SIP Record-Route fix: Addressing Record-Route Issues in the Session Initiation Protocol (SIP)

Network Working Group                                         T. Froment
Request for Comments: 5658                                   Tech-invite
Category: Standards Track                                       C. Lebel
                                                           B. Bonnaerens
                                                          Alcatel-Lucent
                                                            October 2009


                   Addressing Record-Route Issues in
                 the Session Initiation Protocol (SIP)

Abstract

   A typical function of a Session Initiation Protocol (SIP) Proxy is to
   insert a Record-Route header into initial, dialog-creating requests
   in order to make subsequent, in-dialog requests pass through it.
   This header contains a SIP Uniform Resource Identifier (URI) or SIPS
   (secure SIP) URI indicating where and how the subsequent requests
   should be sent to reach the proxy.  These SIP or SIPS URIs can
   contain IPv4 or IPv6 addresses and URI parameters that could
   influence the routing such as the transport parameter (for example,
   transport=tcp), or a compression indication like "comp=sigcomp".
   When a proxy has to change some of those parameters between its
   incoming and outgoing interfaces (multi-homed proxies, transport
   protocol switching, or IPv4 to IPv6 scenarios, etc.), the question
   arises on what should be put in Record-Route header(s).  It is not
   possible to make one header have the characteristics of both
   interfaces at the same time.  This document aims to clarify these
   scenarios and fix bugs already identified on this topic; it formally
   recommends the use of the double Record-Route technique as an
   alternative to the current RFC 3261 text, which describes only a
   Record-Route rewriting solution.

Status of This Memo

   This document specifies an Internet standards track protocol for the
   Internet community, and requests discussion and suggestions for
   improvements.  Please refer to the current edition of the "Internet
   Official Protocol Standards" (STD 1) for the standardization state
   and status of this protocol.  Distribution of this memo is unlimited.

Copyright Notice

   Copyright (c) 2009 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 BSD License.

Table of Contents

   1. Introduction ....................................................3
   2. Terminology .....................................................3
   3. Problem Statement ...............................................4
      3.1. Background: Multi-Homed Proxies ............................4
      3.2. Identified Problems ........................................5
   4. Record-Route Rewriting ..........................................6
   5. Double Record-Routing ...........................................6
   6. Usage of Transport Protocol Parameter ..........................10
      6.1. UA Implementation Problems and Recommendations ............10
      6.2. Proxy Implementation Problems and Recommendations .........14
   7. Conclusion .....................................................15
   8. Security Considerations ........................................16
   9. Acknowledgments ................................................16
   10. References ....................................................17
      10.1. Normative References .....................................17
      10.2. Informative References ...................................17

1.  Introduction

   Over the years, it has been noticed in interoperability events like
   SIPit, that many implementations had interoperability problems due to
   various Record-Routing issues or misinterpretations of [RFC3261]; in
   particular, when a change occurs between the incoming and outgoing
   sides of a proxy: transport protocol switching, "multi-homed" proxies
   (including IPv4 to IPv6 interface changes), etc.  Multiple documents
   have addressed the question, each of them generally providing an
   adequate recommendation for its specific use case, but none of them
   gives a general solution or provides a coherent set of
   clarifications:

      - [RFC3486], Section 6, describes the double Record-Routing as an
        alternative to the Record-Route rewriting in responses.  This
        document is limited in scope to the "comp=sigcomp" parameter
        when doing compression with Signalling Compression (SIGCOMP).

      - [RFC3608], Section 6.2, recommends the usage of double Record-
        Routing instead of the rewriting solution described in [RFC3261]
        for "Dual-homed" proxies.  Those are defined as "proxies
        connected to two (or more) different networks such that requests
        are received on one interface and proxied out through another
        network interface".

      - Section 3.1.1 of [V6Tran] mandates double Record-Routing for
        multi-homed proxies doing IPv4/IPv6 transitions, when the proxy
        inserts IP addresses in the Record-Route header URI.

   The observed interoperability problems can be explained by the fact
   that, despite these multiple documents, the RFC 3261 description has
   not been changed, and many implementations don't support extensions
   like Service-Route ([RFC3608]) or SIGCOMP ([RFC3486]).

   This document also aims to clarify an identified bug referenced in
   [BUG664].  In particular, it takes into account the [BUG664]
   recommendation, which says that "the language that describes this,
   needs to clearly capture that this applies to all types of different
   interface on each side issues, including IPv4 on one side and IPv6 on
   the other".

2.  Terminology

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
   document are to be interpreted as described in RFC 2119 [RFC2119].

3.  Problem Statement

3.1.  Background: Multi-Homed Proxies

   A multi-homed proxy is a proxy connected, like a router, to two or
   more different networks, with an interface into each network, such
   that traffic comes "in" one network and goes "out" a different one.
   A simple example is shown here:

                   +-----+
                   | UA1 |
                   +--+--+
                      | .66
       192.0.2.64/26  |
      ----------------+---+-...
                           |
                           | .65
                         +-+-+
                         | P |
                         +-+-+
                           | .129
                           |          192.0.2.128/26
                     ...---+------+------------------
                                  |
                                  | .130
                               +--+--+
                               | UA2 |
                               +--+--+

                   Figure 1: Multi-Homed Proxy Illustration

   UA1 has one interface with IP address 192.0.2.66.

   The Proxy P has two interfaces and two addresses:

      --192.0.2.65

      --192.0.2.129

   UA2 has one interface with address, 192.0.2.130.  There is
   potentially no IP-level route between UA1 and UA2 (pinging or
   traceroute does not work between these two hosts).  They live in
   entirely different subnetworks.  But they can still exchange SIP
   messages, because P is a SIP Proxy.  This works in SIP because P can
   apply Record-Routing.

   In most cases, there is still some IP connectivity between UA1 and
   UA2, but SIP proxy has to manage the SIP traffic between the two
   different "sides", e.g., with two different IP addresses, or one side
   using SIGCOMP and the other side not, etc.

3.2.  Identified Problems

   Handling of the Record-Route header in SIP Proxies is specified by
   following sections of [RFC3261]:

   On the request processing side, [RFC3261], item 4 of Section 16.6
   states that:

      The URI placed in the Record-Route header field value MUST be a
      SIP or SIPS URI. [...]  The URI SHOULD NOT contain the transport
      parameter unless the proxy has knowledge (such as in a private
      network) that the next downstream element that will be in the path
      of subsequent requests supports that transport.

   Following this statement, it is not clear how to decide when the
   proxy should insert the transport protocol parameter in the Record-
   Route URI.

   On the response processing side, [RFC3261] recommends in step 8 of
   Section 16.7 that:

      If the selected response contains a Record-Route header field
      value originally provided by this proxy, the proxy MAY choose to
      rewrite the value before forwarding the response.  This allows the
      proxy to provide different URIs for itself to the next upstream
      and downstream elements.  A proxy may choose to use this mechanism
      for any reason.  For instance, it is useful for multi-homed hosts.

      If the proxy received the request over Transport Layer Security
      (TLS), and sent it out over a non-TLS connection, the proxy MUST
      rewrite the URI in the Record-Route header field to be a SIPS URI.

   Note that [RFC5630] has weakened the SIP/SIPS URI rewriting
   requirement in the Record-Route header by removing this second
   paragraph.

   Indeed, [RFC3261] suggests rewriting the Record-Route header in
   responses.

   This list highlights the utility of rewriting and double Record-
   Routing techniques that apply for any multi-homed proxy use case:
   whenever the proxy changes its IP address, the transport protocol, or
   the URI scheme between incoming and outgoing interfaces.  Rewriting
   and double Record-Routing are described, compared, and discussed in
   Sections 4 and 5; the specific question of whether or not to insert
   the transport parameter in the Record-Route URI is then discussed in
   Section 6.

4.  Record-Route Rewriting

   As frequently outlined in IETF mailing list discussions, Record-Route
   rewriting in responses is not the optimal way of handling multi-
   homed and transport protocol switching situations.  Additionally, the
   consequence of doing rewriting is that the route set seen by the
   caller is different from the route set seen by the callee, and this
   has at least two negative implications:

   1) The callee cannot sign the route set, because it gets edited by
      the proxy in the response.  Consequently, end-to-end protection of
      the route set cannot be supported by the protocol.  This means the
      Internet's principles of openness and end-to-end connectivity are
      broken.

   2) A proxy must implement special "multi-homed" logic.  During the
      request forwarding phase, it performs an output interface
      calculation and writes information resolving to the output
      interface into the URI of the Record-Route header.  When handling
      responses, the proxy must inspect the Record-Route header(s), look
      for an input interface, and selectively edit them to reference the
      correct output interface.  Since this lookup has to be done for
      all responses forwarded by the proxy, this technique implies a CPU
      drag.

   Therefore, this document recommends using the double Record-Route
   approach to avoid rewriting the Record-Route.  This recommendation
   applies to all uses of Record-Route rewriting by proxies, including
   transport protocol switching and multi-homed proxies.

5.  Double Record-Routing

   The serious drawbacks of the rewriting technique explain why the
   double Record-Routing solution has consequently been recommended in
   SIP extensions like [RFC3486] or [RFC3608].

   This technique consists of inserting before any existing Record-Route
   header, a Record-Route header with the URI reflecting to the input
   interface, including schemes and/or URI parameters, and secondly, a
   Record-Route header with the URI reflecting to the output interface.
   When processing the response, no modification of the recorded route
   is required.  This is completely backward compatible with [RFC3261].
   Generally speaking, the time complexity will be less in double
   Record-Routing since, on processing the response, the proxy does not
   have to do any rewrites (and thus, no searching).  Moreover, the
   handling of in-dialog requests and responses requires no special
   handling anymore.

   When double Record-Routing, the proxy will have to handle the
   subsequent in-dialog request(s) as a spiral, and consequently devote
   resources to maintain transactions required to handle the spiral.
   What is considered to be a spiraling request is explained in Section
   6 of [RFC3261].  In order to avoid a spiral, the proxy can be smart
   and scan an extra Route header ahead to determine whether the request
   will spiral through it.  If it does, it can optimize the second
   spiral through itself.  Even though this is an implementation
   decision, it is much more efficient to avoid spiraling.  So, this
   means, in Section 16.4, "Route Information Preprocessing" [RFC3261],
   implementors can choose that a proxy MAY remove two Route headers
   instead of one when using the double Record-Routing.

   The following example is an extension of the example given in
   [V6Tran].  It illustrates a basic call flow using double Record-
   Routing in a multi-homed IPv4 to IPv6 proxy, and annotates the dialog
   state on each User Agent (UA).  In this example, proxy P1,
   responsible for the domain biloxy.example.com, receives a request
   from an IPv4-only upstream client.  It proxies this request to an
   IPv6-only downstream server.  Proxy P1 is running on a dual-stack
   host; on the IPv4 interface, it has an address of 192.0.2.254.  On
   the IPv6 interface, it is configured with an address of 2001:db8::1.
   Some mandatory SIP headers have been omitted to ease readability.

       UA1              Proxy "P1"               UA2
      (IPv4)            (IPv4/IPv6)             (IPv6)
        |                    |                    |
        |   F1 INVITE        |                    |
        |------------------->|      F2 INVITE     |
        |                    |------------------->|
        |    100 Trying      |                    |
        |<-------------------|                    |
        |                    |    F3 200 OK       |
        |    F4 200 OK       |<-------------------|
        |<-------------------|                    |
        |                    |                    |
        |       F5 ACK       |                    |
        |------------------->|       F6 ACK       |
        |                    |------------------->|
        |                    |                    |
        |                    |        F7 BYE      |
        |       F8 BYE       |<-------------------|
        |<-------------------|                    |

            Figure 2: IPv4 to IPv6 Multi-Homed Proxy Illustration


   F1 INVITE UA1 -> P1 (192.0.2.254:5060)

   INVITE sip:bob@biloxi.example.com SIP/2.0
   Route: <sip:192.0.2.254:5060;lr>
   From: Alice <sip:alice@atlanta.example.com>;tag=1234
   To: Bob <sip:bob@biloxi.example.com>
   Contact: <sip:alice@192.0.2.1>

           F2 INVITE P1 (2001:db8::1) -> UA2

           INVITE sip:bob@biloxi.example.com SIP/2.0
           Record-Route: <sip:[2001:db8::1];lr>
           Record-Route: <sip:192.0.2.254:5060;lr>
           From: Alice <sip:alice@atlanta.example.com>;tag=1234
           To: Bob <sip:bob@biloxi.example.com>
           Contact: <sip:alice@192.0.2.1>

                   Dialog State at UA2:
                   Local URI     = sip:bob@biloxi.example.com
                   Remote URI    = sip:alice@atlanta.example.com
                   Remote target = sip:alice@192.0.2.1
                   Route Set     = sip:[2001:db8::1];lr
                                   sip:192.0.2.254:5060:lr

                   F3 200 OK UA2 -> P1 (2001:db8::1)

                   SIP/2.0 200 OK
                   Record-Route: <sip:[2001:db8::1];lr>
                   Record-Route: <sip:192.0.2.254:5060;lr>
                   From: Alice <sip:alice@atlanta.example.com>;tag=1234
                   To: Bob <sip:bob@biloxi.example.com>;tag=4567
                   Contact: <sip:bob@[2001:db8::33]>

           F4 200 OK P1 -> UA1

           SIP/2.0 200 OK
           Record-Route: <sip:[2001:db8::1];lr>
           Record-Route: <sip:192.0.2.254:5060;lr>
           From: Alice <sip:alice@atlanta.example.com>;tag=1234
           To: Bob <sip:bob@biloxi.example.com>;tag=4567
           Contact: <sip:bob@[2001:db8::33]>

   Dialog State at UA1:
   Local URI     = sip:alice@atlanta.example.com
   Remote URI    = sip:bob@biloxi.example.com
   Remote target = sip:bob@[2001:db8::33]
   Route Set     = sip:192.0.2.254:5060:lr
                   sip:[2001:db8::1];lr

   F5 ACK UA1 -> P1 (192.0.2.254:5060)

   ACK sip:bob@[2001:db8::33] SIP/2.0
   Route: <sip:192.0.2.254:5060:lr>
   Route: <sip:[2001:db8::1];lr>
   From: Alice <sip:alice@atlanta.example.com>;tag=1234
   To: Bob <sip:bob@biloxi.example.com>;tag=4567

           F6 ACK P1 (2001:db8::1) -> UA2

           ACK sip:bob@[2001:db8::33] SIP/2.0
           From: Alice <sip:alice@atlanta.example.com>;tag=1234
           To: Bob <sip:bob@biloxi.example.com>;tag=4567
           (both Route headers have been removed by the proxy)

                   F7 BYE UA2 -> P1 (2001:db8::1)

                   BYE sip:alice@192.0.2.1 SIP/2.0
                   Route: <sip:[2001:db8::1];lr>
                   Route: <sip:192.0.2.254:5060:lr>
                   From: Bob <sip:bob@biloxi.example.com>;tag=4567
                   To: Alice <sip:alice@atlanta.example.com>;tag=1234

           F8 BYE P1 (192.0.2.254:5060) -> UA1

           BYE sip:alice@192.0.2.1 SIP/2.0
           From: Bob <sip:bob@biloxi.example.com>;tag=4567
           To: Alice <sip:alice@atlanta.example.com>;tag=1234

   Figure 3: Multi-Homed IPv4 to IPv6 Double Record-Routing Illustration

6.  Usage of Transport Protocol Parameter

   This section describes a set of problems that is related to the usage
   of transport protocol URI parameters in the Record-Route header.  In
   some circumstances, interoperability problems occur because it is not
   clear whether or not to include the transport parameter on the URI of
   the Record-Route header.  This was identified as a frequent problem
   in past SIPit events.

   [RFC3261], step 8 of Section 16.7 says:

      The URI SHOULD NOT contain the transport parameter unless the
      proxy has knowledge (such as in a private network) that the next
      downstream element that will be in the path of subsequent requests
      supports that transport.

   The preceding seems to confuse implementors, resulting in proxies
   that insert a single Record-Route without a transport URI parameter,
   resulting in the problems described in this section.

6.1.  UA Implementation Problems and Recommendations

   Consider the following scenario: a SIP proxy, doing TCP to UDP
   transport protocol switching.

   In this example, proxy P1, responsible for the domain
   biloxy.example.com, receives a request from Alice UA1, which uses
   TCP.  It proxies this request to Bob UA2, which registered with a
   Contact specifying UDP as transport protocol.  Thus, P1 receives an
   initial request from Alice over TCP and forwards it to Bob over UDP.
   For subsequent requests, it is expected that TCP could continue to be
   used between Alice and P1, and UDP between P1 and Bob, but this can
   not happen if a numeric IP address is used and no transport parameter
   is set on Record-Route URI.  This happens because of procedures
   described in [RFC3263].  Some mandatory SIP headers have been omitted
   to ease readability.

      Alice UA1 ===== TCP ===== Proxy P1 ===== UDP ===== Bob UA2
         |                        |                         |
         |       F1 INVITE        |                         |
         |----------------------->|         F2 INVITE       |
         |                        |------------------------>|
         |      100 Trying        |                         |
         |<-----------------------|                         |
         |                        |        F3 200 OK        |
         |       F4 200 OK        |<------------------------|
         |<-----------------------|                         |
         |                        |                         |
         |        F5 ACK          |                         |
         |---(sent over UDP) X--->|            ACK          |
         |                        |------------------------>|
         |                        |                         |
         |                        |          F6 BYE         |
         |          BYE           |<------------------------|
         |<-----------------------|                         |

                   Figure 4: TCP to UDP Transport Protocol
                         Switching Issue Illustration

   F1 INVITE UA1 -> P1 (192.0.2.1/tcp)

   INVITE sip:bob@biloxi.example.com SIP/2.0
   Route: <sip:192.0.2.1;lr;transport=tcp>
   From: Alice <sip:alice@atlanta.example.com>;tag=1234
   To: Bob <sip:bob@biloxi.example.com>
   Contact: <sip:alice@ua1.atlanta.example.com;transport=tcp>

        F2 INVITE P1 -> UA2 (ua2.biloxi.example.com/udp)

        INVITE sip:bob@ua2.biloxi.example.com;transport=udp SIP/2.0
        Record-Route: <sip:192.0.2.1;lr> (NO transport param)
        From: Alice <sip:alice@atlanta.example.com>;tag=1234
        To: Bob <sip:bob@biloxi.example.com>
        Contact: <sip:alice@ua1.atlanta.example.com;transport=tcp>

        Dialog State at UA2:
        Local URI     = sip:bob@biloxi.example.com
        Remote URI    = sip:alice@atlanta.example.com
        Remote target = sip:alice@ua1.atlanta.example.com;transport=tcp
        Route Set     = sip:192.0.2.1;lr

             F3 200 OK UA2 -> P1 (192.0.2.1/udp)

             SIP/2.0 200 OK
             Record-Route: <sip:192.0.2.1;lr>
             From: Alice <sip:alice@atlanta.example.com>;tag=1234
             To: Bob <sip:bob@biloxi.example.com>;tag=4567
             Contact: <sip:bob@ua2.biloxi.example.com>

        F4 200 OK P1 -> UA1 (ua1.atlanta.example.com/tcp)

        SIP/2.0 200 OK
        Record-Route: <sip:192.0.2.1;lr>
        From: Alice <sip:alice@atlanta.example.com>;tag=1234
        To: Bob <sip:bob@biloxi.example.com>;tag=4567
        Contact: <sip:bob@ua2.biloxi.example.com>

   Dialog State at UA1:
   Local URI     = sip:alice@atlanta.example.com
   Remote URI    = sip:bob@biloxi.example.com
   Remote target = sip:bob@ua2.biloxi.example.com
   Route Set     = sip:192.0.2.1;lr


   F5 ACK UA1 -> P1 (192.0.2.1/udp)

   ACK sip:bob@ua2.biloxi.example.com SIP/2.0
   Route: <sip:192.0.2.1;lr>
   From: Alice <sip:alice@atlanta.example.com>;tag=1234
   To: Bob <sip:bob@biloxi.example.com>;tag=4567

             F6 BYE UA2 -> P1 (192.0.2.1/udp)

             BYE sip:alice@ua1.atlanta.example.com;transport=tcp SIP/2.0
             Route: <sip:192.0.2.1;lr>
             From: Bob <sip:bob@biloxi.example.com>;tag=4567
             To: Alice <sip:alice@atlanta.example.com>;tag=1234

                 Figure 5: TCP to UDP Transport Protocol
                        Switching Issue Description

   Since the proxy P1 does not insert any transport parameter in the
   Record-Route URI, subsequent in-dialog requests of UA1, like the ACK
   sent in F5, will be sent according to the behavior specified in
   Section 12.2 (requests within a Dialog) of [RFC3261].  That means
   that the routeset is used, and then, applying [RFC3263], the Route
   "sip:192.0.2.1" will resolve to a UDP transport by default (since no
   transport parameter is present here), and no Naming Authority Pointer
   (NAPTR) request will be performed since this is a numeric IP address.

   In general, the interoperability problems arise when UA1 is trying to
   send the ACK: it is not ready to change its transport protocol for a
   mid-dialog request and just fails to do so, requiring the proxy
   implementor to insert the transport protocol in the Record-Route URI.

   What happens if the proxy had Record-Routed its logical name
   (biloxi.example.com)?  Since Bob is to be contacted over UDP,
   protocol switching will be avoided only if the resulting transport
   protocol of [RFC3263] procedures is UDP.  For any other resulting
   transport protocol, the transport protocol switching issue described
   above will occur.  Also, if one of the UAs sends an initial request
   using a different transport than the one retrieved from DNS, this
   scenario would be problematic.

   In practice, there are multiple situations where UA implementations
   don't use logical names and NAPTR records when sending an initial
   request to a proxy.  This happens, for instance, when:

   1) UAs offer the ability to "choose" the transport to be used for
      initial requests, even if they support [RFC3263].  This is a
      frequent UA functionality that is justified by the following use
      cases:

      - when it is not possible to change the DNS server configuration
        and the implementation doesn't support all the transport
        protocols that could be configured by default in DNS (e.g.,
        TLS).

      - when the end-user wants to choose his transport protocol for
        whatever reason, e.g., needing to force TCP, avoiding
        UDP/congestion, retransmitting, or fragmenting, etc.

   This ability to force the transport protocol in UAs for initial
   requests SHOULD be avoided: selecting the transport protocol in the
   configuration of an outbound proxy means that [RFC3263] procedure is
   bypassed for initial requests.  As a consequence, if the proxy
   Record-Routed with no transport parameter as is recommended in
   [RFC3261], the UA will be forced to use the [RFC3263]-preferred
   transport for subsequent requests anyway, which leads to the
   problematic scenario described in Figure 4.

   2) UAs decide to always keep the same transport for a given dialog.
      This choice is erratic, since if the proxy is not Record-Routing,
      the callee MAY receive the subsequent request through a transport
      that is not the one put in its Contact.  If a UA really wants to
      avoid transport protocol switching between the initial and
      subsequent request, it SHOULD rely on DNS records for that; thus,
      it SHOULD avoid configuring statically the outbound proxy with a
      numeric IP address.  A logical name, with no transport parameter,
      SHOULD be used instead.

   3) UAs don't support [RFC3263] at all, or don't have any DNS server
      available.  In that case, as illustrated previously, forcing UA1
      to switch from TCP to UDP between initial request and subsequent
      request(s) is clearly not the desired default behavior, and it
      typically leads to interoperability problems.  UA implementations
      SHOULD then be ready to change the transport protocol between
      initial and subsequent requests.  In theory, any UA or proxy using
      UDP must also be prepared to use TCP for requests that exceed the
      size limit of path MTU, as described in Section 18.1.1 of
      [RFC3261].

6.2.  Proxy Implementation Problems and Recommendations

   In order to prevent UA implementation problems, and to maintain a
   reasonable level of interoperability, the situation can be improved
   on the proxy side.  Thus, if the transport protocol changed between
   its incoming and outgoing sides, the proxy SHOULD use the double
   Record-Route technique and SHOULD add a transport parameter to each
   of the Record-Route URIs it inserts.  When TLS is used on the
   transport on either side of the proxy, the URI placed in the Record-
   Route header field MUST encode a next-hop that will be reached using
   TLS.  There are two ways for this to work.  The first way is for the
   URI placed in the Record-Route to be a SIPS URI.  The second is for
   the URI placed in the Record-Route to be constructed such that
   application of [RFC3263] resolution procedures to that URI results in
   TLS being selected.  Proxies compliant with this specification MUST
   NOT use a "transport=tls" parameter on the URI placed in the Record-
   Route because the "transport=tls" usage was deprecated by [RFC3261].
   Record-Route rewriting MAY also be used.  However, the recommendation
   to put a transport protocol parameter on Record-Route URI does not
   apply when the proxy has changed the transport protocol due to the
   size of UDP requests as per Section 18.1.1 of [RFC3261].  As an
   illustration of the previous example, it means one of the following
   processing will be performed:

   - Double Record-Routing: the proxy inserts two Record-Route headers
     into the SIP request.  The first one is set, in this example, to
     Record-Route: <sip:192.0.2.1;lr;transport=tcp>, the second one is
     set to Record-Route: <sip:192.0.2.1;lr> with no transport, or with
     transport=udp, which basically means the same thing.

   - Record-Route rewriting on responses: in the INVITE request sent in
     F2, the proxy puts the outgoing transport protocol in the transport
     parameter of Record-Route URI.  Doing so, UA2 will correctly send
     its BYE request in F6 using the same transport protocol as previous
     messages of the same dialog.  The proxy rewrites the Record-Route
     when processing the 200 OK response, changing the transport
     parameter "on the fly" to "transport=tcp", so that the Route set
     will appear to be <sip:192.0.2.1;lr;transport=tcp> for UA1 and
     <sip:192.0.2.1;lr;transport=udp> for UA2.

   It is a common practice in proxy implementations to support double
   Record-Route AND to insert the transport parameter in the Record-
   Route URI.  This practice is acceptable as long as all SIP elements
   that may be in the path of subsequent requests support that
   transport.  This restriction needs an explanation.  Let's imagine you
   have two proxies, P1 at "p1.biloxi.example.com" and P2 on the path of
   an initial request.  P1 is Record-Routing and changes the transport
   from UDP to Stream Control Transmission Protocol (SCTP) because the
   P2 URI resolves to SCTP applying [RFC3263].  Consequently, the proxy
   P1 inserts two Record-Route headers:

   Record-Route: <sip:p1.biloxi.example.com;transport=udp> and

   Record-Route: <sip:p1.biloxi.example.com;transport=sctp>.

   The problem arises if P2 is not Record-Routing, because the SIP
   element downstream of P2 will be asked to reach P1 using SCTP for any
   subsequent, in-dialog request from the callee, and this downstream
   SIP element may not support that transport.

   In order to handle this situation, this document recommends that a
   proxy SHOULD apply the double Record-Routing technique as soon as it
   changes the transport protocol between its incoming and outgoing
   sides.  If proxy P2 in the example above would follow this
   recommendation, it would perform double Record-Routing and the
   downstream element would not be forced to send requests over a
   transport it does not support.

   By extension, a proxy SHOULD also insert a Record-Route header for
   any multi-homed situation (as the ones described in this document:
   scheme changes, sigcomp, IPv4/IPv6, transport changes, etc.) that may
   impact the processing of proxies being on the path of subsequent
   requests.

7.  Conclusion

   As a conclusion of this document, it is to notice that:

   - Record-Route rewriting is presented as a technique that MAY be
     used, with the drawbacks outlined in Section 4.

   - Double Record-Routing is presented as the technique that SHOULD be
     used, and is documented in Section 5.

   - Record-Route header interoperability problems on transport protocol
     switching scenarios have been outlined and described in Section 6.
     This last section gives some recommendations to UA and proxy
     implementations to improve the situation.  Proxies SHOULD use
     double Record-Routing for any multi-homed situation that MAY impact
     the further processing, and they SHOULD put transport protocol
     parameters on Record-Route URIs in some circumstances.  UAs SHOULD
     NOT offer options to overwrite the transport for initial requests.
     Further, UAs SHOULD rely on DNS to express their desired transport
     and SHOULD avoid IP addresses with transport parameters in this
     case.  Finally, UAs SHOULD be ready to switch transports between
     the initial request and further in-dialog messages.

8.  Security Considerations

   The recommendations in this document describe a way to use the
   existing protocol specified in RFC 3261 rather than introducing any
   new protocol mechanism.  As such, they do not introduce any new
   security concerns, but additional consideration of already existing
   concerns is warranted.  In particular, when a message is transiting
   two interfaces, the double Record-Route technique will carry
   information about both interfaces to each of the involved endpoints
   (and any intermediaries between this proxy and those endpoints),
   where the rewriting technique would only expose information about the
   interface closest to each given endpoint.  If issues such as topology
   hiding or privacy (as described in [RFC3323]) are a concern, the URI
   values placed in the Record-Route for each interface should be
   carefully constructed to avoid exposing more information than was
   intended.

9.  Acknowledgments

   Thank you to Dean Willis, Vijay K. Gurbani, Joel Repiquet, Robert
   Sparks, Jonathan Rosenberg, Cullen Jennings, Juha Heinanen, Paul
   Kyzivat, Nils Ohlmeier, Tim Polk, Francois Audet, Adrian Farrel,
   Ralph Droms, Tom Batsele, Yannick Bourget, Keith Drage, and John
   Elwell for their reviews and comments.

10.  References

10.1.  Normative References

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

   [RFC3261]  Rosenberg, J., Schulzrinne, H., Camarillo, G., Johnston,
              A., Peterson, J., Sparks, R., Handley, M., and E.
              Schooler, "SIP: Session Initiation Protocol", RFC 3261,
              June 2002.

   [RFC3263]  Rosenberg, J. and H. Schulzrinne, "Session Initiation
              Protocol (SIP): Locating SIP Servers", RFC 3263, June
              2002.

   [RFC3323]  Peterson, J., "A Privacy Mechanism for the Session
              Initiation Protocol (SIP)", RFC 3323, November 2002.

   [RFC5630]  Audet, F., "The Use of the SIPS URI Scheme in the Session
              Initiation Protocol (SIP)", RFC 5630, October 2009.

10.2.  Informative References

   [BUG664]   Sparks, RS., "Bug 664: Double record routing,
              http://bugs.sipit.net/show_bug.cgi?id=664", October 2002.

   [RFC3486]  Camarillo, G., "Compressing the Session Initiation
              Protocol (SIP)", RFC 3486, February 2003.

   [RFC3608]  Willis, D. and B. Hoeneisen, "Session Initiation Protocol
              (SIP) Extension Header Field for Service Route Discovery
              During Registration", RFC 3608, October 2003.

   [V6Tran]   Camarillo, G., El Malki, K., and V. Gurbani, "IPv6
              Transition in the Session Initiation Protocol (SIP)", Work
              in Progress, August 2009.

Authors' Addresses

   Thomas Froment
   Tech-invite

   EMail: thomas.froment@tech-invite.com


   Christophe Lebel
   Alcatel-Lucent
   Lieu dit Le Mail
   Orvault  44708
   France

   EMail: christophe.lebel@alcatel-lucent.com


   Ben Bonnaerens
   Alcatel-Lucent
   Copernicuslaan 50
   Antwerpen  2018
   Belgium

   EMail: ben.bonnaerens@alcatel-lucent.com