Network Working Group                                      D. Voyer, ED.
Internet-Draft                                                D. Bernier
Intended status: Informational                               Bell Canada
Expires: December 26, 2018                                      J. Leddy
                                                             C. Filsfils
                                                           D. Dukes, ED.
                                                     Cisco Systems, Inc.
                                                              S. Previdi
                                                  Individual Contributor
                                                           S. Matsushima
                                                            May 25, 2018

    Insertion of IPv6 Segment Routing Headers in a Controlled Domain


   The network operator and vendor community has clearly indicated that
   IPv6 header insertion is useful and required.  This is notably the
   case when the entire journey of the packet remains in its source
   domain.  In such a context, it does not matter where the extension
   header is inserted.  The source domain may decide to place the IPv6
   extension header insertion where it suits its best: at the source of
   the packet or at any midpoint within the source domain.

Voyer, et al.          Expires December 26, 2018                [Page 1]

Internet-Draft             IPv6 SRH Insertion          November 20, 2017

Requirements Language

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   document are to be interpreted as described in RFC 2119 [RFC2119].

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).  Note that other groups may also distribute
   working documents as Internet-Drafts.  The list of current Internet-
   Drafts is at

   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."

Copyright Notice

   Copyright (c) 2017 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
   ( 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
   2.  Source Domain and Packet Journey  . . . . . . . . . . . . . .   3
     2.1.  Example: 6PE  . . . . . . . . . . . . . . . . . . . . . .   4
   3.  Transit through a Source Domain . . . . . . . . . . . . . . .   5
     3.1.  Example: SDWAN  . . . . . . . . . . . . . . . . . . . . .   5
   4.  SRH insertion in Source Domain  . . . . . . . . . . . . . . .   6
   5.  Security Considerations . . . . . . . . . . . . . . . . . . .   7
   6.  IANA Considerations . . . . . . . . . . . . . . . . . . . . .   7
   7.  Manageability Considerations  . . . . . . . . . . . . . . . .   7
   8.  Contributors  . . . . . . . . . . . . . . . . . . . . . . . .   7

Voyer, et al.          Expires December 26, 2018                [Page 2]

Internet-Draft             IPv6 SRH Insertion          November 20, 2017

   9.  Acknowledgements  . . . . . . . . . . . . . . . . . . . . . .   7
   10. References  . . . . . . . . . . . . . . . . . . . . . . . . .   7
     10.1.  Normative References . . . . . . . . . . . . . . . . . .   8
     10.2.  Informative References . . . . . . . . . . . . . . . . .   8
   Authors' Addresses  . . . . . . . . . . . . . . . . . . . . . . .   9

1.  Introduction

   We define the concept of "domain" as the set of nodes under the same
   administration.  For example, a network operator infrastructure
   including routers and links grouped into BGP autonomous systems (ASs)
   and routing domains (running OSFP or IS-IS).

   We define "source domain" as the domain of the source of the packet.

2.  Source Domain and Packet Journey

       (--- Source Domain ---)
       (                     )
       ( 1-----2-----3-----9 )
       (       |     |       )
       (       4-----5       )

                          Figure 1: Source Domain

   In the previous diagram:

   o  All the nodes 1 to 9 are in the same domain D.

   o  Node 1 originates a packet P1 destined to 9 (SA=1, DA=9).

   o  Domain D runs a link-state routing protocols which implements the
      Fast Reroute (FRR) service through the Topology Independent Loop
      Free Alternates (TI-LFA,

   o  All link metrics are set to 10.

   o  Node 2's TI-LFA pre-computed backup path for the destination 9 is
      the Segment Routing Policy <5, 9> via outgoing interface (OIF) to
      node 4 according to [I-D.filsfils-spring-segment-routing-policy],
      [I-D.filsfils-spring-srv6-network-programming] and

   Within the 50 milliseconds of link 2-3 failure detection, node 2
   reroutes the traffic destined to 9 by inserting the pre-computed
   segment routing header (SRH) with SID list <5, 9> and forwards the

Voyer, et al.          Expires December 26, 2018                [Page 3]

Internet-Draft             IPv6 SRH Insertion          November 20, 2017

   packet to node 4.  Node 4 forwards based on DA=5 to neighbor 5.  Node
   5 updates the DA to 9 and removes the SRH.  Node 9 receives the
   packet with (SA=1, DA=9).

   This FRR service is clearly beneficial for the operator of domain D:
   without this FRR operation, depending on the scale of the domain and
   the quality of the routing convergence implementation, traffic could
   be dropped for hundreds to thousands of milliseconds waiting for the
   routing plane to converge.

   This FRR service is largely deployed with MPLS.

   It is important to note that the operators industry is strongly
   requiring the same TI-LFA FRR service without the need to deploy or
   maintain the MPLS layer.

   Obviously, this FRR service increases the size of the packet during
   its journey within domain D.  This is well-known to operators.  Well-
   known mitigation techniques have been deployed for more than 15 years
   for the MPLS-based FRR service and the numerous VPN services.  The
   same exact technique can be used which consists of deploying an
   infrastructure capable of an MTU value higher in the core than at the
   ingress edge.

2.1.  Example: 6PE

   An illustrative example of packet size increasing while in transit is
   given by the 6PE use case ([RFC4798]) and consists of the following:

   o  An ingress router encapsulates incoming IPv6 packets into a MPLS
      label stack.

   o  The ingress router forwards the encapsulated packet across the
      MPLS core infrastructure of the operator.

   o  While transiting in the core, the size of the label stack (hence
      the size of the packet) may increase when additional labels are
      pushed on top of the packet due to services such as traffic
      engineering (TE) and/or fast reroute (FRR).

   o  Once the packet reaches the egress router, the label stack is
      removed and the packet is sent out of the operator network as an
      IPv6 packet.

   Using the example topology in Figure 1:

Voyer, et al.          Expires December 26, 2018                [Page 4]

Internet-Draft             IPv6 SRH Insertion          November 20, 2017

   o  Router 1 receives an incoming IPv6 packet, classifies it and does
      push a label corresponding to the egress router this packet must
      reach (router 9).  At this stage the packet size has already
      increased by the size of one label (32 bits).

   o  In addition, router 1 is configured with a policy consisting of a
      traffic engineered path (2, 4, 5, 9) represented by an additional
      label.  At this stage the packet size has increased by the size of
      two labels).

   o  While in transit in the traffic engineered path, a link may fail
      and fast reroute (FRR) may have been provisioned so that the
      packet is immediately re-routed (e.g., by router 4) using an
      additional label.  At this stage the packet size has increased by
      the size of three labels.

   o  It has also to be noted that the packet may be part of a virtual
      private network (VPN) service in which case it will have an
      additional label corresponding to the VPN the packet belongs to.
      In this case the packet size has increased by the size of four

   o  Once the packet reaches its egress router (router 9) the label
      stack is consumed, the packet shape and size is restored to its
      original state and the packet is forwarded externally, as a plain
      IPv6 packet.

3.  Transit through a Source Domain

        (--- Source Domain ---)
        (                     )
    A---( 1-----2-----3-----9 )---B
        (       |     |       )
        (       4-----5       )

                 Figure 2: Transit Through a Source Domain

   We now consider a packet P2 sent from A to B where A and B are
   external nodes to the domain D.

   We assume that when node 1 receives the packet, for service
   transparency reason (the packet that exits the domain MUST be the
   same as when it entered the domain), node 1 encapsulates the packet
   P2 in an outer IPv6 header with SA=1 and DA=9.  We call the resulting
   packet P3.

Voyer, et al.          Expires December 26, 2018                [Page 5]

Internet-Draft             IPv6 SRH Insertion          November 20, 2017

   Note that node 1 in Domain D is the source of a new IPv6 packet (P3)
   that encapsulates P2, therefore we refer to this scenario as "Transit
   Through a Source domain".

   From the viewpoint of domain D, packet P3 is the same as packet P1 of
   the previous use-case.  Indeed, domain D only considers the outer
   header and from that viewpoint, the outer header is the same: (SA=1,
   DA=9).  Furthermore, as for packet P1, the entire journey of packet
   P3 is contained within domain D.

   Node 2 may thus rightfully insert an SRH on packet P3 to implement a
   sub-50 milliseconds FRR operation upon the loss of the link 2-to-3
   and node 5 can remove this SRH.

   The transparency of the service is guaranteed: the insertion and
   removal of the SRH on packet P3 has no impact on packet P2.  P2 at
   the exit of the domain D is the same as at the entrance of the domain

   Customers of the transit service offered by domain D do demand FRR
   services.  The 50 millisecond FRR operation provides a much better
   service availability than 100's to 1000's of milliseconds of loss for
   each routing transition.

3.1.  Example: SDWAN

   An illustration example of transit through a source domain is given
   by [I-D.dukes-sr-for-sdwan], and consist of the following:

   o  An ingress SDWAN Edge node encrypts and encapsulates incoming IPv4
      or IPv6 packets in an IPv6 header;

   o  The ingress node inserts an SRH with a binding SID;

   o  The use of the binding SID is an explicit agreement within the
      source domain to apply a given SLA to the traffic containing the
      binding SID;

   o  As the packet traverses the source domain, the binding SID and the
      SRH containing it MAY be removed (if the binding SID was a PSP

   o  A new SRH MAY be inserted within the source domain, corresponding
      to the path represented by the binding SID, and this SRH MAY be
      removed before the packet exits the source domain (again depending
      on the last SID in the SID list and whether or not it was a PSP

Voyer, et al.          Expires December 26, 2018                [Page 6]

Internet-Draft             IPv6 SRH Insertion          November 20, 2017

   Using the example topology in Figure 1:

   o  Router 1 receives a packet destined to B, which it encrypts and
      encapsulates in an IPv6 header destined to router 9;

   o  Router 1 inserts an SRH containing a binding SID S2.1 (a binding
      SID at 2,) to receive a path with a negotiated SLA, and sends the
      packet toward S1 with an outer header and SRH of:

   o  Router 2 receives the packet and perform PSP for S2.1

   o  Router 2 inserts a new SRH with segment list <S3,S4>

   o  Once the packet reaches its egress router (router 9) the SRH has
      been removed (if S4 was a PSP SID), the outer IPv6 header is
      removed, the inner packet decrypted and the original packet is
      forwarded externally toward B.

4.  SRH insertion in Source Domain

      This document reminds that for a packet whose journey is
      completely contained within its source domain, it does not matter
      where the IPv6 extension header insertion is done.

      It could be at the source or at the first-hop router or at any
      node of the source domain.

      The source domain owns the packet and decides the location, which
      suits it best.

      During the packet journey in the source domain, any node in the
      source domain may insert an SRH and the egress node MUST remove
      both the outer IPv6 header and inserted SRH.

      In conclusion, in a context of a controlled/trusted domain, the
      insertion of SRHs in packets that are sourced within the domain is
      useful, required and also harmless.  Hence, a context-less ban of
      IPv6 extension header insertion does not make sense.

Voyer, et al.          Expires December 26, 2018                [Page 7]

Internet-Draft             IPv6 SRH Insertion          November 20, 2017

5.  Security Considerations

      This document proposes to insert an SRH to a transit packet if
      such packet is originated and destined within a controlled/trusted
      domain. Typically, when an operator encapsulates an externally
      received packet and sends this packet in a tunneled mode from
      ingress to egress, insertion of SRH is safe and confined within
      the operator's infrastructure.  In such conditions, the security
      of the original packet is not compromised by header insertion and
      the packet leaves the operator's domain without the any header
      that the operator may have added.

      A controlled/trusted domain can operate SRv6-based services for
      internal traffic while preventing any external traffic from
      accessing these internal SRv6-based services.  Several mechanisms
      exists and are currently used by operators.  Here follows a non-
      exhaustive example list:

   o  Access-lists (ACL) on the each externally facing interface in
      order to drop any incoming traffic with SA or DA belonging to the
      internal SID space.

   o  ACL to prevent access to local SIDs from outside the operator's

   o  Support Unicast-RPF on source address on external interface.

6.  IANA Considerations

   This document doesn't introduce any IANA request.

7.  Manageability Considerations


8.  Contributors


9.  Acknowledgements


10.  References

Voyer, et al.          Expires December 26, 2018                [Page 8]

Internet-Draft             IPv6 SRH Insertion          November 20, 2017

10.1.  Normative References

   [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119,
              DOI 10.17487/RFC2119, March 1997,

10.2.  Informative References

              Bashandy, A., Filsfils, C., Decraene, B., Litkowski, S.,
              and P. Francois, "Abstract", draft-bashandy-rtgwg-segment-
              routing-ti-lfa-00 (work in progress), February 2017.

              Filsfils, C., Sivabalan, S., Yoyer, D., Nanduri, M., Lin,
              S.,, b., Horneffer, M., Clad, F.,
              Steinberg, D., Decraene, B., and S. Litkowski, "Segment
              Routing Policy for Traffic Engineering", draft-filsfils-
              spring-segment-routing-policy-00 (work in progress),
              February 2017.

              Filsfils, C., Leddy, J.,, d.,
    , d., Steinberg, D., Raszuk, R.,
              Matsushima, S., Lebrun, D., Decraene, B., Peirens, B.,
              Salsano, S., Naik, G., Elmalky, H., Jonnalagadda, P.,
              Sharif, M., Ayyangar, A., Mynam, S., Bashandy, A., Raza,
              K., Dukes, D., Clad, F., and P. Camarillo, "SRv6 Network
              Programming", draft-filsfils-spring-srv6-network-
              programming-00 (work in progress), March 2017.

              Previdi, S., Filsfils, C., Raza, K., Leddy, J., Field, B.,
    , d.,, d.,
              Matsushima, S., Leung, I., Linkova, J., Aries, E., Kosugi,
              T., Vyncke, E., Lebrun, D., Steinberg, D., and R. Raszuk,
              "IPv6 Segment Routing Header (SRH)", draft-ietf-6man-
              segment-routing-header-06 (work in progress), March 2017.

          Dukes, D., Filsfils, C., Dawra, G., Camirillo, P., Clad, F.,
          Salsano, S., draft-dukes-sr-for-sdwan-00 (work in progress)
          October 2017.

   [RFC4798]  De Clercq, J., Ooms, D., Prevost, S., and F. Le Faucheur,
              "Connecting IPv6 Islands over IPv4 MPLS Using IPv6
              Provider Edge Routers (6PE)", RFC 4798,

Voyer, et al.          Expires December 26, 2018                [Page 9]

Internet-Draft             IPv6 SRH Insertion          November 20, 2017

              DOI 10.17487/RFC4798, February 2007,

Voyer, et al.          Expires December 26, 2018               [Page 10]

Internet-Draft             IPv6 SRH Insertion          November 20, 2017

Authors' Addresses

   Daniel Voyer (editor)
   Bell Canada


   John Leddy
   4100 East Dry Creek Road
   Centennial, CO  80122


   Clarence Filsfils
   Cisco Systems, Inc.


   Stefano Previdi
   Individual Contributor


   Darren Dukes (editor)
   Cisco Systems


   Satoru Matsushima


Voyer, et al.          Expires December 26, 2018               [Page 11]