 skipping to change at page 17, line 32
/ \ / \ / \ / \
[ABR1] Area 0 [H]-+ +-[ABR1] [H] [ABR1] Area 0 [H]-+ +-[ABR1] [H]
/ | | \ / | | \
[F]->[G] V V -<[F]<-[G] [F]->[G] V V -<[F]<-[G]
| | | |
| | | |
[p]<------+ +--------->[p] [p]<------+ +--------->[p]
(d) Blue MRT in Area 0 (e) Red MRT in Area 0 (d) Blue MRT in Area 0 (e) Red MRT in Area 0
Figure 4: ABR Forwarding Behavior and MRTs Figure 3: ABR Forwarding Behavior and MRTs
The other potential forwarding mechanisms require additional The other forwarding mechanism described in Section 6 is using
computation by the penultimate router along the in-local-area MRT Topology-Identification Labels. This mechanism would require that
immediately before the ABR/LBR is reached. The penultimate router any router whose MRT-Red or MRT-Blue next-hop is an ABR/LBR would
can determine that the ABR/LBR will forward the packet out of area/ need to determine whether the ABR/LBR would forward the packet out of
level and, in that case, the penultimate router can remove the MRT the area/level. If so, then that router should pop off the topology-
marking but still forward the packet along the MRT next-hop to reach identification label before forwarding the packet to the ABR/LBR.
the ABR. For instance, in Figure 4, if node H fails, node E has to
put traffic towards prefix p onto the red MRT. But since node D For example, in Figure 3, if node H fails, node E has to put traffic
knows that ABR1 will use a best from another area, it is safe for D towards prefix p onto MRT-Red. But since node D knows that ABR1 will
to remove the MRT marking and just send the packet to ABR1 still on use a best from another area, it is safe for D to pop the Topology-
the red MRT but unmarked. ABR1 will use the shortest path in Area Identification Label and just forward the packet to ABR1 along the
10. MRT-Red next-hop. ABR1 will use the shortest path in Area 10.
In all cases for ISIS and most cases for OSPF, the penultimate router In all cases for ISIS and most cases for OSPF, the penultimate router
can determine what decision the adjacent ABR will make. The one case can determine what decision the adjacent ABR will make. The one case
where it can't be determined is when two ASBRs are in different non- where it can't be determined is when two ASBRs are in different non-
backbone areas attached to the same ABR, then the ASBR's Area ID may backbone areas attached to the same ABR, then the ASBR's Area ID may
be needed for tie-breaking (prefer the route with the largest OPSF be needed for tie-breaking (prefer the route with the largest OPSF
area ID) and the Area ID isn't announced as part of the ASBR link- area ID) and the Area ID isn't announced as part of the ASBR link-
state advertisement (LSA). In this one case, suboptimal forwarding state advertisement (LSA). In this one case, suboptimal forwarding
along the MRT in the other area would happen. If this is a realistic along the MRT in the other area would happen. If that becomes a
deployment scenario, OSPF extensions could be considered. realistic deployment scenario, OSPF extensions could be considered.
This is not covered in [I-D.atlas-ospf-mrt].
10. Issues with Area Abstraction 10. Prefixes Multiply Attached to the MRT Island
MRT fast-reroute provides complete coverage in a area that is How a computing router S determines its local MRT Island for each
2-connected. Where a failure would partition the network, of course, supported MRT profile is already discussed in Section 7.
no alternate can protect against that failure. Similarly, there are
ways of connecting multi-homed prefixes that make it impractical to
protect them without excessive complexity.
50 There are two types of prefixes or FECs that may be multiply attached
|----[ASBR Y]---[B]---[ABR 2]---[C] Backbone Area 0: to an MRT Island. The first type are multi-homed prefixes that
| | ABR 1, ABR 2, C, D usually connect at a domain or protocol boundary. The second type
| | represent routers that do not support the profile for the MRT Island.
| | Area 20: A, ASBR X The key difference is whether the traffic, once out of the MRT
| | Island, remains in the same area/level and might reenter the MRT
p ---[ASBR X]---[A]---[ABR 1]---[D] Area 10: B, ASBR Y Island if a loop-free exit point is not selected.
5 p is a Type 1 AS-external
Figure 5: AS external prefixes in different areas One property of LFAs that is necessary to preserve is the ability to
protect multi-homed prefixes against ABR failure. For instance, if a
prefix from the backbone is available via both ABR A and ABR B, if A
fails, then the traffic should be redirected to B. This can also be
done for backups via MRT.
Consider the network in Figure 5 and assume there is a richer If ASBR protection is desired, this has additonal complexities if the
connective topology that isn't shown, where the same prefix is ASBRs are in different areas. Similarly, protecting labeled BGP
announced by ASBR X and ASBR Y which are in different non-backbone traffic in the event of an ASBR failure has additional complexities
areas. If the link from A to ASBR X fails, then an MRT alternate due to the per-ASBR label spaces involved.
could forward the packet to ABR 1 and ABR 1 could forward it to D,
but then D would find the shortest route is back via ABR 1 to Area
20. The only real way to get it from A to ASBR Y is to explicitly
tunnel it to ASBR Y.
Tunnelling to the backup ASBR is for future consideration. The As discussed in [RFC5286], a multi-homed prefix could be:
previously proposed PHP approach needs to have an exception if BGP
policies (e.g. BGP local preference) determines which ASBR to use.
Consider the case in Figure 6. If the link between A and ASBR X (the
preferred border router) fails, A can put the packets to p onto an
MRT alternate, even tunnel it towards ASBR Y. Node B, however, must
not remove the MRT marking in this case, as nodes in Area 0,
including ASBR Y itself would not know that their preferred ASBR is
down.
Area 20 BB Area 0 o An out-of-area prefix announced by more than one ABR,
p ---[ASBR X]-X-[A]---[B]---[ABR 1]---[D]---[ASBR Y]--- p
BGP prefers ASBR X for prefix p o An AS-External route announced by 2 or more ASBRs,
Figure 6: Failure of path towards ASBR preferred by BGP o A prefix with iBGP multipath to different ASBRs,
The fine details of how to solve multi-area external prefix cases, or o etc.
identifying certain cases as too unlikely and too complex to protect
is for further consideration.
11. Partial Deployment and Islands of Compatible MRT FRR routers There are also two different approaches to protection. The first is
to do endpoint selection to pick a router to tunnel to where that
router is loop-free with respect to the failure-point. Conceptually,
the set of candidate routers to provide LFAs expands to all routers,
with an MRT alternate, attached to the prefix.
A natural concern with new functionality is how to have it be useful The second is to use a proxy-node, that can be named via MPLS label
when it is not deployed across an entire IGP area. In the case of or IP address, and pick the appropriate label or IP address to reach
MRT FRR, where it provides alternates when appropriate LFAs aren't it on either MRT-Blue or MRT-Red as appropriate to avoid the failure
available, there are also deployment scenarios where it may make point. A proxy-node can represent a destination prefix that can be
sense to only enable some routers in an area with MRT FRR. A simple attached to the MRT Island via at least two routers. It is termed a
example of such a scenario would be a ring of 6 or more routers that named proxy-node if there is a way that traffic can be encapsulated
is connected via two routers to the rest of the area. to reach specifically that proxy-node; this could be because there is
an LDP FEC for the associated prefix or because MRT-Red and MRT-Blue
IP addresses are advertised in an as-yet undefined fashion for that
proxy-node. Traffic to a named proxy-node may take a different path
than traffic to the attaching router; traffic is also explicitly
forwarded from the attaching router along a predetermined interface
towards the relevant prefixes.
First, a computing router S must determine its local island of For IP traffic, multi-homed prefixes can use endpoint selection. For
compatible MRT fast-reroute routers. A router that has a common IP traffic that is destined to a router outside the MRT Island, if
profile flag and is connected either to S or to another router that router is the egress for a FEC advertised into the MRT Island,
already determined to be in S's local island can be added to S's then the named proxy-node approach can be used.
local island.
Destinations inside the local island can obviously use MRT For LDP traffic, there is always a FEC advertised into the MRT
alternates. Destinations outside the local island can be treated Island. The named proxy-node approach should be used, unless the
like a multi-homed prefix with caveats to avoid looping. For LDP computing router S knows the label for the FEC at the selected
labels including both destination and topology, the routers at the endpoint.
borders of the local island need to originate labels for the original
FEC and the associated MRT-specific labels. Packets sent to an LDP
label marked as blue or red MRT to a destination outside the local
island will have the last router in the local island swap the label
to one for the destination and forward the packet along the outgoing
interface on the MRT towards a router outside the local island that
was represented by the proxy-node.
For IP in IP encapsulations, remote destinations' loopback addresses If a FEC is advertised from outside the MRT Island into the MRT
for the MRTs cannot be used, even if they were available. Instead, Island and the forwarding mechanism specified in the profile includes
the MRT loopback address of the router attached to a proxy-node, LDP, then the routers learning that FEC MUST also advertise labels
which represents destinations outside the local island, can be used. for (MRT-Red, FEC) and (MRT-Blue, FEC) to neighbors inside the MRT
Packets sent to the router's MRT loopback address will have their Island. If the forwarding mechanism includes LDP, any router
outer IP header removed and will need to be explicitly forwarded receiving a FEC corresponding to a router outside the MRT Island or
along the outgoing interface on the MRT towards a router outside the to a multi-homed prefix MUST compute and install the transit MRT-Blue
local island that was represented by the proxy-node. This behavior and MRT-Red next-hops for that FEC; the associated FECs ( (MT-ID 0,
requires essentially remembering the MT-ID indicated by the outer IP FEC), (MRT-Red, FEC), and (MRT-Blue, FEC)) MUST also be provided via
address. An alternate option would be to advertise different LDP to neighbors inside the MRT Island.
loopback addresses to be associated with the proxy-node; the outer IP
address would still be removed but it would indicate the outgoing
interface to use and no lookup would be necessary on the internal IP
address while maintaining MT-ID context.
A key question is which routers outside the MRT island can packets be 10.1. Endpoint Selection
forwarded to so that they are not forwarded back into the MRT island.
An example of the necessary network graph transformations are given
in Figure 7. There are two parts to the computation. First, the MRT
island is collapsed into a single node; this assumes that the cost of
transiting the MRT island is nothing and is pessimistic but allows
for simpler computation. Then, for each destination (other than the
MRT island), the routers adjacent to the MRT island are checked to
see if they are loop-free with respect to the MRT island and the
destination. The loop-free neighbors of the MRT island that are
closest to the destination are selected. Then, a graph of just the
MRT island is augmented with proxy-nodes that are attached via the
outgoing interfaces to the selected loop-free neighbors. Finally,
the MRTs rooted at each proxy-node are computed on that augmented MRT
island graph. Essentially, the MRT island must have a loop-free
neighbor to be able to have an alternate.
[G]---[E]---(B)---(C)---(D) Endpoint Selection is a local matter for a router in the MRT Island
| \ | | | since it pertains to selecting and using an alternate and does not
| \ | | | affect the transit MRT-Red and MRT-Blue forwarding topologies.
| \ | | |
[H]---[F]---(A)---(S)----|
(1) Network Graph with Partial Deployment Let the computing router be S and the next-hop F be the node whose
failure is to be avoided. Let the destination be prefix p. Have A
be the router to which the prefix p is attached for S's shortest path
to p.
[E],[F],[G],[H] : No support for MRT-FRR The candidates for endpoint selection are those to which the
(A),(B),(C),(D),(S): MRT Island - supports MRT-FRR destination prefix is attached in the area/level. For a particular
candidate B, it is necessary to determine if B is loop-free to reach
p with respect to S and F for node-protection or at least with
respect to S and the link (S, F) for link-protection. If B will
always prefer to send traffic to p via a different area/level, then
this is definitional. Otherwise, distance-based computations are
necessary and an SPF from B's perspective may be necessary. The
following equations give the checks needed; the rationale is similar
to that given in [RFC5286].
[G]---[E]----| |---(B)---(C)---(D) Loop-Free for S: D_opt(B, p) < D_opt(B, S) + D_opt(S, p)
| \ | | | | |
| \ | ( MRT Island ) [ proxy ] | |
| \ | | | | |
[H]---[F]----| |---(A)---(S)----|
(2) Graph for determining (3) Graph for MRT computation Loop-Free for F: D_opt(B, p) < D_opt(B, F) + D_opt(F, p)
loop-free neighbors
Figure 7: Computing alternates to destinations outside the MRT Island The latter is equivalent to the following, which avoids the need to
compute the shortest path from F to p.
Loop-Free for F: D_opt(B, p) < D_opt(B, F) + D_opt(S, p) - D_opt(S,
F)
Finally, the rules for Endpoint selection are given below. The basic
idea is to repair to the prefix-advertising router selected for the
shortest-path and only to select and tunnel to a different endpoint
if necessary (e.g. A=F or F is a cut-vertex or the link (S,F) is a
cut-link).
1. Does S have a node-protecting alternate to A? If so, select
that. Tunnel the packet to A along that alternate. For example,
if LDP is the forwarding mechanism, then push the label (MRT-Red,
A) or (MRT-Blue, A) onto the packet.
2. If not, then is there a router B that is loop-free to reach p
while avoiding both F and S? If so, select B as the end-point.
Determine the MRT alternate to reach B while avoiding F. Tunnel
the packet to B along that alternate. For example, with LDP,
push the label (MRT-Red, B) or (MRT-Blue, B) onto the packet.
3. If not, then does S have a link-protecting alternate to A? If
so, select that.
4. If not, then is there a router B that is loop-free to reach p
while avoiding S and the link from S to F? If so, select B as
the endpoint and the MRT alternate that for reaching B from S
avoiding the link (S,F).
The endpoint selected will receive a packet destined to itself and,
being the egress, will pop that MPLS label (or have signaled Implicit
Null) and forward based on what is underneath. This suffices for IP
traffic where the MPLS labels understood by the endpoint router are
not needed.
10.2. Named Proxy-Nodes
A clear advantage to using a named proxy-node is that it is possible
to explicitly forward from the MRT Island along an interface to a
loop-free island neighbor (LFIN) when that interface may not be a
primary next-hop. For LDP traffic where the label indicates both the
topology and the FEC, it is necessary to either use a named proxy-
node or deal with learning remote MPLS labels.
A named proxy-node represents one or more destinations and, for LDP
forwarding, has a FEC associated with it that is signaled into the
MRT Island. Therefore, it is possible to explicitly label packets to
go to (MRT-Red, FEC) or (MRT-Blue, FEC); at the border of the MRT
Island, the label will swap to meaning (MT-ID 0, FEC). It would be
possible to have named proxy-nodes for IP forwarding, but this would
require extensions to signal two IP addresses to be associated with
MRT-Red and MRT-Blue for the proxy-node. A named proxy-node can be
uniquely represented by the two routers in the MRT Island to which it
is connected. The extensions to signal such IP addresses are not
defined in [I-D.atlas-ospf-mrt]. The details of what label-bindings
must be originated are described in [I-D.atlas-mpls-ldp-mrt].
Computing the MRT next-hops to a named proxy-node and the MRT
alternate for the computing router S to avoid a particular failure
node F is extremely straightforward. The details of the simple
constant-time functions, Select_Proxy_Node_NHs() and
Select_Alternates_Proxy_Node(), are given in
[I-D.enyedi-rtgwg-mrt-frr-algorithm]. A key point is that computing
these MRT next-hops and alternates can be done as new named proxy-
nodes are added or removed without requiring a new MRT computation or
impacting other existing MRT paths. This maps very well to, for
example, how OSPFv2 [[RFC2328] Section 16.5] does incremental updates
for new summary-LSAs.
The key question is how to attach the named proxy-node to the MRT
Island; all the routers in the MRT Island MUST do this consistently.
No more than 2 routers in the MRT Island can be selected; one should
only be selected if there are no others that meet the necessary
criteria. The named proxy-node is logically part of the area/level.
There are two sources for candidate routers in the MRT Island to
connect to the named proxy-node. The first set are those routers
that are advertising the prefix; the cost assigned to each such
router is the announced cost to the prefix. The second set are those
routers in the MRT Island that are connected to routers not in the
MRT Island but in the same area/level; such routers will be defined
as Island Border Routers (IBRs). The routers connected to the IBRs
that are not in the MRT Island and are in the same area/level are
Island Neighbors (INs).
Since packets sent to the named proxy-node along MRT-Red or MRT-Blue
may come from any router inside the MRT Island, it is necessary that
whatever router to which an IBR forwards the packet be loop-free with
regard to the whole MRT Island for the destination. Thus, an IBR is
a candidate router only if it possesses at least one IN whose path to
the prefix does not enter the MRT Island. The cost assigned to each
(IBR, IN) pair is the D_opt(IN, prefix) plus Cost(IBR, IN).
From the set of prefix-advertising routers and the IBRs, the two
lowest cost routers are selected and ties are broken based upon the
lowest Router ID. For ease of discussion, such selected routers are
proxy-node attachment routers and the two selected will be named A
and B.
A proxy-node attachment router has a special forwarding role. When a
packet is received destined to (MRT-Red, prefix) or (MRT-Blue,
prefix), if the proxy-node attachment router is an IBR, it MUST swap
to the default topology (e.g. swap to the label for (MT-ID 0, prefix)
or remove the outer IP encapsulation) and forward the packet to the
IN whose cost was used in the selection. If the proxy-node
attachment router is not an IBR, then the packet MUST be removed from
the MRT forwarding topology and sent along the interface that caused
the router to advertise the prefix; this interface might be out of
the area/level/AS.
10.2.1. Computing if an Island Neighbor (IN) is loop-free
As discussed, the Island Neighbor needs to be loop-free with regard
to the whole MRT Island for the destination. Conceptually, the cost
of transiting the MRT Island should be regarded as 0. This can be
done by collapsing the MRT Island into a single node, as seen in
Figure 4, and then computing SPFs from each Island Neighbor and from
the MRT Island itself.
[G]---[E]---(V)---(U)---(T)
| \ | | |
| \ | | |
| \ | | |
[H]---[F]---(R)---(S)----|
(1) Network Graph with Partial Deployment
[E],[F],[G],[H] : No support for MRT
(R),(S),(T),(U),(V): MRT Island - supports MRT
[G]---[E]----| |---(V)---(U)---(T)
| \ | | | | |
| \ | ( MRT Island ) [ proxy ] | |
| \ | | | | |
[H]---[F]----| |---(R)---(S)----|
(2) Graph for determining (3) Graph for MRT computation
loop-free neighbors
Figure 4: Computing alternates to destinations outside the MRT Island
The simple way to do this without manipulating the topology is to
compute the SPFs from each IN and a node in the MRT Island (e.g. the
GADAG root), but use a link metric of 0 for all links between routers
in the MRT Island. The distances computed via SPF this way will be
refered to as Dist_mrt0.
An IN is loop-free with respect to a destination D if: Dist_mrt0(IN,
D) < Dist_mrt0(IN, MRT Island Router) + Dist_mrt0(MRT Island Router,
D). Any router in the MRT Island can be used since the cost of
transiting between MRT Island routers is 0. The GADAG Root is
recommended for consistency.
10.3. MRT Alternates for Destinations Outside the MRT Island
A natural concern with new functionality is how to have it be useful
when it is not deployed across an entire IGP area. In the case of
MRT FRR, where it provides alternates when appropriate LFAs aren't
available, there are also deployment scenarios where it may make
sense to only enable some routers in an area with MRT FRR. A simple
example of such a scenario would be a ring of 6 or more routers that
is connected via two routers to the rest of the area.
Destinations inside the local island can obviously use MRT
alternates. Destinations outside the local island can be treated
like a multi-homed prefix and either Endpoint Selection or Named
Proxy-Nodes can be used. Named Proxy-Nodes MUST be supported when
LDP forwarding is supported and a label-binding for the destination
is sent to an IBR.
Naturally, there are more complicated options to improve coverage, Naturally, there are more complicated options to improve coverage,
such as connecting multiple MRT islands across tunnels, but it is not such as connecting multiple MRT islands across tunnels, but the need
clear that the additional complexity is necessary. for the additional complexity has not been justified.
12. Network Convergence and Preparing for the Next Failure 11. Network Convergence and Preparing for the Next Failure
After a failure, MRT detours ensure that packets reach their intended After a failure, MRT detours ensure that packets reach their intended
destination while the IGP has not reconverged onto the new topology. destination while the IGP has not reconverged onto the new topology.
As link-state updates reach the routers, the IGP process calculates As link-state updates reach the routers, the IGP process calculates
the new shortest paths. Two things need attention: micro-loop the new shortest paths. Two things need attention: micro-loop
prevention and MRT re-calculation. prevention and MRT re-calculation.
12.1. Micro-forwarding loop prevention and MRTs 11.1. Micro-forwarding loop prevention and MRTs
As is well known[RFC5715], micro-loops can occur during IGP As is well known[RFC5715], micro-loops can occur during IGP
convergence; such loops can be local to the failure or remote from convergence; such loops can be local to the failure or remote from
the failure. Managing micro-loops is an orthogonal issue to having the failure. Managing micro-loops is an orthogonal issue to having
alternates for local repair, such as MRT fast-reroute provides. alternates for local repair, such as MRT fast-reroute provides.
There are two possible micro-loop prevention mechanism discussed in There are two possible micro-loop prevention mechanisms discussed in
[RFC5715]. The first is Ordered FIB [I-D.ietf-rtgwg-ordered-fib]. [RFC5715]. The first is Ordered FIB [I-D.ietf-rtgwg-ordered-fib].
The second is Farside Tunneling which requires tunnels or an The second is Farside Tunneling which requires tunnels or an
alternate topology to reach routers on the farside of the failure. alternate topology to reach routers on the farside of the failure.
Since MRTs provide an alternate topology through which traffic can be Since MRTs provide an alternate topology through which traffic can be
sent and which can be manipulated separately from the SPT, it is sent and which can be manipulated separately from the SPT, it is
possible that MRTs could be used to support Farside Tunneling. possible that MRTs could be used to support Farside Tunneling.
Details of how to do so are outside of this document. Details of how to do so are outside the scope of this document.
12.2. MRT Recalculation Micro-loop mitigation mechanisms can also work when combined with
MRT.
11.2. MRT Recalculation
When a failure event happens, traffic is put by the PLRs onto the MRT When a failure event happens, traffic is put by the PLRs onto the MRT
topologies. After that, each router recomputes its shortest path topologies. After that, each router recomputes its shortest path
tree (SPT) and moves traffic over to that. Only after all the PLRs tree (SPT) and moves traffic over to that. Only after all the PLRs
have switched to using their SPTs and traffic has drained from the have switched to using their SPTs and traffic has drained from the
MRT topologies should each router install the recomputed MRTs into MRT topologies should each router install the recomputed MRTs into
the FIBs. the FIBs.
At each router, therefore, the sequence is as follows: At each router, therefore, the sequence is as follows:
1. Receive failure notification 1. Receive failure notification
2. Recompute SPT 2. Recompute SPT
3. Install new SPT 3. Install new SPT
4. Recompute MRTs 4. If the network was stable before the failure occured, wait a
configured (or advertised) period for all routers to be using
their SPTs and traffic to drain from the MRTs.
5. Wait configured period for all routers to be using their SPTs and 5. Recompute MRTs
traffic to drain from the MRTs.
6. Install new MRTs. 6. Install new MRTs.
While the recomputed MRTs are not installed in the FIB, protection While the recomputed MRTs are not installed in the FIB, protection
coverage is lowered. Therefore, it is important to recalculate the coverage is lowered. Therefore, it is important to recalculate the
MRTs and install them quickly. MRTs and install them quickly.
13. Acknowledgements 12. Acknowledgements
The authors would like to thank Hannes Gredler, Jeff Tantsura, Ted The authors would like to thank Mike Shand for his valuable review
Qian, Kishore Tiruveedhula, Santosh Esale, Nitin Bahadur, Harish and contributions.
Sitaraman and Raveendra Torvi for their suggestions and review.
14. IANA Considerations The authors would like to thank Joel Halpern, Hannes Gredler, Ted
Qian, Kishore Tiruveedhula, Shraddha Hegde, Santosh Esale, Nitin
Bahadur, Harish Sitaraman, Raveendra Torvi and Chris Bowers for their
suggestions and review.
13. IANA Considerations
This doument includes no request to IANA. This doument includes no request to IANA.
15. Security Considerations 14. Security Considerations
This architecture is not currently believed to introduce new security This architecture is not currently believed to introduce new security
concerns. concerns.
16. References 15. References
16.1. Normative References 15.1. Normative References
[I-D.enyedi-rtgwg-mrt-frr-algorithm] [I-D.enyedi-rtgwg-mrt-frr-algorithm]
Atlas, A., Envedi, G., Csaszar, A., and A. Gopalan, Atlas, A., Envedi, G., Csaszar, A., Gopalan, A., and C.
"Algorithms for computing Maximally Redundant Trees for Bowers, "Algorithms for computing Maximally Redundant
IP/LDP Fast- Reroute", Trees for IP/LDP Fast- Reroute", draft-enyedi-rtgwg-mrt-
draft-enyedi-rtgwg-mrt-frr-algorithm-02 (work in frr-algorithm-03 (work in progress), July 2013.
progress), October 2012.
[RFC5286] Atlas, A. and A. Zinin, "Basic Specification for IP Fast [RFC5286] Atlas, A. and A. Zinin, "Basic Specification for IP Fast
Reroute: Loop-Free Alternates", RFC 5286, September 2008. Reroute: Loop-Free Alternates", RFC 5286, September 2008.
[RFC5714] Shand, M. and S. Bryant, "IP Fast Reroute Framework", [RFC5714] Shand, M. and S. Bryant, "IP Fast Reroute Framework", RFC
RFC 5714, January 2010. 5714, January 2010.
16.2. Informative References 15.2. Informative References
[EnyediThesis] [EnyediThesis]
Enyedi, G., "Novel Algorithms for IP Fast Reroute", Enyedi, G., "Novel Algorithms for IP Fast Reroute",
Department of Telecommunications and Media Informatics, Department of Telecommunications and Media Informatics,
Budapest University of Technology and Economics Ph.D. Budapest University of Technology and Economics Ph.D.
Thesis, February 2011, Thesis, February 2011,
<http://timon.tmit.bme.hu/theses/thesis_book.pdf>. <http://timon.tmit.bme.hu/theses/thesis_book.pdf>.
[I-D.atlas-mpls-ldp-mrt]
Atlas, A., Tiruveedhula, K., Tantsura, J., and IJ.
Wijnands, "LDP Extensions to Support Maximally Redundant
Trees", draft-atlas-mpls-ldp-mrt-00 (work in progress),
July 2013.
[I-D.atlas-ospf-mrt]
Atlas, A., Hegde, S., Chris, C., and J. Tantsura, "OSPF
Extensions to Support Maximally Redundant Trees", draft-
atlas-ospf-mrt-00 (work in progress), July 2013.
[I-D.atlas-rtgwg-mrt-mc-arch] [I-D.atlas-rtgwg-mrt-mc-arch]
Atlas, A., Kebler, R., Wijnands, I., Csaszar, A., and G. Atlas, A., Kebler, R., Wijnands, I., Csaszar, A., and G.
Envedi, "An Architecture for Multicast Protection Using Envedi, "An Architecture for Multicast Protection Using
Maximally Redundant Trees", Maximally Redundant Trees", draft-atlas-rtgwg-mrt-mc-
draft-atlas-rtgwg-mrt-mc-arch-00 (work in progress), arch-02 (work in progress), July 2013.
March 2012.
[I-D.bryant-ipfrr-tunnels]
Bryant, S., Filsfils, C., Previdi, S., and M. Shand, "IP
Fast Reroute using tunnels", draft-bryant-ipfrr-tunnels-03
(work in progress), November 2007.
[I-D.ietf-mpls-ldp-multi-topology] [I-D.ietf-mpls-ldp-multi-topology]
Zhao, Q., Fang, L., Zhou, C., Li, L., and K. Raza, "LDP Zhao, Q., Fang, L., Zhou, C., Li, L., and K. Raza, "LDP
Extensions for Multi Topology Routing", Extensions for Multi Topology Routing", draft-ietf-mpls-
draft-ietf-mpls-ldp-multi-topology-06 (work in progress), ldp-multi-topology-08 (work in progress), May 2013.
December 2012.
[I-D.ietf-rtgwg-ipfrr-notvia-addresses] [I-D.ietf-rtgwg-ipfrr-notvia-addresses]
Bryant, S., Previdi, S., and M. Shand, "A Framework for IP Bryant, S., Previdi, S., and M. Shand, "A Framework for IP
and MPLS Fast Reroute Using Not-via Addresses", and MPLS Fast Reroute Using Not-via Addresses", draft-
draft-ietf-rtgwg-ipfrr-notvia-addresses-10 (work in ietf-rtgwg-ipfrr-notvia-addresses-11 (work in progress),
progress), December 2012. May 2013.
[I-D.ietf-rtgwg-lfa-applicability]
Filsfils, C. and P. Francois, "LFA applicability in SP
networks", draft-ietf-rtgwg-lfa-applicability-06 (work in
progress), January 2012.
[I-D.ietf-rtgwg-ordered-fib] [I-D.ietf-rtgwg-ordered-fib]
Shand, M., Bryant, S., Previdi, S., Filsfils, C., Shand, M., Bryant, S., Previdi, S., Filsfils, C.,
Francois, P., and O. Bonaventure, "Framework for Loop-free Francois, P., and O. Bonaventure, "Framework for Loop-free
convergence using oFIB", draft-ietf-rtgwg-ordered-fib-09 convergence using oFIB", draft-ietf-rtgwg-ordered-fib-12
(work in progress), January 2013. (work in progress), May 2013.
[I-D.ietf-rtgwg-remote-lfa] [I-D.ietf-rtgwg-remote-lfa]
Bryant, S., Filsfils, C., Previdi, S., Shand, M., and S. Bryant, S., Filsfils, C., Previdi, S., Shand, M., and S.
Ning, "Remote LFA FRR", draft-ietf-rtgwg-remote-lfa-01 Ning, "Remote LFA FRR", draft-ietf-rtgwg-remote-lfa-02
(work in progress), December 2012. (work in progress), May 2013.
[I-D.litkowski-rtgwg-node-protect-remote-lfa]
Litkowski, S., "Node protecting remote LFA", draft-
litkowski-rtgwg-node-protect-remote-lfa-00 (work in
progress), April 2013.
[LFARevisited] [LFARevisited]
Retvari, G., Tapolcai, J., Enyedi, G., and A. Csaszar, "IP Retvari, G., Tapolcai, J., Enyedi, G., and A. Csaszar, "IP
Fast ReRoute: Loop Free Alternates Revisited", Proceedings Fast ReRoute: Loop Free Alternates Revisited", Proceedings
of IEEE INFOCOM , 2011, <http://opti.tmit.bme.hu/ of IEEE INFOCOM , 2011, <http://opti.tmit.bme.hu/~tapolcai
~tapolcai/papers/retvari2011lfa_infocom.pdf>. /papers/retvari2011lfa_infocom.pdf>.
[LightweightNotVia] [LightweightNotVia]
Enyedi, G., Retvari, G., Szilagyi, P., and A. Csaszar, "IP Enyedi, G., Retvari, G., Szilagyi, P., and A. Csaszar, "IP
Fast ReRoute: Lightweight Not-Via without Additional Fast ReRoute: Lightweight Not-Via without Additional
Addresses", Proceedings of IEEE INFOCOM , 2009, Addresses", Proceedings of IEEE INFOCOM , 2009,
<http://mycite.omikk.bme.hu/doc/71691.pdf>. <http://mycite.omikk.bme.hu/doc/71691.pdf>.
[RFC2119] Bradner, S., "Key words for use in RFCs to Indicate
Requirement Levels", BCP 14, RFC 2119, March 1997.
[RFC2328] Moy, J., "OSPF Version 2", STD 54, RFC 2328, April 1998.
[RFC3137] Retana, A., Nguyen, L., White, R., Zinin, A., and D.
McPherson, "OSPF Stub Router Advertisement", RFC 3137,
June 2001.
[RFC5443] Jork, M., Atlas, A., and L. Fang, "LDP IGP
Synchronization", RFC 5443, March 2009.
[RFC5715] Shand, M. and S. Bryant, "A Framework for Loop-Free [RFC5715] Shand, M. and S. Bryant, "A Framework for Loop-Free
Convergence", RFC 5715, January 2010. Convergence", RFC 5715, January 2010.
[RFC6571] Filsfils, C., Francois, P., Shand, M., Decraene, B.,
Uttaro, J., Leymann, N., and M. Horneffer, "Loop-Free
Alternate (LFA) Applicability in Service Provider (SP)
Networks", RFC 6571, June 2012.
Appendix A. General Issues with Area Abstraction
When a multi-homed prefix is connected in two different areas, it may
be impractical to protect them without adding the complexity of
explicit tunneling. This is also a problem for LFA and Remote-LFA.
50
|----[ASBR Y]---[B]---[ABR 2]---[C] Backbone Area 0:
| | ABR 1, ABR 2, C, D
| |
| | Area 20: A, ASBR X
| |
p ---[ASBR X]---[A]---[ABR 1]---[D] Area 10: B, ASBR Y
5 p is a Type 1 AS-external
Figure 5: AS external prefixes in different areas
Consider the network in Figure 5 and assume there is a richer
connective topology that isn't shown, where the same prefix is
announced by ASBR X and ASBR Y which are in different non-backbone
areas. If the link from A to ASBR X fails, then an MRT alternate
could forward the packet to ABR 1 and ABR 1 could forward it to D,
but then D would find the shortest route is back via ABR 1 to Area
20. This problem occurs because the routers, including the ABR, in
one area are not yet aware of the failure in a different area.
The only way to get it from A to ASBR Y is to explicitly tunnel it to
ASBR Y. If the traffic is unlabeled or the appropriate MPLS labels
are known, then explicit tunneling MAY be used as long as the
shortest-path of the tunnel avoids the failure point. In that case,
A must determine that it should use an explicit tunnel instead of an
MRT alternate.
Authors' Addresses Authors' Addresses
Alia Atlas (editor) Alia Atlas (editor)
Juniper Networks Juniper Networks
10 Technology Park Drive 10 Technology Park Drive
Westford, MA 01886 Westford, MA 01886
USA USA
Email: akatlas@juniper.net Email: akatlas@juniper.net
Robert Kebler Robert Kebler
Juniper Networks Juniper Networks
10 Technology Park Drive 10 Technology Park Drive
Westford, MA 01886 Westford, MA 01886
USA USA
Email: rkebler@juniper.net Email: rkebler@juniper.net
Gabor Sandor Enyedi Gabor Sandor Enyedi
Ericsson Ericsson
Konyves Kalman krt 11. Konyves Kalman krt 11.
Budapest 1097 Budapest 1097
Hungary Hungary
Email: Gabor.Sandor.Enyedi@ericsson.com Email: Gabor.Sandor.Enyedi@ericsson.com
Andras Csaszar Andras Csaszar
Ericsson Ericsson
skipping to change at page 27, line 4 skipping to change at page 29, line 32
300 Holger Way 300 Holger Way
San Jose, CA 95134 San Jose, CA 95134
USA USA
Email: jeff.tantsura@ericsson.com Email: jeff.tantsura@ericsson.com
Maciek Konstantynowicz Maciek Konstantynowicz
Cisco Systems Cisco Systems
Email: maciek@bgp.nu Email: maciek@bgp.nu
Russ White
Verisign
12061 Bluemont Way
Reston, VA 20190
USA
Email: riwhite@verisign.com
Mike Shand Russ White
VCE
Email: mike@mshand.org.uk Email: russw@riw.us
 End of changes. 147 change blocks. 
634 lines changed or deleted 845 lines changed or added

This html diff was produced by rfcdiff 1.41. The latest version is available from http://tools.ietf.org/tools/rfcdiff/