[Search] [txt|pdfized|bibtex] [Tracker] [Email] [Diff1] [Diff2] [Nits]
Versions: 00 01 02 03                                                   
 CCAMP Working Group                            Vishnu Pavan Beeram (Ed)
 Internet Draft                                         Juniper Networks
 Intended status: Standards Track                      Igor Bryskin (Ed)
                                                 ADVA Optical Networking
 Expires: April 21, 2014                                October 21, 2013
                    Mutually Exclusive Link Group (MELG)
 Status of this Memo
    This Internet-Draft is submitted 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-
    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."
    The list of current Internet-Drafts can be accessed at
    The list of Internet-Draft Shadow Directories can be accessed at
    This Internet-Draft will expire on April 21, 2014.
 Copyright Notice
    Copyright (c) 2013 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
 Beeram, et al           Expires April 21, 2014                 [Page 1]

 Internet-Draft                   MELG                      October 2013
    Section 4.e of the Trust Legal Provisions and are provided without
    warranty as described in the Simplified BSD License.
    This document introduces the concept of MELG ("Mutually Exclusive
    Link Group") and discusses its usage in the context of mutually
    exclusive Virtual TE Links.
 Conventions used in this document
    The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
    document are to be interpreted as described in RFC-2119 [RFC2119].
 Table of Contents
    1. Introduction...................................................2
    2. Virtual TE Link - Semantics....................................3
    3. Mutually Exclusive Virtual TE Links............................3
       3.1. Static vs Dynamic.........................................4
    4. Static Mutual Exclusivity......................................4
    5. Mutually Exclusive Link Group..................................7
    6. Protocol Extensions............................................8
       6.1. OSPF......................................................8
       6.2. ISIS......................................................9
    7. Security Considerations.......................................10
    8. IANA Considerations...........................................10
       8.1. OSPF.....................................................10
       8.2. ISIS.....................................................10
    9. Normative References..........................................10
    10. Acknowledgments..............................................11
 1. Introduction
    A Virtual TE Link (as defined in [RFC6001]) advertised into a Client
    Network Domain represents a potentiality to setup an LSP in the
    Server Network Domain to support the advertised TE link. The Virtual
    TE Link gets advertised like any other TE link and follows the same
    rules that are defined for the advertising, processing and use of
    regular TE links [RFC4202]. However, "mutual exclusivity" is one
    attribute that is specific to Virtual TE links. This document
    discusses the different types of mutual exclusivity (Static vs
    Dynamic) that come into play and explains the need to advertise this
 Beeram, et al           Expires April 21, 2014                 [Page 2]

 Internet-Draft                   MELG                      October 2013
    information into the Client TEDB. It then goes onto introduce a new
    TE construct (MELG) to carry static mutual exclusivity information.
 2. Virtual TE Link - Semantics
    A Virtual TE Link (as per existing definitions) represents the
    potentiality to setup a server layer LSP, but there are currently no
    strict guidelines imposed on how the underlying server layer LSP
    would need to get set up. The characteristics of the underlying
    server-path are not necessarily pinned down until the Virtual TE
    Link gets actually committed. This means that some important
    characteristics of the Virtual TE Link like shared-risk and delay
    (and mutual exclusivity information) may not be known until the
    corresponding server layer LSP is set up. This makes resource
    planning (for example - pre-configuring network failure recovery
    schemes) in a multi-layer network that includes Virtual TE Links a
    very hard problem.
    This document uses a slightly enhanced view of a Virtual TE Link. In
    the context of this document, the Virtual TE Link (even when it is
    uncommitted) is always aware of the key characteristics of the
    underlying server-path. The creation and maintenance of this Virtual
    TE Link is strictly driven by policy. Policy not only determines
    which Virtual TE Link to create (What termination points?), but it
    may also constrain how the corresponding underlying server layer LSP
    (What path?) needs to get set up. The basic idea behind this
    "enhanced view" is that it makes the "Virtual TE Link" get as close
    as it can to representing a "Real TE Link".
    Also, as per this document, a Virtual TE Link remains a Virtual TE
    Link through-out its life-time (until it gets deleted by the
    user/policy). It may get committed (underlying server LSP gets set
    up) and uncommitted (underlying server LSP gets deleted) from time
    to time, but it never really loses it "Virtual" property.
 3. Mutually Exclusive Virtual TE Links
    Mutual Exclusivity comes into play when multiple Virtual TE Links
    are dependent on the usage of the same underlying server resource.
    Since not all of these Virtual TE Links can get committed at the
    same time, they are deemed to be mutually exclusive.
    The existence of this "mutual exclusivity" property would need to be
    advertised into the Client TE Domain. This is of relevance to Client
    Path Computation engines; especially those that are capable of doing
    concurrent computations. If this information is absent, there exists
 Beeram, et al           Expires April 21, 2014                 [Page 3]

 Internet-Draft                   MELG                      October 2013
    the risk of the Computation engine yielding erroneous concurrent
    path computation results where only a subset of the computed paths
    get successfully provisioned.
    The "Mutual Exclusivity" property of a Virtual TE Link can be either
    static or dynamic in nature.
 3.1. Static vs Dynamic
    Static Mutual Exclusivity: This type of mutual exclusivity exists
    permanently within a given network configuration. It comes into play
    when two or more Virtual TE Links depend on the usage of the same
    non-shareable underlying server network domain resource. This
    resource gets used up in its entirety by a single Virtual TE Link
    when committed. Such resources exist only in the WDM layer.
    Dynamic Mutual Exclusivity: This type of mutual exclusivity exists
    temporarily within a given network configuration. It comes into play
    when two or more Virtual TE Links depend on the usage of the same
    shareable underlying server network domain resource. Mutual
    Exclusivity exists when the amount of the server resource that is
    available for sharing is limited; it ceases to exist when sufficient
    amount of the resource is available for accommodating all
    corresponding Virtual TE Links. Such resources exist in all layers.
    Because of their inherent difference, the advertisement paradigm of
    the TE construct required to carry static mutual exclusivity
    information is quite different from that of the TE construct
    required to carry dynamic mutual exclusivity information. Static
    mutual exclusivity Information can get advertised per TE-Link using
    a simple sub-TLV construct. There wouldn't be any scaling issues
    with this approach because of the static nature of the information
    that gets advertised. On the contrary, advertising dynamic mutual
    exclusivity information per TE-Link poses serious scaling concerns
    and hence requires a different type of construct/paradigm.
    This document introduces a new TE construct for carrying static
    mutual exclusivity information. The mechanisms to address dynamic
    mutual exclusivity are discussed in a separate document [SRcLG].
 4. Static Mutual Exclusivity
    Consider the network topology depicted in Figure 1a. This is a
    typical packet optical transport deployment scenario where the WDM
    layer network domain serves as a Server Network Domain providing
 Beeram, et al           Expires April 21, 2014                 [Page 4]

 Internet-Draft                   MELG                      October 2013
    transport connectivity to the packet layer network Domain (Client
    Network Domain).
                               | +---+            /-\
                               | |   | Router    (   ) WDM
                               | +---+ Node       \-/  node
  +---+        /-\          /-\           /-\          +---+
  | R1|-------( A )--------( C )---------( E )---------| R3|
  +---+        \-/          \-/           \-/          +---+
                           /   \         /   \
                          /     \       /     \
                         /       \     /       \
                        /         \   /         \
                       /           \ /           \
      +---+          /-\           /-\           /-\          +---+
      | R2|---------( B )---------( D )---------( F )---------| R4|
      +---+          \-/           \-/           \-/          +---+
                     Figure 1a: Sample topology
      -------------                        |  [ ] Client TE Node
      | Client TE |                        |  +++ Client TE Link
      | DataBase  |                        |_____________________
         [R1] ++++++++ [A]                  [E] +++++++++ [R3]
         [R2] ++++++++ [B]                  [F] +++++++++ [R4]
                      Figure 1b: Client TE Database
    Nodes R1, R2, R3 and R4 are IP routers that are connected to an
    Optical WDM transport network. A, B, C, D, E and F are WDM nodes
    that constitute the Server Network Domain. The border nodes (A, B, E
 Beeram, et al           Expires April 21, 2014                 [Page 5]

 Internet-Draft                   MELG                      October 2013
    and F) operate in both the server and client domains. Figure 1b
    depicts how the Client Network Domain TE topology looks like when
    there are no Client TE Links provisioned across the optical domain.
                                | *****  B-F WDM Path
                                | @@@@@  B-E WDM Path
  +---+        /-\          /-\ @@@@@@@@@ /-\          +---+
  | R1|-------( A )--------( C )---------( E )---------| R3|
  +---+        \-/         @\-/           \-/          +---+
                          @/   \         /   \
                         @/     \       /     \
                        @/       \     /       \
                       @/         \   /         \
                      @/           \ /           \
      +---+          /-\ ********* /-\ ********* /-\          +---+
      | R2|---------( B )---------( D )---------( F )---------| R4|
      +---+          \-/           \-/           \-/          +---+
            Figure 2a: Mutually Exclusive potential WDM paths
       ------------   |  TE-Links B-F and B-E are mutually exclusive;
       | Client-TE|   |  They depend on the usage of the same
       | Database |   |  underlying non-shareable server resource
       ------------   |_____________________________________________
         [R1] ++++++++ [A]                      [E] +++++++++ [R3]
         [R2] ++++++++ [B] ++++++++++++++++++++ [F] +++++++++ [R4]
   Figure 2b: Client TE Database - Mutually Exclusive Virtual TE Links
 Beeram, et al           Expires April 21, 2014                 [Page 6]

 Internet-Draft                   MELG                      October 2013
    Now consider augmenting the Client TE topology by creating a couple
    of Virtual TE Links across the optical domain. The potential paths
    in the WDM network catering to these two virtual TE links are as
    shown in Fig 2a and the corresponding augmented Client TE topology
    is as illustrated in Fig 2b.
    In this particular example, the potential paths in the WDM layer
    network supporting the Virtual TE Links require the usage of the
    same source transponder (on "Node B"). Because the Virtual TE Links
    depend on the same uncommitted network resource, only one of them
    could get activated at any given time. In other words they are
    mutually exclusive. This scenario is encountered when the potential
    paths depend on any common physical resource (e.g. transponder,
    regenerator, wavelength converter, etc.) that could be used by only
    one Server Network Domain LSP at a time.
    This document proposes the use of "Mutually Exclusive Link Group
    (MELG)" for catering to this scenario.
 5. Mutually Exclusive Link Group
    The Mutually Exclusive Link Group (MELG) construct defined in this
    document has 2 purposes
    - To indicate via a separate network unique number (MELG ID) an
      element or a situation that makes the advertised Virtual TE Link
      belong to one or more Mutually Exclusive Link Groups. Path
      computing element will be able to decide on whether two or more
      Virtual TE Links are mutually exclusive or not by finding an
      overlap of advertised MELGs (similar to deciding on whether two or
      more TE links share fate or not by finding common SRLGs)
    - To indicate whether the advertised Virtual TE Link is committed or
      not at the moment of the advertising. Such information is
      important for a path computation element: Committing new Virtual
      TE links (vs. re-using already committed ones) has a consequence
      of allocating more server layer resources and disabling other
      Virtual TE Links that have common MELGs with newly committed
      Virtual TE Links; Committing a new Virtual TE Link also means a
      longer setup time for the Client LSP and higher risk of setup-
 Beeram, et al           Expires April 21, 2014                 [Page 7]

 Internet-Draft                   MELG                      October 2013
 6. Protocol Extensions
 6.1. OSPF
    The MELG is a sub-TLV of the top level TE Link TLV. It may occur at
    most once within the Link TLV. The format of the MELGs sub-TLV is
    defined as follows:
    Name: MELG
    Type: TBD
    Length: Variable
    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
    |            Sub-TLV Type       |            Sub-TLV Length     |
    |    VTE-Flags (16 bits)     |U |  Number of MELGs (16 bits)    |
    |                 MELGID1 (64 bits)                             |
    |                 MELGID2 (64 bits)                             |
    |                ........................                       |
    |                 MELGIDn (64 bits)                             |
    Number of MELGs:              number of MELGS advertised for the
                                  Virtual TE Link;
    VTE-Flags:                    Virtual TE Link specific flags;
    MELGID1,MELGID2,...,MELGIDn:  64-bit network domain unique numbers
                                  associated with each of the advertised
    Currently defined Virtual TE Link specific flags are:
       U bit (bit 1): Uncommitted - if set, the Virtual TE Link is
       uncommitted at the time of the advertising (i.e. the server layer
       network LSP is not set up); if cleared, the Virtual TE Link is
       committed (i.e. the server layer LSP is fully provisioned and
       functioning). All other bits of the "VTE-Flags" field are
       reserved for future use and MUST be cleared.
    Note: A Virtual TE Link advertisement MAY include MELGs sub-TLV with
    zero MELGs for the purpose of communicating to the TE domain whether
    the Virtual TE Link is currently committed or not.
 Beeram, et al           Expires April 21, 2014                 [Page 8]

 Internet-Draft                   MELG                      October 2013
 6.2. ISIS
    The MELG TLV (of type TBD) contains a data structure consisting of:
       6        octets of System ID
       1        octet of Pseudonode Number
       1        octet Flag
       4        octets of IPv4 interface address or 4 octets of a Link
                Local Identifier
       4        octets of IPv4 neighbor address or 4 octets of a Link
                Remote Identifier
       2        octets MELG-Flags
       2        octets - Number of MELGs
       variable List of MELG values, where each element in the list
                has 8 octets
    The following illustrates encoding of the value field of the MELG
     0               1               2               3
     0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
    |                           System ID                           |
    |        System ID (cont.)      |Pseudonode num |    Flags      |
    |          Ipv4 interface address/Link Local Identifier         |
    |          Ipv4 neighbor address/Link Remote Identifier         |
    |    VTE-Flags (16 bits)     |U |  Number of MELGs (16 bits)    |
    |                 MELGID1 (64 bits)                             |
    |                 MELGID2 (64 bits)                             |
    |                ........................                       |
    |                 MELGIDn (64 bits)                             |
    The neighbor is identified by its System ID (6 octets), plus one
    octet to indicate the pseudonode number if the neighbor is on a LAN
 Beeram, et al           Expires April 21, 2014                 [Page 9]

 Internet-Draft                   MELG                      October 2013
    The least significant bit of the Flag octet indicates whether the
    interface is numbered (set to 1) or unnumbered (set to 0). All other
    bits are reserved and should be set to 0.
    The length of the TLV is 20 + 8 * (number of MELG values).
    The semantics of "VTE-Flags", "Number of MELGs" and "MELGID Values"
    are the same as the ones defined under OSPF extensions.
    The MELG TLV MAY occur more than once within the IS-IS Link State
    Protocol Data Units.
 7. Security Considerations
 8. IANA Considerations
 8.1. OSPF
    IANA is requested to allocate a new sub-TLV type for MELG (as
    defined in Section 6.1) under the top-level TE Link TLV.
 8.2. ISIS
    IANA is requested to allocate a new IS-IS TLV type for MELG (as
    defined in Section 6.2).
 9. Normative References
    [RFC2119]    Bradner, S., "Key words for use in RFCs to Indicate
                 Requirement Levels", BCP 14, RFC 2119, March 1997.
    [RFC4202]    K.Kompella, Y.Rekhter, "Routing Extensions in Support
                 of Generalized Multi-Protocol Label Switching (GMPLS)",
                 RFC4202, October 2005.
    [RFC6001]    D.Papadimitriou, M.Vigoureax, K.Shiomoto, D.Brungard
                 and JL. Le Roux, "GMPLS Protocol Extensions for Multi-
                 Layer and Multi-Region Networks", RFC 6001, October
    [SRcLG]      Beeram, V., "Shared Resource Link Group",
                 draft-beeram-ccamp-srclg, October 2013
 Beeram, et al           Expires April 21, 2014                [Page 10]

 Internet-Draft                   MELG                      October 2013
 10. Acknowledgments
    Chris Bowers [cbowers@juniper.net]
 Authors' Addresses
    Vishnu Pavan Beeram
    Juniper Networks
    Email: vbeeram@juniper.net
    Igor Bryskin
    ADVA Optical Networking
    Email: ibryskin@advaoptical.com
    John Drake
    Juniper Networks
    Email: jdrake@juniper.net
    Gert Grammel
    Juniper Networks
    Email: ggrammel@juniper.net
    Wes Doonan
    Email: wddlists@gmail.com
    Manuel Paul
    Deutsche Telekom
    Email: Manuel.Paul@telekom.de
    Ruediger Kunze
    Deutsche Telekom
    Email: Ruediger.Kunze@telekom.de
    Oscar Gonzalez de Dios
    Email: ogondio@tid.es
    Cyril Margaria
    Email: cyril.margaria@gmail.com
    Friedrich Armbruster
 Beeram, et al           Expires April 21, 2014                [Page 11]

 Internet-Draft                   MELG                      October 2013
    Coriant GmbH
    Email: friedrich.armbruster@coriant.com
    Daniele Ceccarelli
    Email: daniele.ceccarelli@ericsson.com
    Fatai Zhang
    Huawei Technologies
    Email: zhangfatai@huawei.com
 Beeram, et al           Expires April 21, 2014                [Page 12]