RFC 8372






Internet Engineering Task Force (IETF)                         S. Bryant
Request for Comments: 8372                                        Huawei
Category: Informational                                     C. Pignataro
ISSN: 2070-1721                                                    Cisco
                                                                 M. Chen
                                                                   Z. Li
                                                                  Huawei
                                                               G. Mirsky
                                                               ZTE Corp.
                                                                May 2018


                MPLS Flow Identification Considerations

Abstract



   This document discusses aspects to consider when developing a
   solution for MPLS flow identification.  The key application that
   needs this solution is in-band performance monitoring of MPLS flows
   when MPLS is used to encapsulate user data packets.

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 candidates for any level of Internet
   Standard; see Section 2 of RFC 7841.

   Information about the current status of this document, any errata,
   and how to provide feedback on it may be obtained at
   https://www.rfc-editor.org/info/rfc8372.















Bryant, et al.                Informational                     [Page 1]

RFC 8372                MPLS Flow Identification                May 2018


Copyright Notice



   Copyright (c) 2018 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
   (https://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.  Loss Measurement Considerations . . . . . . . . . . . . . . .   3
   3.  Delay Measurement Considerations  . . . . . . . . . . . . . .   4
   4.  Units of Identification . . . . . . . . . . . . . . . . . . .   4
   5.  Types of LSP  . . . . . . . . . . . . . . . . . . . . . . . .   6
   6.  Network Scope . . . . . . . . . . . . . . . . . . . . . . . .   7
   7.  Backwards Compatibility . . . . . . . . . . . . . . . . . . .   7
   8.  Data Plane  . . . . . . . . . . . . . . . . . . . . . . . . .   7
   9.  Control Plane . . . . . . . . . . . . . . . . . . . . . . . .   9
   10. Privacy Considerations  . . . . . . . . . . . . . . . . . . .   9
   11. Security Considerations . . . . . . . . . . . . . . . . . . .   9
   12. IANA Considerations . . . . . . . . . . . . . . . . . . . . .   9
   13. Informative References  . . . . . . . . . . . . . . . . . . .  10
   Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . .  11
   Authors' Addresses  . . . . . . . . . . . . . . . . . . . . . . .  11

1.  Introduction



   This document discusses the aspects that need to be considered when
   developing a solution for MPLS flow identification.  The key
   application that needs this is in-band performance monitoring of MPLS
   flows when MPLS is used to encapsulate user data packets.

   There is a need to identify flows in MPLS networks for various
   applications such as determining packet loss and packet delay
   measurement.  A method of loss and delay measurement in MPLS networks
   was defined in [RFC6374].  When used to measure packet loss,
   [RFC6374] depends on the use of injected Operations, Administration,
   and Maintenance (OAM) packets to designate the beginning and the end
   of the packet group over which packet loss is being measured.  If the
   misordering of packets from one group relative to the following group



Bryant, et al.                Informational                     [Page 2]

RFC 8372                MPLS Flow Identification                May 2018


   or the misordering of any of the packets being counted relative to
   the Loss Measurement packet [RFC6374] occurs, then an error will
   occur in the packet loss measurement.

   In addition, [RFC6374] did not support different granularities of
   flow or address a number of multipoint cases in which two or more
   ingress Label Switching Routers (LSRs) could send packets to one or
   more destinations.

   Due to the very low loss rate in normal operation, improvements in
   link and transmission technologies have made it more difficult to
   assess packet loss using active performance measurement methods with
   synthetic traffic.  That, together with more demanding service-level
   requirements, means that network operators now need to be able to
   measure the loss of the actual user data traffic using passive
   performance measurement methods.  Any technique deployed needs to be
   transparent to the end user, and it needs to be assumed that they
   will not take any active part in the measurement process.  Indeed, it
   is important that any flow identification technique be invisible to
   them and that no remnant of the measurement process leaks into their
   network.

   Additionally, when there are multiple traffic sources, such as in
   multipoint-to-point and multipoint-to-multipoint network
   environments, there needs to be a method whereby the sink can
   distinguish between packets from the various sources; that is to say,
   a multipoint measurement model needs to be developed.

2.  Loss Measurement Considerations



   Modern networks, if not oversubscribed, generally drop relatively few
   packets; thus, packet loss measurement is highly sensitive to the
   common demarcation of the exact set of packets to be measured for
   loss.  Without some form of coloring or batch marking such as that
   proposed in [RFC8321], it may not be possible to achieve the required
   accuracy in the loss measurement of customer data traffic.  Thus,
   when accurate measurement of packet loss is required, it may be
   economically advantageous, or even be a technical requirement, to
   include some form of marking in the packets to assign each packet to
   a particular counter for loss measurement purposes.

   When this level of accuracy is required and the traffic between a
   source-destination pair is subject to Equal-Cost Multipath (ECMP), a
   demarcation mechanism is needed to group the packets into batches.
   Once a batch is correlated at both ingress and egress, the packet
   accounting mechanism is then able to operate on the batch of packets
   that can be accounted for at both the packet ingress and the packet




Bryant, et al.                Informational                     [Page 3]

RFC 8372                MPLS Flow Identification                May 2018


   egress.  Errors in the accounting are particularly acute in Label
   Switched Paths (LSPs) subjected to ECMP because the network transit
   time will be different for the various ECMP paths since:

   1.  the packets may traverse different sets of LSRs;

   2.  the packets may depart from different interfaces on different
       line cards on LSRs; and

   3.  the packets may arrive at different interfaces on different line
       cards on LSRs.



   A consideration with this solution on modifying the identity label
   (the MPLS label ordinarily used to identify the LSP, Virtual Private
   Network, Pseudowire, etc.) to indicate the batch is the impact that
   this has on the path chosen by the ECMP mechanism.  When the member
   of the ECMP path set is chosen by deep packet inspection, a change of
   batch represented by a change of identity label will have no impact
   on the ECMP path.  If the path member is chosen by reference to an
   entropy label [RFC6790], then changing the batch identifier will not
   result in a change to the chosen ECMP path.  ECMP is so pervasive in
   multipoint-to-(multi)point networks that some method of avoiding
   accounting errors introduced by ECMP needs to be supported.

3.  Delay Measurement Considerations

   Most of the existing delay measurement methods are active methods
   that depend on the extra injected test packet to evaluate the delay
   of a path.  With the active measurement method, the rate, numbers,
   and interval between the injected packets may affect the accuracy of
   the results.  Due to ECMP (or link aggregation techniques), injected
   test packets may traverse different links from the ones used by the
   data traffic.  Thus, measuring the delay of the real traffic is
   required.

   For combined loss and delay measurements, both the loss and the delay
   considerations apply.

4.  Units of Identification



   The most basic unit of identification is the identity of the node
   that processed the packet on its entry to the MPLS network.  However,
   the required unit of identification may vary depending on the use
   case for accounting, performance measurement, or other types of
   packet observations.  In particular, note that there may be a need to
   impose identity at several different layers of the MPLS label stack.





Bryant, et al.                Informational                     [Page 4]

RFC 8372                MPLS Flow Identification                May 2018


   This document considers the identification of the following traffic
   components:

   o  Per source LSR: everything from one source is aggregated

   o  Per group of LSPs chosen by an ingress LSR: an ingress LSP
      aggregates a group of LSPs (e.g., all the LSPs of a tunnel)

   o  Per LSP: the basic form

   o  Per flow [RFC6790] within an LSP: a fine-grained method

   Note that a fine-grained identity resolution is needed when there is
   a need to perform these operations on a flow not readily identified
   by some other element in the label stack.  Such a fine-grained
   resolution may be possible by deep packet inspection.  However, this
   may not always be possible, or it may be desired to minimize
   processing costs by doing this only on entry to the network.  Adding
   a suitable identifier to the packet for reference by other network
   elements minimizes the processing needed by other network elements.
   An example of such a fine-grained case might be traffic belonging to
   a certain service or from a specific source, particularly if matters
   related to service level agreement or application performance were
   being investigated.

   We can thus characterize the identification requirement in the
   following broad terms:

   o  There needs to be some way for an egress LSR to identify the
      ingress LSR with an appropriate degree of scope.  This concept is
      discussed further in Section 6.

   o  There needs to be a way to identify a specific LSP at the egress
      node.  This allows for the case of instrumenting multiple LSPs
      operating between the same pair of nodes.  In such cases, the
      identity of the ingress LSR is insufficient.

   o  In order to conserve resources such as labels, counters, and/or
      compute cycles, it may be desirable to identify an LSP group so
      that an operation can be performed on the group as an aggregate.

   o  There needs to be a way to identify a flow within an LSP.  This is
      necessary when investigating a specific flow that has been
      aggregated into an LSP.

   The unit of identification and the method of determining which
   packets constitute a flow will be specific to the application or use
   case and are out of scope of this document.



Bryant, et al.                Informational                     [Page 5]

RFC 8372                MPLS Flow Identification                May 2018


5.  Types of LSP



   We need to consider a number of types of LSP.  The two simplest types
   to monitor are point-to-point LSPs and point-to-multipoint LSPs.  The
   ingress LSR for a point-to-point LSP, such as those created using the
   Resource Reservation Protocol - Traffic Engineering (RSVP-TE)
   [RFC5420] signaling protocol or those that conform to the MPLS
   Transport Profile (MPLS-TP) [RFC5654], may be identified by
   inspection of the top label in the stack because, at any provider-
   edge (PE) or provider (P) router on the path, the top label is unique
   to the ingress-egress pair at every hop at a given layer in the LSP
   hierarchy.  Provided that Penultimate Hop Popping (PHP) is disabled,
   the identity of the ingress LSR of a point-to-point LSP is available
   at the egress LSR; thus, determining the identity of the ingress LSR
   must be regarded as a solved problem.  Note, however, that the
   identity of a flow cannot to be determined without further
   information being carried in the packet or gleaned from some aspect
   of the packet payload.

   In the case of a point-to-multipoint LSP, and in the absence of PHP,
   the identity of the ingress LSR may also be inferred from the top
   label.  However, it may not possible to adequately identify the flow
   from the top label alone; thus, further information may need to be
   carried in the packet or gleaned from some aspect of the packet
   payload.  In designing any solution, it is desirable that a common
   flow identification solution be used for both point-to-point and
   point-to-multipoint LSP types.  Similarly, it is desirable that a
   common method of LSP group identification be used.  In the above
   cases, a context label [RFC5331] needs to be used to provide the
   required identity information.  This is a widely supported MPLS
   feature.

   A more interesting case is the case of a multipoint-to-point LSP.  In
   this case, the same label is normally used by multiple ingress or
   upstream LSRs; hence, source identification is not possible by
   inspection of the top label by the egress LSRs.  It is therefore
   necessary for a packet to be able to explicitly convey any of the
   identity types described in Section 4.

   Similarly, in the case of a multipoint-to-multipoint LSP, the same
   label is normally used by multiple ingress or upstream LSRs; hence,
   source identification is not possible by inspection of the top label
   by egress LSRs.  The various identity types described in Section 4
   are again needed.  Note, however, that the scope of the identity may
   be constrained to be unique within the set of multipoint-to-
   multipoint LSPs terminating on any common node.





Bryant, et al.                Informational                     [Page 6]

RFC 8372                MPLS Flow Identification                May 2018


6.  Network Scope



   The scope of identification can be constrained to the set of flows
   that are uniquely identifiable at an ingress LSR or some aggregation
   thereof.  There is no need for an ingress LSR to seek assistance from
   outside the MPLS protocol domain.

   In any solution that constrains itself to carrying the required
   identity in the MPLS label stack rather than in some different
   associated data structure, constraints on the choice of label and
   label stack size imply that the scope of identity resides within that
   MPLS domain.  For similar reasons, the identity scope of a component
   of an LSP is constrained to the scope of that LSP.

7.  Backwards Compatibility



   In any network, it is unlikely that all LSRs will have the same
   capability to support the methods of identification discussed in this
   document.  It is therefore an important constraint on any flow
   identity solution that it is backwards compatible with deployed MPLS
   equipment to the extent that deploying the new feature will not
   disable anything that currently works on the legacy equipment.

   This is particularly the case when the deployment is incremental or
   the feature is not required for all LSRs or all LSPs.  Thus, the flow
   identification design must support the coexistence of LSRs that can
   identify the traffic components described in Section 4 and those that
   cannot.  In addition, the identification of the traffic components
   described in Section 4 must be an optional feature that is disabled
   by default.  As a design simplification, a solution may require that
   all egress LSRs of a point-to-multipoint or a multipoint-to-
   multipoint LSP support the identification type in use so that a
   single packet can be correctly processed by all egress devices.  The
   corollary of this last point is that either all egress LSRs are
   enabled to support the required identity type or none of them are.

8.  Data Plane



   There is a huge installed base of MPLS equipment; typically, this
   type of equipment remains in service for an extended period of time,
   and in many cases, hardware constraints mean that it is not possible
   to upgrade its data-plane functionality.  Changes to the MPLS data
   plane are therefore expensive to implement, add complexity to the
   network, and may significantly impact the deployability of a solution
   that requires such changes.  For these reasons, MPLS users have set a
   very high bar to changes to the MPLS data plane, and only a very
   small number have been adopted.  Hence, it is important that the
   method of identification must minimize changes to the MPLS data



Bryant, et al.                Informational                     [Page 7]

RFC 8372                MPLS Flow Identification                May 2018


   plane.  Ideally, method(s) of identification that require no changes
   to the MPLS data plane should be given preferential consideration.
   If a method of identification that makes a change to the data plane
   is chosen, it will need to have a significant advantage over any
   method that makes no change, and the advantage of the approach will
   need to be carefully evaluated and documented.  If a change to the
   MPLS data plane proves necessary, it should be (a) as small a change
   as possible and (b) a general-purpose method, so as to maximize its
   use for future applications.  It is imperative that, as far as can be
   foreseen, any necessary change made to the MPLS data plane does not
   impose any foreseeable future limitation on the MPLS data plane.

   Stack size is an issue with many MPLS implementations both as a
   result of hardware limitations and due to the impact on networks and
   applications in which a large number of small payloads need to be
   transported.  In particular, one MPLS payload may be carried inside
   another.  For example, one LSP may be carried over another LSP, or a
   Pseudowire (PW) or similar multiplexing construct may be carried over
   an LSP, and identification may be required at both layers.  Of
   particular concern is the implementation of low-cost edge LSRs that,
   for cost reasons, have a significant limit on the number of Label
   Stack Entries (LSEs) that they can impose or dispose.  Therefore, any
   method of identity must not consume an excessive number of unique
   labels and must not result in an excessive increase in the size of
   the label stack.

   The design of the MPLS data plane provides two types of special-
   purpose labels: the original 16 reserved labels and the much larger
   set of special-purpose labels defined in [RFC7274].  The original
   reserved labels need one LSE, and the newer special-purpose labels
   [RFC7274] need two LSEs.  Given the tiny number of original reserved
   labels, it is core to the MPLS design philosophy that this scarce
   resource is only used when it is absolutely necessary.  Using a
   special-purpose label to encode flow identity requires two label
   stack entries, one for the reserved label and one for the flow
   identity.  Use of extended special-purpose labels [RFC7274] requires
   a total of three label stack entries to encode the flow identity.
   The larger set of [RFC7274] labels requires two label stack entries
   for the special-purpose label itself; hence, a total of three label
   stack entries is needed to encode the flow identity.

   The use of special-purpose labels [RFC7274] as part of a method to
   encode the identity information therefore has a number of undesirable
   implications for the data plane.  Thus, while a solution may use
   special-purpose labels, methods that do not require special-purpose
   labels need to be carefully considered.





Bryant, et al.                Informational                     [Page 8]

RFC 8372                MPLS Flow Identification                May 2018


9.  Control Plane



   Any flow identity design should both seek to minimize the complexity
   of the control plane and minimize the amount of label coordination
   needed amongst LSRs.

10.  Privacy Considerations



   The inclusion of originating and/or flow information in a packet
   provides more identity information and hence potentially degrades the
   privacy of the communication.

   Recent IETF concerns on pervasive monitoring [RFC7258] have resulted
   in a preference for a solution that does not degrade the privacy of
   user traffic below that of an MPLS network not implementing the flow
   identification feature.  The choice of using MPLS technology for this
   OAM solution has a privacy advantage, as the choice of the label
   identifying a flow is limited to the scope of the MPLS domain and
   does not have any dependency on the identification of the user data.
   This minimizes the observability of the flow characteristics.

11.  Security Considerations



   Any flow identification solution must not degrade the security of the
   MPLS network below that of an equivalent network not deploying the
   specified identity solution.  In order to preserve present
   assumptions about MPLS privacy properties, propagation of
   identification information outside the MPLS network imposing it must
   be disabled by default.  Any solution should provide for the
   restriction of the identity information to those components of the
   network that need to know it.  It is thus desirable to limit the
   knowledge of the identify of an endpoint to only those LSRs that need
   to participate in traffic flow.  The choice of using MPLS technology
   for this OAM solution, with MPLS encapsulation of user traffic,
   provides for a key advantage over other data-plane solutions, as it
   provides for a controlled access and trusted domain within a service
   provider's network.

   For a more comprehensive discussion of MPLS security and attack
   mitigation techniques, please see "Security Framework for MPLS and
   GMPLS Networks" [RFC5920].

12.  IANA Considerations



   This document has no IANA considerations.






Bryant, et al.                Informational                     [Page 9]

RFC 8372                MPLS Flow Identification                May 2018


13.  Informative References



   [RFC5331]  Aggarwal, R., Rekhter, Y., and E. Rosen, "MPLS Upstream
              Label Assignment and Context-Specific Label Space",
              RFC 5331, DOI 10.17487/RFC5331, August 2008,
              <https://www.rfc-editor.org/info/rfc5331>.

   [RFC5420]  Farrel, A., Ed., Papadimitriou, D., Vasseur, JP., and A.
              Ayyangarps, "Encoding of Attributes for MPLS LSP
              Establishment Using Resource Reservation Protocol Traffic
              Engineering (RSVP-TE)", RFC 5420, DOI 10.17487/RFC5420,
              February 2009, <https://www.rfc-editor.org/info/rfc5420>.

   [RFC5654]  Niven-Jenkins, B., Ed., Brungard, D., Ed., Betts, M., Ed.,
              Sprecher, N., and S. Ueno, "Requirements of an MPLS
              Transport Profile", RFC 5654, DOI 10.17487/RFC5654,
              September 2009, <https://www.rfc-editor.org/info/rfc5654>.

   [RFC5920]  Fang, L., Ed., "Security Framework for MPLS and GMPLS
              Networks", RFC 5920, DOI 10.17487/RFC5920, July 2010,
              <https://www.rfc-editor.org/info/rfc5920>.

   [RFC6374]  Frost, D. and S. Bryant, "Packet Loss and Delay
              Measurement for MPLS Networks", RFC 6374,
              DOI 10.17487/RFC6374, September 2011,
              <https://www.rfc-editor.org/info/rfc6374>.

   [RFC6790]  Kompella, K., Drake, J., Amante, S., Henderickx, W., and
              L. Yong, "The Use of Entropy Labels in MPLS Forwarding",
              RFC 6790, DOI 10.17487/RFC6790, November 2012,
              <https://www.rfc-editor.org/info/rfc6790>.

   [RFC7258]  Farrell, S. and H. Tschofenig, "Pervasive Monitoring Is an
              Attack", BCP 188, RFC 7258, DOI 10.17487/RFC7258, May
              2014, <https://www.rfc-editor.org/info/rfc7258>.

   [RFC7274]  Kompella, K., Andersson, L., and A. Farrel, "Allocating
              and Retiring Special-Purpose MPLS Labels", RFC 7274,
              DOI 10.17487/RFC7274, June 2014,
              <https://www.rfc-editor.org/info/rfc7274>.

   [RFC8321]  Fioccola, G., Ed., Capello, A., Cociglio, M., Castaldelli,
              L., Chen, M., Zheng, L., Mirsky, G., and T. Mizrahi,
              "Alternate-Marking Method for Passive and Hybrid
              Performance Monitoring", RFC 8321, DOI 10.17487/RFC8321,
              January 2018, <https://www.rfc-editor.org/info/rfc8321>.





Bryant, et al.                Informational                    [Page 10]

RFC 8372                MPLS Flow Identification                May 2018


Acknowledgments

   The authors thank Nobo Akiya, Nagendra Kumar Nainar, George Swallow,
   and Deborah Brungard for their comments.

Authors' Addresses



   Stewart Bryant
   Huawei

   Email: stewart.bryant@gmail.com


   Carlos Pignataro
   Cisco Systems, Inc.

   Email: cpignata@cisco.com


   Mach(Guoyi) Chen
   Huawei

   Email: mach.chen@huawei.com


   Zhenbin Li
   Huawei

   Email: lizhenbin@huawei.com


   Gregory Mirsky
   ZTE Corp.

   Email: gregimirsky@gmail.com
















Bryant, et al.                Informational                    [Page 11]