RFC 7190

Internet Engineering Task Force (IETF)                     C. Villamizar
Request for Comments: 7190             Outer Cape Cod Network Consulting
Category: Informational                                       March 2014
ISSN: 2070-1721

    Use of Multipath with MPLS and MPLS Transport Profile (MPLS-TP)


   Many MPLS implementations have supported multipath techniques, and
   many MPLS deployments have used multipath techniques, particularly in
   very high-bandwidth applications, such as provider IP/MPLS core
   networks.  MPLS Transport Profile (MPLS-TP) has strongly discouraged
   the use of multipath techniques.  Some degradation of MPLS-TP
   Operations, Administration, and Maintenance (OAM) performance cannot
   be avoided when operating over many types of multipath

   Using MPLS Entropy Labels (RFC 6790), MPLS Label Switched Paths
   (LSPs) can be carried over multipath links while also providing a
   fully MPLS-TP-compliant server layer for MPLS-TP LSPs.  This document
   describes the means of supporting MPLS as a server layer for MPLS-TP.
   The use of MPLS-TP LSPs as a server layer for MPLS LSPs is also

Status of This Memo

   This document is not an Internet Standards Track specification; it is
   published for informational purposes.

   This document is a product of the Internet Engineering Task Force
   (IETF).  It represents the consensus of the IETF community.  It has
   received public review and has been approved for publication by the
   Internet Engineering Steering Group (IESG).  Not all documents
   approved by the IESG are a candidate for any level of Internet
   Standard; see Section 2 of RFC 5741.

   Information about the current status of this document, any errata,
   and how to provide feedback on it may be obtained at

Villamizar                    Informational                     [Page 1]

RFC 7190               MPLS and MPLS-TP Multipath             March 2014

Copyright Notice

   Copyright (c) 2014 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  . . . . . . . . . . . . . . . . . . . . . . . .   2
   2.  Definitions . . . . . . . . . . . . . . . . . . . . . . . . .   4
   3.  MPLS as a Server Layer for MPLS-TP  . . . . . . . . . . . . .   5
     3.1.  MPLS-TP Forwarding and Server-Layer Requirements  . . . .   5
     3.2.  Methods of Supporting MPLS-TP Client LSPs over MPLS . . .   7
   4.  MPLS-TP as a Server Layer for MPLS  . . . . . . . . . . . . .  11
   5.  Summary . . . . . . . . . . . . . . . . . . . . . . . . . . .  11
   6.  Acknowledgements  . . . . . . . . . . . . . . . . . . . . . .  12
   7.  Security Considerations . . . . . . . . . . . . . . . . . . .  13
   8.  References  . . . . . . . . . . . . . . . . . . . . . . . . .  13
     8.1.  Normative References  . . . . . . . . . . . . . . . . . .  13
     8.2.  Informative References  . . . . . . . . . . . . . . . . .  13

1.  Introduction

   Today the requirement to handle large aggregations of traffic can be
   met by a number of techniques that we will collectively call
   "multipath".  Multipath applied to parallel links between the same
   set of nodes includes Ethernet Link Aggregation [IEEE-802.1AX], link
   bundling [RFC4201], or other aggregation techniques, some of which
   could be vendor specific.  Multipath applied to diverse paths rather
   than parallel links includes Equal-Cost Multipath (ECMP) as applied
   to OSPF, IS-IS, or BGP, and equal-cost Label Switched Paths (LSPs).
   Some vendors support load splitting across equal-cost MPLS LSPs where
   the load is split proportionally to the reserved bandwidth of the set
   of LSPs.

   RFC 5654 requirement 33 requires the capability to carry a client
   MPLS Transport Profile (MPLS-TP) or MPLS layer over a server MPLS-TP
   or MPLS layer [RFC5654].  This is possible in all cases with one
   exception.  When an MPLS LSP exceeds the capacity of any single

Villamizar                    Informational                     [Page 2]

RFC 7190               MPLS and MPLS-TP Multipath             March 2014

   component link, it MAY be carried by a network using multipath
   techniques, but it SHOULD NOT be carried by a single MPLS-TP LSP due
   to the inherent MPLS-TP capacity limitation imposed by MPLS-TP
   Operations, Administration, and Maintenance (OAM) fate-sharing
   constraints and MPLS-TP Loss Measurement OAM packet-ordering
   constraints (see Section 3.1).  Instead, multiple MPLS-TP LSPs SHOULD
   be used to carry a large MPLS LSP (see Section 4).

   The term "composite link" is more general than terms such as "link
   aggregation" (which is specific to Ethernet) or "ECMP" (which implies
   equal-cost paths within a routing protocol).  The use of the term
   "composite link" here is consistent with the broad definition in
   [ITU-T.G.800].  Multipath is very similar to composite link as
   defined by ITU-T but specifically excludes inverse multiplexing.

   MPLS LSPs today are able to function as a server layer and carry
   client MPLS LSPs.  When control-plane signaling is used, forwarding
   adjacency (FA) advertisements are used to inform the set of Label
   Switching Routers (LSRs) of Packet Switching Capable (PSC) LSPs
   within the MPLS topology [RFC4206].  Client MPLS LSP at a higher
   layer (lower PSC number) may signal their intention to use PSC LSPs
   as hops in the RSVP-TE Explicit Route Object (ERO).  LSRs with no
   explicit support for RFC 4206 see the PSC LSPs as ordinary links and
   therefore use them.

   An MPLS LSP that has been set up using RSVP-TE appears to its ingress
   LSR as a viable IP next hop to a distant LSR.  If LDP is used and
   bidirectional RSVP-TE LSP connectivity is available, then LDP
   signaling can be set up among the RSVP-TE LSP endpoints, and LDP can
   make use of the RSVP-TE LSP as an LDP hop.  This is another form of
   existing MPLS-in-MPLS use.  MPLS LSPs may also make use of hierarchy
   that is configured through the management plane rather than signaled
   using RSVP-TE.

   These existing forms of MPLS-in-MPLS may traverse multipath hops such
   as Ethernet Link Aggregation Group (LAG) [IEEE-802.1AX] or MPLS Link
   Bundling [RFC4201].  MPLS-TP brings with it a new set of requirements
   not considered in past deployments of the various forms of MPLS-in-
   MPLS where multipath was in use.  This document merely discusses use
   of existing forwarding and protocol mechanisms that can support the
   case where either the client-layer LSPs or the server-layer LSPs are
   MPLS-TP and where multipath is used.

Villamizar                    Informational                     [Page 3]

RFC 7190               MPLS and MPLS-TP Multipath             March 2014

2.  Definitions

   Please refer to the terminology related to multipath introduced in
   [ADV-MULTIPATH-REQ].  The following additional terms are used in this
   document; related terms are grouped together.

   Link Bundle
       Link bundling is a multipath technique specific to MPLS
       [RFC4201].  Link bundling supports two modes of operations.
       Either an LSP can be placed on one component link of a link
       bundle, or an LSP can be load-split across all members of the
       bundle.  There is no signaling defined that allows a per-LSP
       preference regarding load split, therefore whether to load split
       is generally configured per bundle and applied to all LSPs across
       the bundle.

   All-Ones Component
       Within the context of link bundling, [RFC4201] defines a special
       case where the same label is to be valid across all component
       links.  This case is indicated in signaling by a bit value of
       "all ones" when identifying a component link.  Following the
       publication of RFC 4201, for brevity this special case has been
       referred to as the "all-ones component".

   Equal-Cost Multipath (ECMP)
       Equal-Cost Multipath (ECMP) is a specific form of multipath in
       which the costs of the links or paths must be equal in a given
       routing protocol.  The load may be split equally across all
       available links (or available paths), or the load may be split
       proportionally to the capacity of each link (or path).

   Loop-Free Alternate Paths (LFA)
       "Loop-free alternate paths" (LFA) are defined in Section 5.2 of
       RFC 5714 [RFC5714] as follows: "Such a path exists when a direct
       neighbor of the router adjacent to the failure has a path to the
       destination that can be guaranteed not to traverse the failure."
       Further detail can be found in [RFC5286].  LFA as defined for IP
       Fast Reroute (IPFRR) can be used to load balance by relaxing the
       equal-cost criteria of ECMP, though IPFRR defined LFA for use in
       selecting protection paths.  When used with IP, proportional
       split is generally not used.  LFA use in load balancing is
       implemented by some vendors, though it may be rare or non-
       existent in deployments.

Villamizar                    Informational                     [Page 4]

RFC 7190               MPLS and MPLS-TP Multipath             March 2014

   Link Aggregation
       The term "link aggregation" generally refers to Ethernet Link
       Aggregation as defined by [IEEE-802.1AX].  Ethernet Link
       Aggregation defines a Link Aggregation Control Protocol (LACP)
       which coordinates inclusion of Link Aggregation Group (LAG)
       members in the LAG.

   Link Aggregation Group (LAG)
       A group of physical Ethernet interfaces that are treated as a
       logical link when using Ethernet Link Aggregation is referred to
       as a Link Aggregation Group (LAG).

   LAG Member
       Ethernet Link Aggregation as defined in [IEEE-802.1AX] refers to
       an individual link in a LAG as a LAG member.  A LAG member is a
       component link.  An Ethernet LAG is a composite link.  IEEE does
       not use the terms "composite link" or "component link".

   A small set of requirements are discussed.  These requirements make
   use of keywords such as MUST and SHOULD as described in [RFC2119].

3.  MPLS as a Server Layer for MPLS-TP

   An MPLS LSP may be used as a server layer for MPLS-TP LSPs as long as
   all MPLS-TP requirements are met.  Section 3.1 reviews the basis for
   requirements of a server layer that supports MPLS-TP as a client
   layer.  Key requirements include OAM "fate-sharing" and that packets
   within an MPLS-TP LSP (including both payload and OAM packets) not be
   reordered.  Section 3.2 discusses implied requirements where MPLS is
   the server layer for MPLS-TP client LSPs and describes a set of
   solutions that use existing MPLS mechanisms.

3.1.  MPLS-TP Forwarding and Server-Layer Requirements

   [RFC5960] defines the data-plane requirements for MPLS-TP.  Two very
   relevant paragraphs in Section 3.1.1 ("LSP Packet Encapsulation and
   Forwarding") are the following:

   RFC 5960, Section 3.1.1, Paragraph 3
       Except for transient packet reordering that may occur, for
       example, during fault conditions, packets are delivered in order
       on L-LSPs, and on E-LSPs within a specific ordered aggregate.

Villamizar                    Informational                     [Page 5]

RFC 7190               MPLS and MPLS-TP Multipath             March 2014

   RFC 5960, Section 3.1.1, Paragraph 6
       Equal-Cost Multi-Path (ECMP) load-balancing MUST NOT be performed
       on an MPLS-TP LSP.  MPLS-TP LSPs as defined in this document MAY
       operate over a server layer that supports load-balancing, but
       this load-balancing MUST operate in such a manner that it is
       transparent to MPLS-TP.  This does not preclude the future
       definition of new MPLS-TP LSP types that have different
       requirements regarding the use of ECMP in the server layer.

   [RFC5960], Section 3.1.1, Paragraph 3 requires that packets within a
   specific ordered aggregate be delivered in order.  This same
   requirement is already specified by Differentiated Services
   [RFC2475].  [RFC5960], Section 3.1.1, Paragraph 6 explicitly allows a
   server layer to use ECMP, provided that it is transparent to the
   MPLS-TP client layer.

   [RFC6371] adds a requirement for data traffic and OAM traffic "fate-
   sharing".  The following paragraph in Section 1 ("Introduction")
   summarizes this requirement.

   RFC 6371, Section 1, Paragraph 7
       OAM packets that instrument a particular direction of a transport
       path are subject to the same forwarding treatment (i.e., fate-
       share) as the user data packets and in some cases, where
       Explicitly TC-encoded-PSC LSPs (E-LSPs) are employed, may be
       required to have common per-hop behavior (PHB) Scheduling Class
       (PSC) End-to-End (E2E) with the class of traffic monitored.  In
       case of Label-Only-Inferred-PSC LSP (L-LSP), only one class of
       traffic needs to be monitored, and therefore the OAM packets have
       common PSC with the monitored traffic class.

   [RFC6371] does not prohibit multilink techniques in Section 4.6
   ("Fate-Sharing Considerations for Multilink"), where multilink is
   defined as Ethernet Link Aggregation and the use of Link Bundling for
   MPLS, but it does declare that such a network would be only partially
   MPLS-TP compliant.  The characteristic that is to be avoided is
   contained in the following sentence in that section.

   RFC 6371, Section 4.6, Paragraph 1, last sentence
       These techniques frequently share the characteristic that an LSP
       may be spread over a set of component links and therefore be
       reordered, but no flow within the LSP is reordered (except when
       very infrequent and minimally disruptive load rebalancing

   A declaration that implies that Link Bundling for MPLS yields a
   partially MPLS-TP-compliant network is perhaps overstated since only
   the Link Bundling all-ones component link has this characteristic.

Villamizar                    Informational                     [Page 6]

RFC 7190               MPLS and MPLS-TP Multipath             March 2014

   [RFC6374] defines a direct Loss Measurement (LM) where LM OAM packets
   cannot be reordered with respect to payload packets.  This will
   require that payload packets themselves not be reordered.  The
   following paragraph in Section 2.9.4 ("Equal Cost Multipath") gives
   the reason for this restriction.

   RFC 6374, Section 2.9.4, Paragraph 2
       The effects of ECMP on loss measurement will depend on the LM
       mode.  In the case of direct LM, the measurement will account for
       any packets lost between the sender and the receiver, regardless
       of how many paths exist between them.  However, the presence of
       ECMP increases the likelihood of misordering both of LM messages
       relative to data packets and of the LM messages themselves.  Such
       misorderings tend to create unmeasurable intervals and thus
       degrade the accuracy of loss measurement.  The effects of ECMP
       are similar for inferred LM, with the additional caveat that,
       unless the test packets are specially constructed so as to probe
       all available paths, the loss characteristics of one or more of
       the alternate paths cannot be accounted for.

3.2.  Methods of Supporting MPLS-TP Client LSPs over MPLS

   Supporting MPLS-TP LSPs over a fully MPLS-TP conformant MPLS LSP
   server layer where the MPLS LSPs are making use of multipath requires
   special treatment of the MPLS-TP LSPs such that those LSPs meet MPLS-
   TP forwarding requirements (see Section 3.1).  This implies the
   following brief set of requirements.

   MP#1  It MUST be possible for a midpoint MPLS-TP Label Switching
         Router (LSR) that is serving as ingress to a server-layer MPLS
         LSP to identify MPLS-TP LSPs, so that MPLS-TP forwarding
         requirements can be applied, or to otherwise accommodate the
         MPLS-TP forwarding requirements.

   MP#2  The ability to completely exclude MPLS-TP LSPs from the
         multipath hash and load split SHOULD be supported.  If the
         selected component link no longer meets requirements, an LSP is
         considered down, which may trigger protection and/or may
         require that the ingress LSR select a new path and signal a new

   MP#3  It SHOULD be possible to ensure that MPLS-TP LSPs will not be
         moved to another component link as a result of a load-
         rebalancing operation for multipath.  If the selected component
         link no longer meets requirements, another component link may
         be selected; however, a change in path SHOULD NOT occur solely
         for load balancing.

Villamizar                    Informational                     [Page 7]

RFC 7190               MPLS and MPLS-TP Multipath             March 2014

   MP#4  Where a Resource Reservation Protocol - Traffic Engineering
         (RSVP-TE) control plane is used, it MUST be possible for an
         ingress LSR that is setting up an MPLS-TP or an MPLS LSP to
         determine at path selection time whether a link or Forwarding
         Adjacency (FA; see [RFC4206]) within the topology can support
         the MPLS-TP requirements of the LSP.

   The reason for requirement MP#1 may not be obvious.  An MPLS-TP LSP
   may be aggregated along with other client LSPs by a midpoint LSR into
   a very large MPLS server-layer LSP, as would be the case in a core-
   node-to-core-node MPLS LSP between major cities.  In this case, the
   ingress of the MPLS LSP, being a midpoint LSR for a set of client
   LSPs, has no signaling mechanism that can be used to determine
   whether one of its specific client LSPs is using MPLS or MPLS-TP.
   Multipath load splitting can be avoided for MPLS-TP LSPs if at the
   MPLS server-layer LSP ingress LSR an Entropy Label Indicator (ELI)
   and Entropy Label (EL) are added to the label stack by the midpoint
   LSR for the client MPLS-TP LSP, at the ingress of the MPLS LSP
   [RFC6790].  For those client LSPs that are MPLS-TP LSPs, a single
   per-LSP EL value must be chosen.  For those client LSPs that are MPLS
   LSPs, per-packet entropy below the top label must, for practical
   reasons, be used to determine the entropy label value.  The resulting
   label stack contains the server MPLS LSP label, ELI, EL and the
   client LSP label.  Requirement MP#1 simply states that there must be
   a means to make this decision.

   There is currently no signaling mechanism defined to support
   requirement MP#1, though that does not preclude a new extension being
   defined later.  In the absence of a signaling extension, MPLS-TP can
   be identified through some form of configuration, such as
   configuration that provides an MPLS-TP-compatible server layer to all
   LSPs arriving on a specific interface or originating from a specific
   set of ingress LSRs.

   Alternatively, the need for requirement MP#1 can be eliminated if
   every MPLS-TP LSP created by an MPLS-TP ingress makes use of an
   Entropy Label Indicator (ELI) and Entropy Label (EL) below the MPLS-
   TP label [RFC6790].  This would require that all MPLS-TP LSRs in a
   deployment support Entropy Label, which may render it impractical in
   many deployments.

   Some hardware that exists today can support requirement MP#2.
   Signaling in the absence of MPLS Entropy Labels can make use of link
   bundling with the path pinned to a specific component for MPLS-TP
   LSPs and link bundling using the all-ones component for MPLS LSPs.
   This prevents MPLS-TP LSPs from being carried within MPLS LSPs but
   does allow the coexistence of MPLS-TP and very large MPLS LSPs.

Villamizar                    Informational                     [Page 8]

RFC 7190               MPLS and MPLS-TP Multipath             March 2014

   When Entropy Label Indicators (ELIs) and Entropy Labels (ELs) are not
   applied by MPLS-TP ingresses, MPLS-TP LSPs can be carried as client
   LSPs within an MPLS server LSP if the ingress of the MPLS server-
   layer LSP pushes an Entropy Label Indicator (ELI) and Entropy Label
   (EL) below the server-layer LSP label(s) in the label stack, just
   above the MPLS-TP LSP label entry [RFC6790].  The value of EL can be
   randomly selected at the client MPLS-TP LSP setup time, and the same
   EL value can be used for all packets of that MPLS-TP LSP.  This
   allows MPLS-TP LSPs to be carried as client LSPs within MPLS LSPs and
   satisfies MPLS-TP forwarding requirements but requires that MPLS LSRs
   be able to identify MPLS-TP LSPs (requirement MP#1).

   MPLS-TP traffic can be protected from degraded performance due to an
   imperfect load split if the MPLS-TP traffic is given queuing
   priority.  For example, using (1) strict priority and policing,
   shaping at ingress, or per-LSP shaping locally, or (2) per-LSP
   weighted queuing locally.  This can be accomplished using the Traffic
   Class (TC) field and Diffserv treatment of traffic [RFC5462]
   [RFC2475].  In the event of congestion due to load imbalance, only
   non-prioritized traffic will suffer as long as there is a low
   percentage of prioritized traffic.

   If MPLS-TP LSPs are carried within MPLS LSPs and ELI and EL are used,
   requirement MP#3 is satisfied (1) for uncongested links where load
   balancing is not required, or (2) for MPLS-TP LSPs using Traffic
   Class (TC) and Diffserv, where the load rebalancing implementation
   rebalances only the less preferred traffic.  Load rebalance is
   generally needed only when congestion occurs; therefore, restricting
   MPLS-TP to be carried over MPLS LSPs that are known to traverse only
   links that are expected to be uncongested can satisfy requirement

   An MPLS-TP LSP can be pinned to a Link Bundle component link if the
   behavior of requirement MP#2 is preferred.  An MPLS-TP LSP can be
   assigned to a Link Bundle but not pinned if the behavior of
   requirement MP#3 is preferred.  In both of these cases, the MPLS-TP
   LSP must be the top-level LSP, except as noted above.

   If MPLS-TP LSPs can be moved among component links, then the Link
   Bundle all-ones component link can be used or server-layer MPLS LSPs
   can be used with no restrictions on the server-layer MPLS use of
   multipath, except that Entropy Labels must be supported along the
   entire path.  An Entropy Label must be used to ensure that all of the
   MPLS-TP payload and OAM traffic are carried on the same component,
   except during very infrequent transitions due to load balancing.
   Since the Entropy Label Indicator and Entropy Label are always placed
   above the Generic Associated Channel Label (GAL) in the stack, the

Villamizar                    Informational                     [Page 9]

RFC 7190               MPLS and MPLS-TP Multipath             March 2014

   presence of a GAL will not affect the selection of a component link
   as long as the LSR does not hash on the label stack entries below the
   Entropy Label.

   An MPLS-TP LSP may not traverse multipath links on the path where
   MPLS-TP forwarding requirements cannot be met.  Such links include
   any using pre-[RFC6790] Ethernet Link Aggregation, pre-[RFC6790] Link
   Bundling using the all-ones component link, or any other form of
   multipath that does not support termination of the entropy search at
   the EL as called for in [RFC6790].  An MPLS-TP LSP MUST NOT traverse
   a server-layer MPLS LSP that traverses any form of multipath that
   does not support termination of the entropy search at the EL.  For
   this to occur, the MPLS-TP ingress LSR MUST be aware of these links.
   This is the reason for requirement MP#4.

   Requirement MP#4 can be supported using administrative attributes.
   Administrative attributes are defined in [RFC3209].  Some
   configuration is required to support this.

   In MPLS Link Bundling the requirement for bidirectional co-routing
   can be interpreted as meaning that the same set of LSRs must be
   traversed or can be interpreted to mean that the same set of
   component links must be traversed [RFC4201] [RFC3473].  Following the
   procedures of Section 3 of RFC 3473 where Link Bundling is used only
   ensures that the same set of LSRs are traversed and that acceptable
   labels are created in each direction.

   When an MPLS-TP LSP is set up over a MPLS LSP, if the MPLS-TP LSP is
   a bidirectional LSP, then providers who want to only set these MPLS-
   TP LSPs over bidirectional co-routed MPLS LSPs can make use of
   administrative attributes [RFC3209] to ensure that this occurs.  If
   MPLS-TP LSPs are carried by unidirectional MPLS LSPs, the MPLS-TP OAM
   will be unaffected, as only the MPLS LSP endpoints will appear as
   MPLS-TP OAM Maintenance Entity Group Intermediate Points (MIPs).

   Two methods of adding an Entropy Label are described above.  The
   MPLS-TP ingress must have a means to determine which links can
   support MPLS-TP in selecting a path (MP#4).  Administrative
   attributes can satisfy that requirement.  If the MPLS-TP LSR is
   capable of adding ELI/EL to the label stack, this method is
   preferred.  However, equipment furthest from a provider's network
   core is the least likely to support RFC 6790 in the near term.  For
   portions of the topology where an MPLS-TP is carried within a server-
   layer MPLS LSP, the ingress of the server-layer MPLS LSP can add ELI/
   EL using a fixed EL value per client LSP, except those known not to
   require MPLS-TP treatment.  There are numerous ways to determine
   which client LSPs are MPLS-TP LSPs and which are not.  While this

Villamizar                    Informational                    [Page 10]

RFC 7190               MPLS and MPLS-TP Multipath             March 2014

   determination is out of scope and will vary among deployments,
   configuration or the presence of specific attribute affinities in
   RSVP-TE signaling are among the likely means to do so.

4.  MPLS-TP as a Server Layer for MPLS

   Carrying MPLS LSPs that are larger than a component link over an
   MPLS-TP server layer requires that the large MPLS client-layer LSP be
   accommodated by multiple MPLS-TP server-layer LSPs.  MPLS multipath
   can be used in the client-layer MPLS.

   Creating multiple MPLS-TP server-layer LSPs places a greater Incoming
   Label Map (ILM) scaling burden on the LSR.  High-bandwidth MPLS cores
   with a smaller amount of nodes have the greatest tendency to require
   LSPs in excess of component links; therefore, the reduction in the
   number of nodes offsets the impact of increasing the number of
   server-layer LSPs in parallel.  Today, only in cases where deployed
   LSR ILMs are small would this be an issue.

   The most significant disadvantage of MPLS-TP as a server layer for
   MPLS is that the use of MPLS-TP server-layer LSPs reduces the
   efficiency of carrying the MPLS client layer.  The service that
   provides by far the largest offered load in provider networks is the
   Internet, for which the LSP capacity reservations are predictions of
   expected load.  Many of these MPLS LSPs may be smaller than component
   link capacity.  Using MPLS-TP as a server layer results in bin-
   packing problems for these smaller LSPs.  For those LSPs that are
   larger than component link capacity, the LSP capacities need not be
   (and often are not) integer multiples of convenient capacity
   increments such as 10 Gbit/s.  Using MPLS-TP as an underlying server
   layer greatly reduces the ability of the client-layer MPLS LSPs to
   share capacity.  For example, when one MPLS LSP is underutilizing its
   predicted capacity, the fixed allocation of MPLS-TP to component
   links may not allow another LSP to exceed its predicted capacity.
   Using MPLS-TP as a server layer may result in less efficient use of
   resources and may result in a less cost-effective network.

   No additional requirements beyond MPLS-TP as it is now currently
   defined are required to support MPLS-TP as a server layer for MPLS.
   It is therefore viable but has some undesirable characteristics
   discussed above.

5.  Summary

   MPLS equipment deployed in the core currently supports multipath.
   For large service providers, core LSR must support some form of
   multipath to be deployable.  Deployed MPLS access and edge equipment
   is often oblivious to the use of multipath in the core.  It is

Villamizar                    Informational                    [Page 11]

RFC 7190               MPLS and MPLS-TP Multipath             March 2014

   expected that at least first-generation MPLS-TP equipment will be
   oblivious to the use of multipath in the core.  This first-generation
   MPLS-TP equipment is deployable in a core using multipath, with no
   adverse impact to RSVP-TE signaling, if:

   1.  the edge equipment can support administrative attributes (RFC

   2.  the core equipment can support ELI/EL, and

   3.  the core equipment can put a per-LSP fixed EL value on any LSP
       that indicates a particular attribute affinity or can identify a
       client MPLS-TP LSP through some other means.

   There are no issues carrying MPLS over MPLS-TP, except when the MPLS
   LSP is too big to be carried by a single MPLS-TP LSP.  Most MPLS core
   equipment and some edge equipment can configure an MPLS Link Bundle
   [RFC4201] over multiple component links where the component links are
   themselves MPLS LSP.  This existing capability can be used to carry
   large MPLS LSPs and overcome the limited capacity of any single
   server-layer MPLS-TP LSP.

   MPLS OAM and MPLS-TP OAM are unaffected in the following cases
   proposed in this document:

   1.  Where MPLS is carried over a single MPLS-TP, all traffic flows on
       one link, MPLS OAM is unaffected and need not use multipath
       support in LSP Ping [RFC4379].

   2.  Where MPLS-TP is carried over MPLS, all traffic for that MPLS-TP
       LSP is carried over one link thanks to the fixed EL value.  In
       this case, MPLS-TP OAM is unaffected.

   3.  Where MPLS LSPs are carried over MPLS LSPs (an existing case) or
       over multiple MPLS-TP LSPs, the multipath support in LSP Ping is
       used and LSP Ping operation is unaffected [RFC4379] [RFC6425].

6.  Acknowledgements

   Carlos Pignataro, Dave Allan, and Mach Chen provided valuable
   comments and suggestions.  Carlos suggested that MPLS-TP requirements
   in RFC 5960 be explicitly referenced or quoted.  An email
   conversation with Dave led to the inclusion of references and quotes
   from RFCs 6371 and 6374.  Mach made suggestions to improve the
   clarity of the document.

Villamizar                    Informational                    [Page 12]

RFC 7190               MPLS and MPLS-TP Multipath             March 2014

7.  Security Considerations

   This document specifies use of existing MPLS and MPLS-TP mechanisms
   to support MPLS and MPLS-TP as client and server layers for each
   other.  This use of existing mechanisms supports coexistence of MPLS/
   GMPLS (without MPLS-TP) when used over a packet network, MPLS-TP, and
   multipath.  The combination of MPLS, MPLS-TP, and multipath does not
   introduce any new security threats.  The security considerations for
   MPLS/GMPLS and for MPLS-TP are documented in [RFC5920] and [RFC6941].

8.  References

8.1.  Normative References

   [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119, March 1997.

   [RFC5654]  Niven-Jenkins, B., Brungard, D., Betts, M., Sprecher, N.,
              and S. Ueno, "Requirements of an MPLS Transport Profile",
              RFC 5654, September 2009.

   [RFC5960]  Frost, D., Bryant, S., and M. Bocci, "MPLS Transport
              Profile Data Plane Architecture", RFC 5960, August 2010.

   [RFC6371]  Busi, I. and D. Allan, "Operations, Administration, and
              Maintenance Framework for MPLS-Based Transport Networks",
              RFC 6371, September 2011.

   [RFC6374]  Frost, D. and S. Bryant, "Packet Loss and Delay
              Measurement for MPLS Networks", RFC 6374, September 2011.

   [RFC6790]  Kompella, K., Drake, J., Amante, S., Henderickx, W., and
              L. Yong, "The Use of Entropy Labels in MPLS Forwarding",
              RFC 6790, November 2012.

8.2.  Informative References

              Villamizar, C., McDysan, D., Ning, S., Malis, A., and L.
              Yong, "Requirements for Advanced Multipath in MPLS
              Networks", Work in Progress, February 2014.

              IEEE, "IEEE Standard for Local and Metropolitan Area
              Networks - Link Aggregation", IEEE Std 802.1AX-2008, 2006,

Villamizar                    Informational                    [Page 13]

RFC 7190               MPLS and MPLS-TP Multipath             March 2014

              ITU-T, "Unified functional architecture of transport
              networks", ITU-T G.800, 2007, <http://www.itu.int/rec/

   [RFC2475]  Blake, S., Black, D., Carlson, M., Davies, E., Wang, Z.,
              and W. Weiss, "An Architecture for Differentiated
              Services", RFC 2475, December 1998.

   [RFC3209]  Awduche, D., Berger, L., Gan, D., Li, T., Srinivasan, V.,
              and G. Swallow, "RSVP-TE: Extensions to RSVP for LSP
              Tunnels", RFC 3209, December 2001.

   [RFC3473]  Berger, L., "Generalized Multi-Protocol Label Switching
              (GMPLS) Signaling Resource ReserVation Protocol-Traffic
              Engineering (RSVP-TE) Extensions", RFC 3473, January 2003.

   [RFC4201]  Kompella, K., Rekhter, Y., and L. Berger, "Link Bundling
              in MPLS Traffic Engineering (TE)", RFC 4201, October 2005.

   [RFC4206]  Kompella, K. and Y. Rekhter, "Label Switched Paths (LSP)
              Hierarchy with Generalized Multi-Protocol Label Switching
              (GMPLS) Traffic Engineering (TE)", RFC 4206, October 2005.

   [RFC4379]  Kompella, K. and G. Swallow, "Detecting Multi-Protocol
              Label Switched (MPLS) Data Plane Failures", RFC 4379,
              February 2006.

   [RFC5286]  Atlas, A. and A. Zinin, "Basic Specification for IP Fast
              Reroute: Loop-Free Alternates", RFC 5286, September 2008.

   [RFC5462]  Andersson, L. and R. Asati, "Multiprotocol Label Switching
              (MPLS) Label Stack Entry: "EXP" Field Renamed to "Traffic
              Class" Field", RFC 5462, February 2009.

   [RFC5714]  Shand, M. and S. Bryant, "IP Fast Reroute Framework", RFC
              5714, January 2010.

   [RFC5920]  Fang, L., "Security Framework for MPLS and GMPLS
              Networks", RFC 5920, July 2010.

   [RFC6425]  Saxena, S., Swallow, G., Ali, Z., Farrel, A., Yasukawa,
              S., and T. Nadeau, "Detecting Data-Plane Failures in
              Point-to-Multipoint MPLS - Extensions to LSP Ping", RFC
              6425, November 2011.

Villamizar                    Informational                    [Page 14]

RFC 7190               MPLS and MPLS-TP Multipath             March 2014

   [RFC6941]  Fang, L., Niven-Jenkins, B., Mansfield, S., and R.
              Graveman, "MPLS Transport Profile (MPLS-TP) Security
              Framework", RFC 6941, April 2013.

Author's Address

   Curtis Villamizar
   Outer Cape Cod Network Consulting

   EMail: curtis@occnc.com

Villamizar                    Informational                    [Page 15]