draft-ietf-tsvwg-sctp-failover-08.txt   draft-ietf-tsvwg-sctp-failover-09.txt 
Network Working Group Y. Nishida Network Working Group Y. Nishida
Internet-Draft GE Global Research Internet-Draft GE Global Research
Intended status: Standards Track P. Natarajan Intended status: Standards Track P. Natarajan
Expires: April 27, 2015 Cisco Systems Expires: June 27, 2015 Cisco Systems
A. Caro A. Caro
BBN Technologies BBN Technologies
P. Amer P. Amer
University of Delaware University of Delaware
K. Nielsen K. Nielsen
Ericsson Ericsson
October 24, 2014 December 24, 2014
Quick Failover Algorithm in SCTP SCTP-PF: Quick Failover Algorithm in SCTP
draft-ietf-tsvwg-sctp-failover-08.txt draft-ietf-tsvwg-sctp-failover-09.txt
Abstract Abstract
One of the major advantages of SCTP is that it supports multi-homed One of the major advantages of SCTP is the support of multi-homed
communication. A multi-homed SCTP end-point has the ability to communication. A multi-homed SCTP end-point has the ability to
withstand network failures by migrating the traffic from an inactive withstand network failures by migrating the traffic from an inactive
network to an active one. However, if the [RFC4960] specified network to an active one. However, if the failover operation as
failover operation is followed there can be a significant delay in specified in [RFC4960] is followed, there can be a significant delay
the migration to the active destination addresses, thus severely in the migration to the active destination addresses, thus severely
reducing the effectiveness of SCTP multi-homed operation. reducing the effectiveness of the SCTP failover operation.
The memo complements RFC4960 by the introduction of the Potentially This memo complements [RFC4960] by the introduction of the
Failed state and associated new Quick Failover operation to apply Potentially Failed path state and the associated new failover
during network failure and specifies for SCTP senders to support this operation called SCTP-PF to apply during a network failure. In
more performance optimal failover procedure as an add-on to the addition, the memo complements [RFC4960] by introducing of
[RFC4960] failover operation. The memo in addition complements alternative switchover operation modes for the data transfer path
[RFC4960] by introduction of alternative switchover operation modes management after the recovery of a failed primary path. These modes
for the data transfer path management after a failover. These offers for more performance optimal operation in some network
operation modes offer for more performance optimal operation in some environments. The implementation of the additional switchover
network environments. From the perspective of this memo the operation modes is optional.
implementation of the additional switchover operation modes is
considered optional.
The procedures defined require only minimal modifications to the The procedures defined in the document require only minimal
current specification. The procedures are sender-side only and do modifications to the current specification. The procedures are
not impact the SCTP receiver. sender-side only and do not impact the SCTP receiver.
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 April 27, 2015. This Internet-Draft will expire on June 27, 2015.
Copyright Notice Copyright Notice
Copyright (c) 2014 IETF Trust and the persons identified as the Copyright (c) 2014 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 . . . . . . . . . . . . . . . . . . . . . . . . 3 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3
2. Conventions and Terminology . . . . . . . . . . . . . . . . . 4 2. Conventions and Terminology . . . . . . . . . . . . . . . . . 3
3. Issues with the SCTP Path Management . . . . . . . . . . . . 4 3. Issues with the SCTP Path Management . . . . . . . . . . . . 4
4. SCTP with Potentially-Failed Destination State (SCTP-PF) . . 5 4. SCTP with Potentially-Failed Destination State (SCTP-PF) . . 5
4.1. SCTP-PF Concept . . . . . . . . . . . . . . . . . . . . . 5 4.1. SCTP-PF Concept . . . . . . . . . . . . . . . . . . . . . 5
4.2. SCTP-PF Algorithm Detail . . . . . . . . . . . . . . . . 6 4.2. SCTP-PF Algorithm in Detail . . . . . . . . . . . . . . . 6
4.3. Optional Feature: Permanent Failover . . . . . . . . . . 9 4.3. Optional Feature: Permanent Failover . . . . . . . . . . 9
5. Socket API Considerations . . . . . . . . . . . . . . . . . . 10 5. Socket API Considerations . . . . . . . . . . . . . . . . . . 11
5.1. Support for the Potentially Failed Path State . . . . . . 11 5.1. Support for the Potentially Failed Path State . . . . . . 11
5.2. Peer Address Thresholds (SCTP_PEER_ADDR_THLDS) Socket 5.2. Peer Address Thresholds (SCTP_PEER_ADDR_THLDS) Socket
Option . . . . . . . . . . . . . . . . . . . . . . . . . 12 Option . . . . . . . . . . . . . . . . . . . . . . . . . 12
5.3. Exposing the Potentially Failed Path State 5.3. Exposing the Potentially Failed Path State
(SCTP_EXPOSE_POTENTIALLY_FAILED_STATE) Socket Option . . 13 (SCTP_EXPOSE_POTENTIALLY_FAILED_STATE) Socket Option . . 13
6. Security Considerations . . . . . . . . . . . . . . . . . . . 13 6. Security Considerations . . . . . . . . . . . . . . . . . . . 14
7. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 14 7. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 14
8. Proposed Change of Status (to be Deleted before Publication) 14 8. Proposed Change of Status (to be Deleted before Publication) 14
9. References . . . . . . . . . . . . . . . . . . . . . . . . . 14 9. References . . . . . . . . . . . . . . . . . . . . . . . . . 14
9.1. Normative References . . . . . . . . . . . . . . . . . . 14 9.1. Normative References . . . . . . . . . . . . . . . . . . 14
9.2. Informative References . . . . . . . . . . . . . . . . . 14 9.2. Informative References . . . . . . . . . . . . . . . . . 15
Appendix A. Discussions of Alternative Approaches . . . . . . . 16
Appendix A. Discussions of Alternative Approaches . . . . . . . 15 A.1. Reduce Path.Max.Retrans (PMR) . . . . . . . . . . . . . . 16
A.1. Reduce Path.Max.Retrans (PMR) . . . . . . . . . . . . . . 15
A.2. Adjust RTO related parameters . . . . . . . . . . . . . . 16 A.2. Adjust RTO related parameters . . . . . . . . . . . . . . 16
Appendix B. Discussions for Path Bouncing Effect . . . . . . . . 17 Appendix B. Discussions for Path Bouncing Effect . . . . . . . . 17
Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 17 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 17
1. Introduction 1. Introduction
The Stream Control Transmission Protocol (SCTP) as specified in The Stream Control Transmission Protocol (SCTP) as specified in
[RFC4960] supports multihoming at the transport layer -- an SCTP [RFC4960] supports multihoming at the transport layer -- an SCTP
association can bind to multiple IP addresses at each endpoint. endpoint can bind to multiple IP addresses. SCTP's multihoming
SCTP's multihoming features include failure detection and failover features include failure detection and failover procedures to provide
procedures to provide network interface redundancy and improved end- network interface redundancy and improved end-to-end fault tolerance.
to-end fault tolerance.
In SCTP's current failure detection procedure, the sender must In SCTP's current failure detection procedure, the sender must
experience Path.Max.Retrans (PMR) number of consecutive failed experience Path.Max.Retrans (PMR) number of consecutive failed timer-
retransmissions on a destination before detecting a path failure. based retransmissions on a destination address before detecting a
The sender fails over to an alternate active destination only after path failure. The sender fails over to an alternate active
failure detection. Until detecting the failover, the sender destination address only after failure detection. Until detecting
continues to transmit data on the failed path, which degrades the the failover, the sender continues to transmit data on the failed
SCTP performance. Concurrent Multipath Transfer (CMT) [IYENGAR06] is path, which degrades the SCTP performance. Concurrent Multipath
an extension to SCTP and allows the sender to transmit data on Transfer (CMT) [IYENGAR06] is an extension to SCTP that allows the
multiple paths simultaneously. Research [NATARAJAN09] shows that the sender to transmit data on multiple paths simultaneously. Research
current failure detection procedure worsens CMT performance during [NATARAJAN09] shows that the current failure detection procedure
failover and can be significantly improved by employing a better worsens CMT performance during failover and can be significantly
failover algorithm. improved by employing a better failover algorithm.
This document specifies an alternative failure detection procedure This document specifies an alternative failure detection procedure
for SCTP that improves the SCTP performance during a failover. for SCTP that improves the SCTP performance during a failover.
Also the operation after a failover impacts the performance of the Also the operation after the recovery of a failed path impacts the
protocol. With [RFC4960] procedures, SCTP will, after a failover performance of the protocol. With procedures specified in [RFC4960],
from the primary path, switch back to use the primary path for data SCTP will, after a failover from the primary path, switch back to the
transfer as soon as this path becomes available. From a performance primary path for data transfer as soon as this path becomes available
perspective, as confirmed in research [CARO02], such a switchback of again. From a performance perspective, as confirmed in research
the data transmission path is not optimal in general. As an optional [CARO02], such a switchback of the data transmission path is not
alternative to the switchback operation of [RFC4960], this document optimal in general. As an optional alternative to the switchback
specifies for SCTP to support the Permanent Failover switchover operation of [RFC4960], this document specifies the Permanent
procedures proposed by [CARO02]. Additional discussions for Failover procedures proposed by [CARO02].
alternative approach that does not require modifications to [RFC4960]
and path bouncing effects that might be caused by frequent switchover Additional discussions for alternative approaches that do not require
are provided in Appendix. modifications to [RFC4960] and path bouncing effects that might be
caused by frequent switchover are provided in the Appendices.
2. Conventions and Terminology 2. Conventions and Terminology
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in [RFC2119]. document are to be interpreted as described in [RFC2119].
3. Issues with the SCTP Path Management 3. Issues with the SCTP Path Management
This section describes issues in the current SCTP to be fixed by the This section describes issues in the SCTP as specified in [RFC4960]
approach described in this document. to be fixed by the approach described in this document.
SCTP can utilize multiple IP addresses for a single SCTP association. An SCTP endpoint can support multiple IP addresses. Each SCTP
Each SCTP endpoint exchanges the list of its usable addresses during endpoint exchanges the list of its usable addresses during the
initial negotiation with its peer. Then the endpoints select one initial negotiation with its peer. Then the endpoints select one
address from the peer's list and define this as the primary address from the peer's list and use this as the primary destination
destination. During normal transmission, SCTP sends all user data to address. During normal transmission, an SCTP endpoint sends all user
the primary destination. Also, it sends heartbeat packets to all data to the primary destination address. Also, it sends packets
idle destinations at a certain interval to check the reachability of containing a HEARTBEAT chunk to all idle destination addresses at a
the path. Idle destinations normally include all non-primary certain interval to check the reachability of these destination
destinations. addresses. Idle destination addresses normally include all non-
primary destination addresses.
If a sender has multiple active destination addresses, it can If a sender has multiple active destination addresses, it can
retransmit data to secondary destination address, when the retransmit data to an non-primary destination address, if the
transmission to the primary times out. transmission to the primary times out.
When a sender receives an acknowledgment for DATA or HEARTBEAT chunks When a sender receives an acknowledgment for DATA or HEARTBEAT chunks
sent to one of the destination addresses, it considers that sent to one of the destination addresses, it considers that
destination to be active. If it fails to receive acknowledgments, destination address to be active and clears the error counter for the
the error count for the address is increased. If the error counter destination address. If it fails to receive acknowledgments, the
exceeds the protocol parameter 'Path.Max.Retrans', SCTP endpoint error count for the destination address is increased. If the error
considers the address to be inactive. counter exceeds the tunable protocol parameter Path.Max.Retrans
(PMR), the SCTP endpoint considers the destination address to be
inactive.
The failover process of SCTP is initiated when the primary path The failover process of SCTP is initiated when the primary path
becomes inactive (error counter for the primary path exceeds becomes inactive (the error counter for the primary path exceeds
Path.Max.Retrans). If the primary path is marked inactive, SCTP Path.Max.Retrans). If the primary path is marked inactive, SCTP
chooses a new destination address from one of the active destinations chooses a new destination address from one of the active destinations
and start using this address to send data to. If the primary path and start using this address to send data to. If the primary path
becomes active again, SCTP uses the primary destination for becomes active again, SCTP uses the primary destination address for
subsequent data transmissions and stop using non-primary one. subsequent data transmissions and stop using the non-primary one.
One issue with this failover process is that it usually takes One issue with this failover process is that it usually takes a
significant amount of time before SCTP switches to the new significant amount of time before SCTP switches to the new
destination. Let's say the primary path on a multi-homed host destination address. Let's say the primary path on a multi-homed
becomes unavailable and the RTO value for the primary path at that host becomes unavailable and the RTO value for the primary path at
time is around 1 second, it usually takes over 60 seconds before SCTP that time is around 1 second, it usually takes over 60 seconds before
starts to use the secondary path. This is because the recommended SCTP starts to use the non-primary path for initial data
value for Path.Max.Retrans in the standard is 5, which requires 6 transmission. This is because the recommended value for
consecutive timeouts before failover takes place. Before SCTP Path.Max.Retrans in the [RFC4960] is 5, which requires 6 consecutive
switches to the secondary address, SCTP keeps trying to send packets timeouts before the failover takes place. Before SCTP switches to
to the primary and only retransmitted packets are sent to the the non-primary address, SCTP keeps trying to send packets to the
secondary and can thus be reached at the receiver. This slow primary address and only retransmitted packets are sent to the non-
failover process can cause significant performance degradation and primary address and thus can be received by the receiver. This slow
will not be acceptable in some situations. failover process can cause significant performance degradation and is
not acceptable in some situations.
Another issue is that once the primary path is active again, the Another issue is that once the primary path becomes active again, the
traffic is switched back. This is not optimal in some situations. traffic is switched back. This is not optimal in some situations.
This is further discussed in Section 4.3. This is further discussed in Section 4.3.
4. SCTP with Potentially-Failed Destination State (SCTP-PF) 4. SCTP with Potentially-Failed Destination State (SCTP-PF)
To address the issues described in Section 3, this section updates To address the issues described in Section 3, this section extends
SCTP path management scheme with the Potentially Failed state and SCTP path management scheme by adding the Potentially Failed state
associated Quick Failover operation. We use the term SCTP-PF to and the associated failover operation. We use the term SCTP-PF to
denote the resulting SCTP path management operation. denote the resulting SCTP path management operation.
4.1. SCTP-PF Concept 4.1. SCTP-PF Concept
SCTP-PF as defined stems from the following two observations about SCTP-PF as defined stems from the following two observations about
SCTP's failure detection procedure: SCTP's failure detection procedure:
o To minimize performance impact during failover, the sender should o To minimize the performance impact during failover, the sender
avoid transmitting data to the failed destination as early as should avoid transmitting data to the failed destination address
possible. In the current SCTP path management scheme, the sender as early as possible. In the current SCTP path management scheme,
stops transmitting data to a destination only after the the sender stops transmitting data to a destination destination
destination is marked Failed (inactive). Thus, a smaller PMR only after the destination is marked Failed (inactive). Thus, a
value is ideal so that the sender transitions a destination to the smaller PMR value is better because the sender can transition a
Failed (inactive) state quicker. destination address to the Failed (inactive) state quicker.
o Smaller PMR values increase the chances of spurious failure o Smaller PMR values increase the chances of spurious failure
detection where the sender incorrectly marks a destination as detection where the sender incorrectly marks a destination address
Failed (inactive) during periods of temporary congestion. As as Failed (inactive) during periods of temporary congestion. As
[RFC4960] recommends for a coupling of the PMR value and the AMR [RFC4960] recommends for a coupling of the PMR value and the
value such spurious failure detection risks to carry over to protocol parameter Association.Max.Retrans (AMR) value such
spurious association failure detection and closure. Larger PMR spurious failure detection risks to carry over to spurious
values are preferable to avoid spurious failure detection. association failure detection and closure. Larger PMR values are
preferable to avoid spurious failure detection.
From the above observations it is clear that tuning the PMR value From the above observations it is clear that tuning the PMR value
involves the following tradeoff -- a lower value improves performance involves the following tradeoff -- a lower value improves performance
but increases the chances of spurious failure detection, whereas a but increases the chances of spurious failure detection, whereas a
higher value degrades performance and reduces spurious failure higher value degrades performance and reduces spurious failure
detection in a wide range of path conditions. Thus, tuning the detection in a wide range of path conditions. Thus, tuning the
association's PMR value is an incomplete solution to address association's PMR value is an incomplete solution to address the
performance impact during failure. performance impact during failure.
This new method introduces a new "Potentially-Failed" (PF) SCTP-PF defined in this document introduces a new "Potentially-
destination state in SCTP's path management procedure. The PF state Failed" (PF) destination state in SCTP's path management procedure.
was originally proposed to improve CMT performance [NATARAJAN09]. The PF state was originally proposed to improve CMT performance
The PF state is an intermediate state between Active and Failed
states. SCTP's failure detection procedure is modified to include
the PF state. The new failure detection algorithm assumes that loss
detected by a timeout implies either severe congestion or failure en-
route. After a number of consecutive timeouts on a path, the sender
is unsure, and marks the corresponding destination as PF. A PF
destination is not used for data transmission except in special cases
(discussed below). The new failure detection algorithm requires only
sender-side changes.
4.2. SCTP-PF Algorithm Detail [NATARAJAN09]. The PF state is an intermediate state between the
Active and Failed states. SCTP's failure detection procedure is
modified to include the PF state. The new failure detection
algorithm assumes that loss detected by a timeout implies either
severe congestion or failure en-route. After a number of consecutive
timeouts on a path, the sender is unsure, and marks the corresponding
destination address as PF. A PF destination address is not used for
data transmission except in special cases (discussed below). The new
failure detection algorithm requires only sender-side changes.
SCTP PF operation is specified as follows: 4.2. SCTP-PF Algorithm in Detail
The SCTP-PF operation is specified as follows:
1. The sender maintains a new tunable parameter called Potentially- 1. The sender maintains a new tunable parameter called Potentially-
Failed.Max.Retrans (PFMR). The RECOMMENDED value of PFMR = 0 Failed.Max.Retrans (PFMR). The RECOMMENDED value of PFMR = 0
when Quick Failover is used. When PFMR is larger or equal to when SCTP-PF is used. When PFMR is larger or equal to PMR,
PMR, Quick Failover is turned off. SCTP-PF is turned off.
2. The error counter of an active destination address is 2. The error counter of an active destination address is
incremented as specified in [RFC4960]. This means that the incremented as specified in [RFC4960]. This means that the
error counter of the destination address will be incremented error counter of the destination address will be incremented
each time the T3-rtx timer expires, or at times where a each time the T3-rtx timer expires, or at times where a
HEARTBEAT sent to an idle, active address is not acknowledged HEARTBEAT sent to an idle, active address is not acknowledged
within an RTO. When the value in the destination address error within an RTO. When the value in the destination address error
counter exceeds PFMR, the endpoint MUST mark the destination counter exceeds PFMR, the endpoint MUST mark the destination
transport address as PF. transport address as PF.
3. The sender SHOULD avoid data transmission to PF destinations. 3. The sender SHOULD avoid data transmission to PF destination
When the destinations are all in PF state or some in PF state addresses. When the destination addresses are all in PF state
and some in inactive state, the sender MUST choose one or some in PF state and some in inactive state, the sender MUST
destination in PF state and transmit data to this destination. choose one destination address in PF state and transmit data to
The sender SHOULD choose the destination in PF state with the this destination. The sender SHOULD choose the destination
lowest error count (fewest consecutive timeouts) for data address in PF state with the lowest error count (fewest
transmission and transmit data to this destination. When there consecutive timeouts) for data transmission and transmit data to
are multiple PF destinations with same error count, the sender this destination. When there are multiple PF destinations with
SHOULD let the choice among the multiple PF destination with same error count, the sender SHOULD let the choice among the
equal error count be based on the [RFC4960], section 6.4.1, multiple PF destination address with equal error count be based
principles of choosing most divergent source-destination pairs on the [RFC4960], section 6.4.1, principles of choosing most
when executing (potentially consecutive) retransmission. This divergent source-destination pairs when executing (potentially
means that the sender SHOULD attempt to pick the most divergent consecutive) retransmission. This means that the sender SHOULD
source - destination pair from the last source - destination attempt to pick the most divergent source - destination pair
pair on which data were transmitted or retransmitted. Rules for from the last source - destination pair on which data were
picking the most divergent source-destination pair are an transmitted or retransmitted. Rules for picking the most
implementation decision and are not specified within this divergent source-destination pair are an implementation decision
document. A sender may choose to deploy other strategies than and are not specified within this document. A sender may choose
the above when choosing among multiple PF destinations with to deploy other strategies than the above when choosing among
equal error count. In all cases the sender MUST NOT change the multiple PF destinations with equal error count. In all cases,
state of chosen destination and it MUST NOT clear the the sender MUST NOT change the state of chosen destination
destination's error counter as a result of choosing the address and it MUST NOT clear the destination's error counter as
destination for data transmission. a result of choosing the destination address for data
transmission.
4. Heartbeats SHOULD be sent to PF destination(s) once per RTO. 4. HEARTBEAT chunks SHOULD be sent to PF destination(s) once per
This means the sender MUST ignore HB.interval for PF RTO, which requires to ignore HB.interval for PF destinations.
destinations. If an heartbeat is unanswered, the sender SHOULD If a HEARTBEAT chunk is not acknowledged, the sender SHOULD
increment the error counter and exponentially back off the RTO increment the error counter and exponentially back off the RTO
value. If error counter is less than PMR, the sender SHOULD value. If error counter is less than PMR, the sender SHOULD
transmit another heartbeat immediately after T3-timer transmit another packet containing HEARTBEAT chunk immediately
expiration. When data is transmitted to a PF destination, the after T3-timer expiration. When data is transmitted to a PF
transmission of heartbeats may be omitted as SACK or T3-rtx destination, the transmission of HEARTBEAT chunk MAY be omitted
timer expiration can provide equivalent information. It is as receipt of SACK chunks or a T3-rtx timer expiration can
RECOMMENDED that heartbeats be send to PF destinations provide equivalent information. It is RECOMMENDED that
regardless of whether the Path Heartbeat function (Section 8.3 HEARTBEAT chunks are send to PF destinations regardless of
of [RFC4960]) is enabled for the destination address or not. whether the Path Heartbeat function (Section 8.3 of [RFC4960])
is enabled for the destination address or not.
5. When the sender receives an heartbeat ACK from a PF destination, 5. When the sender receives a HEARTBEAT ACK from a PF destination,
the sender MUST clear the destination's error counter and the sender MUST clear the destination's error counter and
transition the PF destination back to Active state. When the transition the PF destination address back to Active state.
sender resumes data transmission on the destination it MUST do When the sender resumes data transmission on the destination
this following the prescriptions of Section 7.2 of [RFC4960]. address, it MUST do this following the prescriptions of
Section 7.2 of [RFC4960].
6. Additional (PMR - PFMR) consecutive timeouts on a PF destination 6. Additional (PMR - PFMR) consecutive timeouts on a PF destination
confirm the path failure, upon which the destination transitions address confirm the path failure, upon which the destination
to the Inactive state. As described in [RFC4960], the sender address transitions to the Inactive state. As described in
(i) SHOULD notify ULP about this state transition, and (ii) [RFC4960], the sender (i) SHOULD notify ULP about this state
transmit heartbeats to the Inactive destination at a lower transition, and (ii) transmit HEARTBEAT chunks to the Inactive
frequency as described in Section 8.3 of [RFC4960] (when this destination address at a lower frequency as described in
function is enabled for the destination address). Section 8.3 of [RFC4960] (when this function is enabled for the
destination address).
7. When all destinations are in inactive state (association dormant 7. When all destinations are in inactive state (association dormant
state) the sender MUST also choose one destination to transmit state) the sender MUST also choose one destination address to
data to. The sender SHOULD choose the destination in inactive transmit data to. The sender SHOULD choose the destination
state with the lowest error count (fewest consecutive timeouts) address in inactive state with the lowest error count (fewest
for data transmission and transmit data to this destination. consecutive timeouts) for data transmission and transmit data to
When there are multiple destinations with same error count in this destination. When there are multiple destination addresses
inactive state, the sender SHOULD attempt to pick the most with same error count in inactive state, the sender SHOULD
divergent source - destination pair from the last source - attempt to pick the most divergent source - destination pair
destination pair on which data were transmitted or retransmitted from the last source - destination pair on which data were
following [RFC4960]. Rules for picking the most divergent transmitted or retransmitted following [RFC4960]. Rules for
source-destination pair are an implementation decision and are picking the most divergent source-destination pair are an
not specified within this document. Therefore, a sender SHOULD implementation decision and are not specified within this
allow for incrementing the destination error counters up to some document. Therefore, a sender SHOULD allow for incrementing the
reasonable limit larger than PMR+1, thus changing the destination error counters up to some reasonable limit larger
prescriptions of [RFC4960], section 8.3, in this respect. The than PMR+1, thus changing the prescriptions of [RFC4960],
exact limit to apply is not specified in this document but it is section 8.3, in this respect. The exact limit to apply is not
considered reasonable to require for such to be an order of specified in this document but it is considered reasonable to
magnitude higher than the PMR value. A sender MAY choose to require for such to be an order of magnitude higher than the PMR
deploy other strategies than the above. For example, a sender value. A sender MAY choose to deploy other strategies than the
could choose to prioritize the last active destination during above. For example, a sender could choose to prioritize the
dormant state. The strategy to prioritize the last active last active destination address during dormant state. The
destination is optimal when some paths are permanently inactive, strategy to prioritize the last active destination address is
but suboptimal when paths' instability is transient. While the optimal when some paths are permanently inactive, but suboptimal
increment of the error counters above PMR+1 is a prerequisite when paths' instability is transient. While the increment of
for the error counter values to serve to guide the path the error counters above PMR+1 is a prerequisite for the error
selection in dormant state, then it is noted that by virtue of counter values to serve to guide the path selection in dormant
the introduction of the Potentially Failed state, one may deploy state, then it is noted that by virtue of the introduction of
higher values of PMR without compromising the efficiency of the the Potentially Failed state, one may deploy higher values of
failover operation, and thus making the increase of path error PMR without compromising the efficiency of the failover
counters above PMR+1 less critical as the dormant state will be operation, and thus making the increase of path error counters
less likely to happen. The downside of increasing the PMR value above PMR+1 less critical as the dormant state will be less
likely to happen. The downside of increasing the PMR value
relative to the AMR value, however, is that the per destination relative to the AMR value, however, is that the per destination
address failure detection and notification of such to ULP address failure detection and notification of such to ULP
thereby is weakened. In all cases the sender MUST NOT change thereby is weakened. In all cases the sender MUST NOT change
the state of the chosen destination and it MUST NOT clear the the state of the chosen destination address and it MUST NOT
destination's error counter as a result of choosing the clear the destination's error counter as a result of choosing
destination for data transmission. the destination address for data transmission.
8. ACKs for chunks that have been transmitted to multiple 8. Acknowledgments for chunks that have been transmitted to
destinations (i.e., a chunk which has been retransmitted to a multiple destinations (i.e., a chunk which has been
different destination than the destination to which the chunk retransmitted to a different destination address than the
was first transmitted) SHOULD NOT clear the error count of an destination address to which the chunk was first transmitted)
inactive destination and SHOULD NOT transition a PF destination SHOULD NOT clear the error count of an inactive destination
back to Active state, since a sender cannot disambiguate whether address and SHOULD NOT transition a PF destination address back
the ACK was for the original transmission or the to Active state, since a sender cannot disambiguate whether the
retransmission(s). The same ambiguity concerns the related ACK was for the original transmission or the retransmission(s).
congestion window growth. The bytes of a newly acknowledged The same ambiguity concerns the related congestion window
chunk which has been transmitted to multiple destinations SHOULD growth. The bytes of a newly acknowledged chunk which has been
be considered for contribution to the congestion window growth transmitted to multiple destination addresses SHOULD be
towards the destination where the chunk was last sent. The considered for contribution to the congestion window growth
contribution of the acked bytes to the window growth is subject towards the destination address where the chunk was last sent.
to the prescriptions described in Section 7.2 of [RFC4960] is The contribution of the ACKed bytes to the window growth is
fulfilled. A SCTP sender MAY apply a different approach for subject to the prescriptions described in Section 7.2 of
both the error count handling and the congestion control growth [RFC4960] is fulfilled. A SCTP sender MAY apply a different
handling based on unequivocally information on which destination approach for both the error count handling and the congestion
(including multiple destinations) the chunk reached. This control growth handling based on unequivocally information on
document makes no reference to what such unequivocally which destination (including multiple destination addresses) the
information could consist of, neither how such unequivocally chunk reached. This document makes no reference to what such
information could be obtained. The implementation of such an unequivocally information could consist of, neither how such
alternative approach is left to implementations. unequivocally information could be obtained. The implementation
of such an alternative approach is left to implementations.
9. ACKs for chunks which has been transmitted to one destination 9. Acknowledgments for chunks that has been transmitted to one
address only MUST clear the error counter of the destination destination address only MUST clear the error counter of the
address and MUST transition a PF destination back to Active destination address and MUST transition a PF destination address
state. This situation can happen when new data is sent to a back to Active state. This situation can happen when new data
destination address in PF state. It can also happen in is sent to a destination address in PF state. It can also
situations where the destination address is in PF state due to happen in situations where the destination address is in PF
the occurrence of a spurious T3-rtx timer and ACKs start to state due to the occurrence of a spurious T3-rtx timer and
arrive for data sent prior to occurrence of the spurious T3-rtx Acknowledgments start to arrive for data sent prior to
and data has not yet been retransmitted towards other occurrence of the spurious T3-rtx and data has not yet been
destinations. This document does not specify special handling retransmitted towards other destinations. This document does
for detection of or reaction to spurious T3-rtx timeouts, e.g., not specify special handling for detection of or reaction to
for special operation vis-a-vis the congestion control handling spurious T3-rtx timeouts, e.g., for special operation vis-a-vis
or data retransmission operation towards a destination address the congestion control handling or data retransmission operation
which undergoes a transition from active to PF to active state towards a destination address which undergoes a transition from
due to a spurious T3-rtx timeout. But it is noted that this is active to PF to active state due to a spurious T3-rtx timeout.
an area which would benefit from additional attention, But it is noted that this is an area which would benefit from
experimentation and specification for Single Homed SCTP as well additional attention, experimentation and specification for
as for Multi Homed SCTP protocol operation. Single Homed SCTP as well as for Multi Homed SCTP protocol
operation.
10. SCTP stack SHOULD provide the ULP with the means to expose the 10. SCTP stack SHOULD provide the ULP with the means to expose the
PF state of its destinations as well as the means to notify the PF state of its destinations as well as the means to notify the
state transitions from Active to PF, and vice-versa. When doing state transitions from Active to PF, and vice-versa. When doing
this, such SCTP stack MUST provide the ULP with the means to this, such an SCTP stack MUST provide the ULP with the means to
suppress exposure of PF state and association state transitions suppress exposure of PF state and associated state transitions
as well. as well.
4.3. Optional Feature: Permanent Failover 4.3. Optional Feature: Permanent Failover
In [RFC4960], an SCTP sender migrates the traffic back to the In [RFC4960], an SCTP sender migrates the traffic back to the
original primary destination once this destination becomes active original primary destination address once this address becomes active
again. As the CWND towards the original primary destination has to again. As the CWND towards the original primary destination address
be rebuilt once data transfer resumes, the switch back to use the has to be rebuilt once data transfer resumes, the switch back to use
original primary path is not always optimal. Indeed [CARO02] shows the original primary address is not always optimal. Indeed [CARO02]
that the switch back to the original primary may degrade SCTP shows that the switch back to the original primary may degrade SCTP
performance compared to continuing data transmission on the same performance compared to continuing data transmission on the same
path, especially, but not only, in scenarios where this path's path, especially, but not only, in scenarios where this path's
characteristics are better. In order to mitigate this performance characteristics are better. In order to mitigate this performance
degradation, Permanent Failover operation was proposed in [CARO02]. degradation, the Permanent Failover operation was proposed in
When SCTP changes the destination due to failover, Permanent Failover [CARO02]. When SCTP changes the destination address due to failover,
operation allows SCTP sender to continue data transmission on the new Permanent Failover operation allows SCTP sender to continue data
working path even if the old primary destination becomes active transmission on the new working path even when the old primary
again. This is achieved by having SCTP perform a switch over of the destination address becomes active again. This is achieved by having
primary path to the alternative working path rather than having SCTP SCTP perform a switch over of the primary path to the alternative
switch back data transfer to the (previous) primary path. working path rather than having SCTP switch back data transfer to the
(previous) primary path.
The manner of switch over operation that is most optimal in a given The manner of switch over operation that is most optimal in a given
scenario depends on the relative quality of a set primary path versus scenario depends on the relative quality of a set primary path versus
the quality of alternative paths available as well as it depends on the quality of alternative paths available as well as it depends on
the extent to which it is desired for the mode of operation to the extent to which it is desired for the mode of operation to
enforce traffic distribution over a number of network paths. I.e., enforce traffic distribution over a number of network paths. I.e.,
load distribution of traffic from multiple SCTP associations may be load distribution of traffic from multiple SCTP associations may be
sought to be enforced by distribution of the set primary paths with sought to be enforced by distribution of the set primary paths with
[RFC4960] switchback operation. However as [RFC4960] switchback [RFC4960] switchback operation. However as [RFC4960] switchback
behavior is suboptimal in certain situations, especially in scenarios behavior is suboptimal in certain situations, especially in scenarios
skipping to change at page 10, line 41 skipping to change at page 10, line 50
failed primary path is performed. An implementation of Permanent failed primary path is performed. An implementation of Permanent
Failover MUST support the setting of PSMR = PFMR. An Failover MUST support the setting of PSMR = PFMR. An
implementation of Permanent Failover MAY support setting of PSMR implementation of Permanent Failover MAY support setting of PSMR
> PFMR. > PFMR.
5. It MUST be possible to disable the Permanent Failover and obtain 5. It MUST be possible to disable the Permanent Failover and obtain
the standard switchback operation of [RFC4960]. the standard switchback operation of [RFC4960].
This specifications RECOMMENDS a default configuration that uses This specifications RECOMMENDS a default configuration that uses
standard RFC4960 switchback, i.e., switch back to the old primary standard RFC4960 switchback, i.e., switch back to the old primary
destination once the destination becomes active again. However, to destination once the destination address becomes active again.
support optimal operation in a wider range of network scenarios, an However, to support optimal operation in a wider range of network
implementation MAY implement Permanent Failover operation as detailed scenarios, an implementation MAY implement Permanent Failover
above and MAY enable it based on network configurations or users' operation as detailed above and MAY enable it based on network
requests. configurations or users' requests.
5. Socket API Considerations 5. Socket API Considerations
This section describes how the socket API defined in [RFC6458] is This section describes how the socket API defined in [RFC6458] is
extended to provide a way for the application to control and observe extended to provide a way for the application to control and observe
the quick failover behavior. the SCTP-PF behavior.
Please note that this section is informational only. Please note that this section is informational only.
A socket API implementation based on [RFC6458] is, by means of the A socket API implementation based on [RFC6458] is, by means of the
existing SCTP_PEER_ADDR_CHANGE event, extended to provide the event existing SCTP_PEER_ADDR_CHANGE event, extended to provide the event
notification when a peer address enters or leaves the potentially notification when a peer address enters or leaves the potentially
failed state as well as the socket API implementation is extended to failed state as well as the socket API implementation is extended to
expose the potentially failed state of a peer address in the existing expose the potentially failed state of a peer address in the existing
SCTP_GET_PEER_ADDR_INFO structure. SCTP_GET_PEER_ADDR_INFO structure.
skipping to change at page 12, line 23 skipping to change at page 12, line 34
}; };
[RFC6458] defines the constants SCTP_UNCONFIRMED, SCTP_ACTIVE, and [RFC6458] defines the constants SCTP_UNCONFIRMED, SCTP_ACTIVE, and
SCTP_INACTIVE to be provided in the spinfo_state field. This SCTP_INACTIVE to be provided in the spinfo_state field. This
document defines in addition to that the new constant document defines in addition to that the new constant
SCTP_POTENTIALLY_FAILED, which is reported if the peer address is SCTP_POTENTIALLY_FAILED, which is reported if the peer address is
potentially failed. potentially failed.
5.2. Peer Address Thresholds (SCTP_PEER_ADDR_THLDS) Socket Option 5.2. Peer Address Thresholds (SCTP_PEER_ADDR_THLDS) Socket Option
Applications can control the quick failover behavior by getting or Applications can control the SCTP-PF behavior by getting or setting
setting the number of consecutive timeouts before a peer address is the number of consecutive timeouts before a peer address is
considered potentially failed or unreachable and before the primary considered potentially failed or unreachable and before the primary
path is changed automatically. This socket option uses the level path is changed automatically. This socket option uses the level
IPPROTO_SCTP and the name SCTP_PEER_ADDR_THLDS. IPPROTO_SCTP and the name SCTP_PEER_ADDR_THLDS.
The following structure is used to access and modify the thresholds: The following structure is used to access and modify the thresholds:
struct sctp_paddrthlds { struct sctp_paddrthlds {
sctp_assoc_t spt_assoc_id; sctp_assoc_t spt_assoc_id;
struct sockaddr_storage spt_address; struct sockaddr_storage spt_address;
uint16_t spt_pathmaxrxt; uint16_t spt_pathmaxrxt;
skipping to change at page 14, line 17 skipping to change at page 14, line 28
This document does not create any new registries or modify the rules This document does not create any new registries or modify the rules
for any existing registries managed by IANA. for any existing registries managed by IANA.
8. Proposed Change of Status (to be Deleted before Publication) 8. Proposed Change of Status (to be Deleted before Publication)
Initially this work looked to entail some changes of the Congestion Initially this work looked to entail some changes of the Congestion
Control (CC) operation of SCTP and for this reason the work was Control (CC) operation of SCTP and for this reason the work was
proposed as Experimental. These intended changes of the CC operation proposed as Experimental. These intended changes of the CC operation
have since been judged to be irrelevant and are no longer part of the have since been judged to be irrelevant and are no longer part of the
specification. As the specification entails no other potential specification. As the specification entails no other potential
harmful features, consensus exists in the wg to bring the work harmful features, consensus exists in the WG to bring the work
forward as PS. forward as PS.
Initially concerns have been expressed about the possibility for the Initially concerns have been expressed about the possibility for the
mechanism to introduce path bouncing with potential harmful network mechanism to introduce path bouncing with potential harmful network
impacts. These concerns are believed to be unfounded. This issue is impacts. These concerns are believed to be unfounded. This issue is
addressed in Appendix B. addressed in Appendix B.
It is noted that the feature specified by this document is It is noted that the feature specified by this document is
implemented by multiple SCTP SW implementations and furthermore that implemented by multiple SCTP SW implementations and furthermore that
various variants of the solution have been deployed in Telco various variants of the solution have been deployed in Telco
skipping to change at page 15, line 50 skipping to change at page 16, line 16
This section lists alternative approaches for the issues desribed in This section lists alternative approaches for the issues desribed in
this document. Although these approaches do not require to update this document. Although these approaches do not require to update
RFC4960, we do not recommend them from the reasons described below. RFC4960, we do not recommend them from the reasons described below.
A.1. Reduce Path.Max.Retrans (PMR) A.1. Reduce Path.Max.Retrans (PMR)
Smaller values for Path.Max.Retrans shorten the failover duration. Smaller values for Path.Max.Retrans shorten the failover duration.
In fact, this is recommended in some research results [JUNGMAIER02] In fact, this is recommended in some research results [JUNGMAIER02]
[GRINNEMO04] [FALLON08]. For example, if when Path.Max.Retrans=0, [GRINNEMO04] [FALLON08]. For example, if when Path.Max.Retrans=0,
SCTP switches to another destination on a single timeout. This SCTP switches to another destination address on a single timeout.
smaller value for Path.Max.Retrans can results in spurious failover, This smaller value for Path.Max.Retrans can results in spurious
which might be a problem. failover, which might be a problem.
Unlike SCTP-PF, the interval for heartbeat packets is governed by Unlike SCTP-PF, the interval for heartbeat packets is governed by
'HB.interval' even during failover process. 'HB.interval' is usually 'HB.interval' even during failover process. 'HB.interval' is usually
set in the order of seconds (recommended value is 30 seconds). When set in the order of seconds (recommended value is 30 seconds). When
the primary path becomes inactive, the next HB can be transmitted the primary path becomes inactive, the next HB can be transmitted
only seconds later. Meanwhile, the primary path may have recovered. only seconds later. Meanwhile, the primary path may have recovered.
In such situations, post failover, an endpoint is forced to wait on In such situations, post failover, an endpoint is forced to wait on
the order of seconds before the endpoint can resume transmission on the order of seconds before the endpoint can resume transmission on
the primary path. However, using smaller value for 'HB.interval' the primary path. However, using smaller value for 'HB.interval'
might help this situation, but it will be the waste of bandwidth in might help this situation, but it will be the waste of bandwidth in
skipping to change at page 17, line 23 skipping to change at page 17, line 33
There are two main reasons for this. First, SCTP is basically There are two main reasons for this. First, SCTP is basically
designed for multipath communication, which means SCTP maintains all designed for multipath communication, which means SCTP maintains all
path related parameters (CWND, ssthresh, RTT, error count, etc) per path related parameters (CWND, ssthresh, RTT, error count, etc) per
each destination address. These parameters cannot be affected by each destination address. These parameters cannot be affected by
path bouncing. In addition, when SCTP migrates the data transfer to path bouncing. In addition, when SCTP migrates the data transfer to
another path, it starts with the minimal or the initial CWND. Hence, another path, it starts with the minimal or the initial CWND. Hence,
there is little chance for packet reordering or duplicating. there is little chance for packet reordering or duplicating.
Second, even if all communication paths between the end-nodes share Second, even if all communication paths between the end-nodes share
the same bottleneck, the quick failover results in a behavior already the same bottleneck, the SCTP-PF results in a behavior already
allowed by [RFC4960]. allowed by [RFC4960].
Authors' Addresses Authors' Addresses
Yoshifumi Nishida Yoshifumi Nishida
GE Global Research GE Global Research
2623 Camino Ramon 2623 Camino Ramon
San Ramon, CA 94583 San Ramon, CA 94583
USA USA
 End of changes. 53 change blocks. 
266 lines changed or deleted 277 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/