Internet DRAFT - draft-allan-spme-smp-fmwk
draft-allan-spme-smp-fmwk
MPLS Working Group Dave Allan, Greg Mirsky
Internet Draft Ericsson
Intended status: Informational
Expires: May 2012
November 2011
A framework for the use of SPMEs for shared mesh protection
draft-allan-spme-smp-fmwk-00
Abstract
Shared mesh protection allows a set of diversely routed paths with
diverse endpoints to collectively oversubcribe protection resources.
Under normal conditions no single failure will result in the capacity
of the associated protection resources to be exhausted.
When multiple failures occur such that more than one path in the set
of paths utilizing shared protection resources is affected, the
necessity arises of pre-empting traffic on the basis of business
priority rather than application priority.
This memo describes the use of SPMEs and TC marking as a means of
indicating business priority for shared mesh protection.
Requirements Language
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in RFC2119 [1].
Status of this Memo
This Internet-Draft is submitted to IETF in full conformance
with the provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet
Engineering Task Force (IETF), its areas, and its working
groups. Note that other groups may also distribute working
documents as Internet-Drafts.
Internet-Drafts are draft documents valid for a maximum of six
months and may be updated, replaced, or obsoleted by other
documents at any time. It is inappropriate to use Internet-
Drafts as reference material or to cite them other than as "work
in progress".
Allan et al., Expires May 2012 [Page 1]
Internet-Draft draft-allan-spme-smp-fmwk-00 November 2011
The list of current Internet-Drafts can be accessed at
http://www.ietf.org/ietf/1id-abstracts.txt.
The list of Internet-Draft Shadow Directories can be accessed at
http://www.ietf.org/shadow.html.
This Internet-Draft will expire on August 2nd 2011.
Copyright Notice
Copyright (c) 2011 IETF Trust and the persons identified as the
document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal
Provisions Relating to IETF Documents
(http://trustee.ietf.org/license-info) in effect on the date of
publication of this document. Please review these documents
carefully, as they describe your rights and restrictions with
respect to this document. Code Components extracted from this
document must include Simplified BSD License text as described
in Section 4.e of the Trust Legal Provisions and are provided
without warranty as described in the Simplified BSD License.
Table of Contents
1. Introduction...................................................3
1.1. Authors......................................................3
2. Conventions used in this document..............................3
2.1. Terminology..................................................3
3. Overview.......................................................4
3.1. Architectural Overview.......................................4
4. Signalling Implications........................................5
5. IANA Considerations............................................5
6. Security Considerations........................................5
7. References.....................................................6
7.1. Normative References.........................................6
7.2. Informative References.......................................6
8. Authors' Addresses.............................................6
Allan et al., Expires May 2012 [Page 2]
Internet-Draft draft-allan-spme-smp-fmwk-00 November 2011
1. Introduction
Shared mesh protection is described in [2]. A common interpretation
of the behavior of shared mesh protection emerges from the circuit
switched world whereby subtending path selectors and selector
coordination functions support path preemption to ensure that the
highest priority path needing the protection resources is granted
ownership of the shared segment, all others being preempted, and such
functionality can be successfully delegated to dataplane OAM.
Ultimately this resolves into a business priority decision vs. an
application priority decision in how customer traffic is handled. The
packet world is different from the circuit world in that there is no
guarantee of convenient alignment of resource requirements between
preempting and preempted paths. Nor in a packet environment is there
the need to completely preempt all the traffic in a lower priority
path simply because a higher priority path lays claim to the
resources. Finally it is useful to obviate the requirement for
preempting and preemptable traffic to be co-routed.
This memo proposes the use of SPMEs with the pipe model of TC copying
as an alternative to the use of path pre-emption, path selectors and
selector coordination functions for the purposes of implementing
business policy.
1.1. Authors
David Allan, Greg Mirsky
2. Conventions used in this document
2.1. Terminology
MPLS-TP: MPLS Transport Profile
MPLS-TP LSP: Uni-directional or Bidirectional Label Switch Path
representing a circuit
SMP: Shared Mesh Protection
SPME: Sub-Path Maintenance Entity
TC: Traffic Class
TTL: Time To Live
Allan et al., Expires May 2012 [Page 3]
Internet-Draft draft-allan-spme-smp-fmwk-00 November 2011
3. Overview
Shared mesh protection is described in [2]. A common interpretation
of the behavior of shared mesh protection emerges from the circuit
switched world. In that interpretation subtending path selectors and
selector coordination support path preemption functionality to ensure
that the highest priority path needing the protection resources is
the one granted ownership of the shared segment; all others being
preempted. It also assumes that all paths sharing the protection
resources conveniently all need exactly the same size pipe.
In packet transport networks there will frequently not be a
convenient 1:1 equivalence of the bandwidth requirements of the set
of transport paths sharing protection resources such that a simple
pre-emption decision can be made. For example 3 paths: A, B, and C
sized "n", "n/2" and "n/2" respectively could have a shared segment
size "3n/2" such that simultaneous failures necessitating the
activation of any two of the protection paths could be accommodated
without path preemption. When one ranks A, B and C with a variety of
priorities and considers all failure combinations a rather large
matrix of possible required behaviors emerges.
If one pursues this line of thinking to its logical conclusion, and
envisions a significant set of paths of diverse sizes and diverse
priorities, the policy associated with successful path prioritization
and preemption becomes quite complex, and ensuring multiple selectors
make timely and of necessity common preemption decisions starts to
impose network design constraints or require additional coordination
protocols that severely impact the utility of SMP.
Further in a packet network there can be a difference in the
bandwidth reserved and the bandwidth actually used at any given
instant in time. One consequence is that there is no need to
completely preempt all the traffic in a lower priority path simply
because a higher priority path lays a preferential claim to the
bandwidth.
To obviate these problems, this memo proposes an alternative to how
business priority can be implemented for shared mesh protection that
obviates the need for path preemption and the limitations such an
approach imposes.
3.1. Architectural Overview
This memo pre-supposes an operational mode of behavior along the
lines of the following:
Allan et al., Expires May 2012 [Page 4]
Internet-Draft draft-allan-spme-smp-fmwk-00 November 2011
1) As a matter of network design, specific links (or engineered
virtual links) are set aside for the purpose of acting as shared
protection resources. The key attribute of these links is that
the processing of TC markings will be exclusively for shared
protection.
2) The arrangement of the shared protection links can be arbitrary
such that contiguous domains can be constructed with an arbitrary
number of ingress and egress points. A set of contiguous
protection links is known as a protection domain.
3) Either an apriori or on-demand mesh of SPMEs that connect all
ingress and egress points in a protection domain is required.
These are logically forwarding adjacencies for the purposes of
routing protection paths
4) The instantiation of protection paths requires the mapping of the
incoming path at an ingress node for the protection domain to an
SPME that connects the ingress to the required egress node from
the domain.
5) The pipe model of TC copying is used such that the SPME gets the
TC marking associated with the business priority for the path
associated with the incoming label value. As the SPME only
transits resources where the TC marking has been overloaded in
this fashion business priority does not conflict with application
requirements.
6) Admission control for the protection paths transiting the
protection domain is performed such that the sum of the bandwidth
for a given business priority does not oversubscribe any links in
the protected domain, but the sum of the bandwidth for all
business priorities can. In this way no traffic of the highest
business priority using the shared mesh pool will be contended.
4. Signalling Implications
For a future version of this document.
5. IANA Considerations
No IETF protocols were harmed in the publishing of this memo.
6. Security Considerations
For a future version of this document.
Allan et al., Expires May 2012 [Page 5]
Internet-Draft draft-allan-spme-smp-fmwk-00 November 2011
7. References
7.1. Normative References
[1] Bradner, S., "Key words for use in RFCs to Indicate
Requirement Levels", BCP 14, RFC 2119, March 1997.
7.2. Informative References
[2] Sprecher, N., et al. "MPLS Transport Profile (MPLS-TP)
Survivability Framework", RFC 6372, September 2011
8. Authors' Addresses
Dave Allan
Ericsson
Email: david.i.allan@ericsson.com
Greg Mirsky
Ericsson
Email: Gregory.mirsky@ericsson.com
Allan et al., Expires May 2012 [Page 6]
