TOC 
Network Working GroupC. Jennings, Ed.
Internet-DraftCisco Systems
Updates:3261, 3327 (if approved)R. Mahy, Ed.
Expires: April 25, 2007Plantronics
 October 22, 2006

Managing Client Initiated Connections in the Session Initiation Protocol (SIP)

draft-ietf-sip-outbound-05

Status of this Memo

By submitting this Internet-Draft, each author represents that any applicable patent or other IPR claims of which he or she is aware have been or will be disclosed, and any of which he or she becomes aware will be disclosed, in accordance with Section 6 of BCP 79.

Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet-Drafts.

Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as “work in progress.”

The list of current Internet-Drafts can be accessed at http://www.ietf.org/ietf/1id-abstracts.txt.

The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html.

This Internet-Draft will expire on April 25, 2007.

Copyright Notice

Copyright © The Internet Society (2006).

Abstract

The Session Initiation Protocol (SIP) allows proxy servers to initiate TCP connections and send asynchronous UDP datagrams to User Agents in order to deliver requests. However, many practical considerations, such as the existence of firewalls and Network Address Translators (NATs), prevent servers from connecting to User Agents in this way. This specification defines behaviors for User Agents, registrars and proxy servers that allow requests to be delivered on existing connections established by the User Agent. It also defines keep alive behaviors needed to keep NAT bindings open and specifies the usage of multiple connections.



Table of Contents

1.  Introduction
2.  Conventions and Terminology
    2.1  Definitions
3.  Overview
    3.1  Summary of Mechanism
    3.2  Single Registrar and UA
    3.3  Multiple Connections from a User Agent
    3.4  Edge Proxies
    3.5  Keepalive Technique
4.  User Agent Mechanisms
    4.1  Instance ID Creation
    4.2  Initial Registrations
        4.2.1  Registration by Other Instances
    4.3  Sending Requests
    4.4  Detecting Flow Failure
        4.4.1  Keepalive with TCP KEEPALIVE
        4.4.2  Keepalive with STUN
        4.4.3  Flow Recovery
5.  Edge Proxy Mechanisms
    5.1  Processing Register Requests
    5.2  Generating Flow Tokens
    5.3  Forwarding Requests
    5.4  Edge Proxy Keepalive Handling
6.  Registrar Mechanisms: Processing REGISTER Requests
7.  Authoritative Proxy Mechansims: Forwarding Requests
8.  STUN Keepalive Processing
    8.1  Explicit Probes
    8.2  Use with Sigcomp
9.  Example Message Flow
10.  Grammar
11.  Definition of 430 Flow Failed response code
12.  IANA Considerations
    12.1  Contact Header Field
    12.2  SIP/SIPS URI Parameters
    12.3  SIP Option Tag
    12.4  Response Code
    12.5  Media Feature Tag
13.  Security Considerations
14.  Requirements
15.  Changes
    15.1  Changes from 04 Version
    15.2  Changes from 03 Version
    15.3  Changes from 02 Version
    15.4  Changes from 01 Version
    15.5  Changes from 00 Version
16.  Acknowledgments
A.  Default Flow Registration Backoff Times
17.  References
    17.1  Normative References
    17.2  Informative References
§  Authors' Addresses
§  Intellectual Property and Copyright Statements




 TOC 

1. Introduction

There are many environments for SIP (Rosenberg, J., Schulzrinne, H., Camarillo, G., Johnston, A., Peterson, J., Sparks, R., Handley, M., and E. Schooler, “SIP: Session Initiation Protocol,” June 2002.)[RFC3261] deployments in which the User Agent (UA) can form a connection to a Registrar or Proxy but in which connections in the reverse direction to the UA are not possible. This can happen for several reasons. Connections to the UA can be blocked by a firewall device between the UA and the proxy or registrar, which will only allow new connections in the direction of the UA to the Proxy. Similarly there may be a NAT, which are only capable of allowing new connections from the private address side to the public side. This specification allows SIP registration when the UA is behind such a firewall or NAT.

Most IP phones and personal computers get their network configurations dynamically via a protocol such as DHCP (Dynamic Host Configuration Protocol). These systems typically do not have a useful name in the Domain Name System (DNS), and they definitely do not have a long-term, stable DNS name that is appropriate for use in the subjectAltName of a certificate, as required by [RFC3261] (Rosenberg, J., Schulzrinne, H., Camarillo, G., Johnston, A., Peterson, J., Sparks, R., Handley, M., and E. Schooler, “SIP: Session Initiation Protocol,” June 2002.). However, these systems can still act as a TLS client and form connections to a proxy or registrar which authenticates with a server certificate. The server can authenticate the UA using a shared secret in a digest challenge over that TLS connection.

The key idea of this specification is that when a UA sends a REGISTER request, the proxy can later use this same network "flow", whether this is a bidirectional stream of UDP datagrams, a TCP connection, or an analogous concept of another transport protocol to forward any requests that need to go to this UA. For a UA to receive incoming requests, the UA has to connect to a server. Since the server can't connect to the UA, the UA has to make sure that a flow is always active. This requires the UA to detect when a flow fails. Since such detection takes time and leaves a window of opportunity for missed incoming requests, this mechanism allows the UA to use multiple flows to the proxy or registrar. This mechanism also uses a keep alive mechanism over each flow so that the UA can detect when a flow has failed.



 TOC 

2. Conventions and 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 (Bradner, S., “Key words for use in RFCs to Indicate Requirement Levels,” March 1997.)[RFC2119].

2.1 Definitions

Authoritative Proxy:
A proxy that handles non-REGISTER requests for a specific Address-of-Record (AOR), performs the logical Location Server lookup described in RFC 3261, and forwards those requests to specific Contact URIs.
Edge Proxy:
An Edge Proxy is any proxy that is located topologically between the registering User Agent and the Authoritative Proxy.
Flow:
A Flow is a network protocol layer (layer 4) association between two hosts that is represented by the network address and port number of both ends and by the protocol. For TCP, a flow is equivalent to a TCP connection. For UDP a flow is a bidirectional stream of datagrams between a single pair of IP addresses and ports of both peers. With TCP, a flow often has a one to one correspondence with a single file descriptor in the operating system.
reg-id:
This refers to the value of a new header field parameter value for the Contact header field. When a UA registers multiple times, each simultaneous registration gets a unique reg-id value.
instance-id:
This specification uses the word instance-id to refer to the value of the "sip.instance" media feature tag in the Contact header field. This is a Uniform Resource Name (URN) that uniquely identifies this specific UA instance.
outbound-proxy-set
A set of SIP URIs (Uniform Resource Identifiers) that represents each of the outbound proxies (often Edge Proxies) with which the UA will attempt to maintain a direct flow. The first URI in the set is often referred to as the primary outbound proxy and the second as the secondary outbound proxy. There is no difference between any of the URIs in this set, nor does the primary/secondary terminology imply that one is preferred over the other.



 TOC 

3. Overview

Several scenarios in which this technique is useful are discussed below, including the simple co-located registrar and proxy, a User Agent desiring multiple connections to a resource (for redundancy, for example), and a system that uses Edge Proxies.

3.1 Summary of Mechanism

The overall approach is fairly simple. Each UA has a unique instance-id that stays the same for this UA even if the UA reboots or is power cycled. Each UA can register multiple times over different connections for the same SIP Address of Record (AOR) to achieve high reliability. Each registration includes the instance-id for the UA and a reg-id label that is different for each flow. The registrar can use the instance-id to recognize that two different registrations both reach the same UA. The registrar can use the reg-id label to recognize that a UA is registering after a reboot or a network failure.

When a proxy goes to route a message to a UA for which it has a binding, it can use any one of the flows on which a successful registration has been completed. A failure on a particular flow can be tried again on an alternate flow. Proxies can determine which flows go to the same UA by comparing the instance-id. Proxies can tell that a flow replaces a previously abandoned flow by looking at the reg-id.

UAs use the STUN (Simple Traversal of UDP through NATs) protocol as the keepalive mechanism to keep their flow to the proxy or registrar alive.

3.2 Single Registrar and UA

In the topology shown below, a single server is acting as both a registrar and proxy.

   +-----------+
   | Registrar |
   | Proxy     |
   +-----+-----+
         |
         |
    +----+--+
    | User  |
    | Agent |
    +-------+

User Agents which form only a single flow continue to register normally but include the instance-id as described in Section 4.1 (Instance ID Creation). The UA can also include a reg-id parameter which is used to allow the registrar to detect and avoid keeping invalid contacts when a UA reboots or reconnects after its old connection has failed for some reason.

For clarity, here is an example. Bob's UA creates a new TCP flow to the registrar and sends the following REGISTER request.

REGISTER sip:example.com SIP/2.0
Via: SIP/2.0/TCP 192.0.2.1;branch=z9hG4bK-bad0ce-11-1036
Max-Forwards: 70
From: Bob <sip:bob@example.com>;tag=d879h76
To: Bob <sip:bob@example.com>
Call-ID: 8921348ju72je840.204
CSeq: 1 REGISTER
Supported: path
Contact: <sip:line1@192.168.0.2>; reg-id=1;
 ;+sip.instance="<urn:uuid:00000000-0000-0000-0000-000A95A0E128>"
Content-Length: 0

The registrar challenges this registration to authenticate Bob. When the registrar adds an entry for this contact under the AOR for Bob, the registrar also keeps track of the connection over which it received this registration.

The registrar saves the instance-id ("urn:uuid:00000000-0000-0000-0000-000A95A0E128") and reg-id ("1") along with the rest of the Contact header field. If the instance-id and reg-id are the same as a previous registration for the same AOR, the registrar replaces the old Contact URI and flow information. This allows a UA that has rebooted to replace its previous registration for each flow with minimal impact on overall system load.

When Alice sends a request to Bob, his authoritative proxy selects the target set. The proxy forwards the request to elements in the target set based on the proxy's policy. The proxy looks at the target set and uses the instance-id to understand if two targets both end up routing to the same UA. When the proxy goes to forward a request to a given target, it looks and finds the flows over which it received the registration. The proxy then forwards the request on that flow instead of trying to form a new flow to that contact. This allows the proxy to forward a request to a particular contact over the same flow that the UA used to register this AOR. If the proxy has multiple flows that all go to this UA, it can choose any one of registration bindings for this AOR that has the same instance-id as the selected UA.

3.3 Multiple Connections from a User Agent

There are various ways to deploy SIP to build a reliable and scalable system. This section discusses one such design that is possible with the mechanisms in this specification. Other designs are also possible.

In the example system below, the logical outbound proxy/registrar for the domain is running on two hosts that share the appropriate state and can both provide registrar and outbound proxy functionality for the domain. The UA will form connections to two of the physical hosts that can perform the outbound proxy/registrar function for the domain. Reliability is achieved by having the UA form two TCP connections to the domain.

Scalability is achieved by using DNS SRV to load balance the primary connection across a set of machines that can service the primary connection, and also using DNS SRV to load balance across a separate set of machines that can service the secondary connection. The deployment here requires that DNS is configured with one entry that resolves to all the primary hosts and another entry that resolves to all the secondary hosts. While this introduces additional DNS configuration, the approach works and requires no addition SIP extensions.

Note: Approaches which select multiple connections from a single DNS SRV set were also considered, but cannot prevent two connections from accidentally resolving to the same host. The approach in this document does not prevent future extensions, such as the SIP UA configuration framework (Petrie, D., “A Framework for Session Initiation Protocol User Agent Profile Delivery,” Mar 2006.)[I-D.ietf-sipping-config-framework], from adding other ways for a User Agent to discover its outbound-proxy-set.

    +-------------------+
    | Domain            |
    | Logical Proxy/Reg |
    |                   |
    |+-----+     +-----+|
    ||Host1|     |Host2||
    |+-----+     +-----+|
    +---\------------/--+
         \          /
          \        /
           \      /
            \    /
           +------+
           | User |
           | Agent|
           +------+

The UA is configured with multiple outbound proxy registration URIs. These URIs are configured into the UA through whatever the normal mechanism is to configure the proxy or registrar address in the UA. If the AOR is Alice@example.com, the outbound-proxy-set might look something like "sip:primary.example.com;keepalive=stun" and "sip:secondary.example.com;keepalive=stun". The "keepalive=stun" tag indicates that a SIP server supports STUN and SIP multiplexed over the same flow, as described later in this specification. Note that each URI in the outbound-proxy-set could resolve to several different physical hosts. The administrative domain that created these URIs should ensure that the two URIs resolve to separate hosts. These URIs are handled according to normal SIP processing rules, so mechanisms like SRV can be used to do load balancing across a proxy farm.

The domain also needs to ensure that a request for the UA sent to host1 or host2 is then sent across the appropriate flow to the UA. The domain might choose to use the Path header approach (as described in the next section) to store this internal routing information on host1 or host2.

When a single server fails, all the UAs that have a flow through it will detect a flow failure and try to reconnect. This can cause large loads on the server. When large numbers of hosts reconnect nearly simultaneously, this is referred to as the avalanche restart problem, and is further discussed in Section 4.4.3 (Flow Recovery). The multiple flows to many servers help reduce the load caused by the avalanche restart. If a UA has multiple flows, and one of the servers fails, the UA delays the specified time before trying to form a new connection to replace the flow to the server that failed. By spreading out the time used for all the UAs to reconnect to a server, the load on the server farm is reduced.

When used in this fashion to achieve high reliability, the operator will need to configure DNS such that the various URIs in the outbound proxy set do not resolve to the same host.

3.4 Edge Proxies

Some SIP deployments use edge proxies such that the UA sends the REGISTER to an Edge Proxy that then forwards the REGISTER to the Registrar. The Edge Proxy includes a Path header (Willis, D. and B. Hoeneisen, “Session Initiation Protocol (SIP) Extension Header Field for Registering Non-Adjacent Contacts,” December 2002.)[RFC3327] so that when the registrar later forwards a request to this UA, the request is routed through the Edge Proxy. There could be a NAT or firewall between the UA and the Edge Proxy.

             +---------+
             |Registrar|
             |Proxy    |
             +---------+
              /      \
             /        \
            /          \
         +-----+     +-----+
         |Edge1|     |Edge2|
         +-----+     +-----+
            \           /
             \         /
     ----------------------------NAT/FW
               \     /
                \   /
               +------+
               |User  |
               |Agent |
               +------+

These systems can use effectively the same mechanism as described in the previous sections but need to use the Path header. When the Edge Proxy receives a registration, it needs to create an identifier value that is unique to this flow (and not a subsequent flow with the same addresses) and put this identifier in the Path header URI. This identifier has two purposes. First, it allows the Edge Proxy to map future requests back to the correct flow. Second, because the identifier will only be returned if the user authentication with the registrar succeeds, it allows the Edge Proxy to indirectly check the user's authentication information via the registrar. The identifier SHOULD be placed in the user portion of a loose route in the Path header. If the registration succeeds, the Edge Proxy needs to map future requests that are routed to the identifier value from the Path header, to the associated flow.

The term Edge Proxy is often used to refer to deployments where the Edge Proxy is in the same administrative domain as the Registrar. However, in this specification we use the term to refer to any proxy between the UA and the Registrar. For example the Edge Proxy may be inside an enterprise that requires its use and the registrar could be from a service provider with no relationship to the enterprise. Regardless if they are in the same administrative domain, this specification requires that Registrars and Edge proxies support the Path header mechanism in RFC 3327 (Willis, D. and B. Hoeneisen, “Session Initiation Protocol (SIP) Extension Header Field for Registering Non-Adjacent Contacts,” December 2002.)[RFC3327].

3.5 Keepalive Technique

A keepalive mechanism needs to detect failure of a connection and changes to the NAT public mapping, as well as keeping any NAT bindings refreshed. This specification describes using STUN (Rosenberg, J., “Simple Traversal of UDP Through Network Address Translators (NAT) (STUN),” July 2005.)[I-D.ietf-behave-rfc3489bis] over the same flow as the SIP traffic to perform the keepalive. For connection-oriented transports (e.g. TCP and TLS over TCP), the UAC MAY use TCP keepalives to detect flow failure if the UAC can send these keepalives and detect a keepalive failure according to the time frames described in Section 4.4 (Detecting Flow Failure).

Note: when TCP is being used, it's natural to think of using TCP KEEPALIVE. Unfortunately, many operating systems and programming environments do not allow the keepalive time to be set on a per-connection basis. Thus, applications may not be able to set an appropriate time.

For connection-less transports, a flow definition could change because a NAT device in the network path reboots and the resulting public IP address or port mapping for the UA changes. To detect this, requests are sent over the same flow that is being used for the SIP traffic. The proxy or registrar acts as a STUN server on the SIP signaling port.

Note: The STUN mechanism is very robust and allows the detection of a changed IP address. Many other options were considered, but the SIP Working Group selected the STUN-based approach, since it works over any transport. Approaches using SIP requests were abandoned because to achieve the required performance, the server needs to deviate from the SIP specification in significant ways. This would result in many undesirable and non-deterministic behaviors in some environments.

Another approach considered to detect a changed flow was using OPTIONS messages and the rport parameter. Although the OPTIONS approach has the advantage of being backwards compatible, it also significantly increases the load on the proxy or registrar server. Related to this idea was an idea of creating a new SIP PING method that was like OPTIONS but faster. It would be critical that this PING method did not violate the processing requirements of a proxies and UAS so it was never clear how it would be significantly faster than OPTIONS given it would still have to obey things like checking the Proxy-Require header. After considerable consideration the working group came to some consensus that the STUN approach was a better solution that these alternative designs.

When the UA detects that a flow has failed or that the flow definition has changed, the UA needs to re-register and will use the back-off mechanism described in Section 4 (User Agent Mechanisms) to provide congestion relief when a large number of agents simultaneously reboot.



 TOC 

4. User Agent Mechanisms

4.1 Instance ID Creation

Each UA MUST have an Instance Identifier URN that uniquely identifies the device. Usage of a URN provides a persistent and unique name for the UA instance. It also provides an easy way to guarantee uniqueness within the AOR. This URN MUST be persistent across power cycles of the device. The Instance ID MUST NOT change as the device moves from one network to another.

A UA SHOULD use a UUID URN [RFC4122] (Leach, P., Mealling, M., and R. Salz, “A Universally Unique IDentifier (UUID) URN Namespace,” July 2005.). The UUID URN allows for non-centralized computation of a URN based on time, unique names (such as a MAC address), or a random number generator.

A device like a soft-phone, when first installed, can generate a UUID (Leach, P., Mealling, M., and R. Salz, “A Universally Unique IDentifier (UUID) URN Namespace,” July 2005.)[RFC4122] and then save this in persistent storage for all future use. For a device such as a hard phone, which will only ever have a single SIP UA present, the UUID can include the MAC address and be generated at any time because it is guaranteed that no other UUID is being generated at the same time on that physical device. This means the value of the time component of the UUID can be arbitrarily selected to be any time less than the time when the device was manufactured. A time of 0 (as shown in the example in Section 3.2 (Single Registrar and UA)) is perfectly legal as long as the device knows no other UUIDs were generated at this time.

If a URN scheme other than UUID is used, the URN MUST be selected such that the instance can be certain that no other instance registering against the same AOR would choose the same URN value. An example of a URN that would not meet the requirements of this specification is the national bibliographic number [RFC3188] (Hakala, J., “Using National Bibliography Numbers as Uniform Resource Names,” October 2001.). Since there is no clear relationship between a SIP UA instance and a URN in this namespace, there is no way a selection of a value can be performed that guarantees that another UA instance doesn't choose the same value.

The UA SHOULD include a "sip.instance" media feature tag as a UA characteristic [RFC3840] (Rosenberg, J., Schulzrinne, H., and P. Kyzivat, “Indicating User Agent Capabilities in the Session Initiation Protocol (SIP),” August 2004.) in requests and responses. As described in [RFC3840] (Rosenberg, J., Schulzrinne, H., and P. Kyzivat, “Indicating User Agent Capabilities in the Session Initiation Protocol (SIP),” August 2004.), this media feature tag will be encoded in the Contact header field as the "+sip.instance" Contact header field parameter. The value of this parameter MUST be a URN [RFC2141] (Moats, R., “URN Syntax,” May 1997.). One case where a UA may not want to include the URN in the sip.instance media feature tag is when it is making an anonymous request or some other privacy concern requires that the UA not reveal its identity.

RFC 3840 (Rosenberg, J., Schulzrinne, H., and P. Kyzivat, “Indicating User Agent Capabilities in the Session Initiation Protocol (SIP),” August 2004.)[RFC3840] defines equality rules for callee capabilities parameters, and according to that specification, the "sip.instance" media feature tag will be compared by case-sensitive string comparison. This means that the URN will be encapsulated by angle brackets ("<" and ">") when it is placed within the quoted string value of the +sip.instance Contact header field parameter. The case-sensitive matching rules apply only to the generic usages defined in RFC 3840 (Rosenberg, J., Schulzrinne, H., and P. Kyzivat, “Indicating User Agent Capabilities in the Session Initiation Protocol (SIP),” August 2004.)[RFC3840] and in the caller preferences specification [RFC3841] (Rosenberg, J., Schulzrinne, H., and P. Kyzivat, “Caller Preferences for the Session Initiation Protocol (SIP),” August 2004.). When the instance ID is used in this specification, it is effectively "extracted" from the value in the "sip.instance" media feature tag. Thus, equality comparisons are performed using the rules for URN equality that are specific to the scheme in the URN. If the element performing the comparisons does not understand the URN scheme, it performs the comparisons using the lexical equality rules defined in RFC 2141 [RFC2141] (Moats, R., “URN Syntax,” May 1997.). Lexical equality may result in two URNs being considered unequal when they are actually equal. In this specific usage of URNs, the only element which provides the URN is the SIP UA instance identified by that URN. As a result, the UA instance SHOULD provide lexically equivalent URNs in each registration it generates. This is likely to be normal behavior in any case; clients are not likely to modify the value of the instance ID so that it remains functionally equivalent yet lexigraphically different from previous registrations.

4.2 Initial Registrations

UAs obtain at configuration time one or more SIP URIs representing the default outbound-proxy-set. This specification assumes the set is determined via any of a number of configuration mechanisms, and future specifications may define additional mechanisms such as using DNS to discover this set. How the UA is configured is outside the scope of this specification. However, a UA MUST support sets with at least two outbound proxy URIs and SHOULD support sets with up to four URIs. For each outbound proxy URI in the set, the UA SHOULD send a REGISTER in the normal way using this URI as the default outbound proxy. Forming the route set for the request is outside the scope of this document, but typically results in sending the REGISTER such that the topmost Route header field contains a loose route to the outbound proxy URI. Other issues related to outbound route construction are discussed in [I-D.rosenberg-sip-route-construct] (Rosenberg, J., “Clarifying Construction of the Route Header Field in the Session Initiation Protocol (SIP),” July 2005.).

Registration requests, other than those described in Section 4.2.1 (Registration by Other Instances), MUST include an instance-id media feature tag as specified in Section 4.1 (Instance ID Creation).

These ordinary registration requests MUST also add a distinct reg-id parameter to the Contact header field. Each one of these registrations will form a new flow from the UA to the proxy. The reg-id sequence does not have to be sequential but MUST be exactly the same reg-id sequence each time the device power cycles or reboots so that the reg-id values will collide with the previously used reg-id values. This is so the registrar can replace the older registration.

The UAC MUST indicate that it supports the Path header (Willis, D. and B. Hoeneisen, “Session Initiation Protocol (SIP) Extension Header Field for Registering Non-Adjacent Contacts,” December 2002.)[RFC3327] mechanism, by including the 'path' option-tag in a Supported header field value in its REGISTER requests. Other than optionally examining the Path vector in the response, this is all that is required of the UAC to support Path.

The UAC MAY examine successful registrations for the presence of an 'outbound' option-tag in a Supported header field value. Presence of this option-tag indicates that the registrar is compliant with this specification, and that any edge proxies which need to partcipate are also compliant.

Note that the UA needs to honor 503 responses to registrations as described in RFC 3261 and RFC 3263 (Rosenberg, J. and H. Schulzrinne, “Session Initiation Protocol (SIP): Locating SIP Servers,” June 2002.)[RFC3263]. In particular, implementors should note that when receiving a 503 response with a Retry-After header field, the UA should wait the indicated amount of time and retry the registration. A Retry-After header field value of 0 is valid and indicates the UA should retry the REGISTER immediately. Implementations need to ensure that when retrying the REGISTER, they revisit the DNS resolution results such that the UA can select an alternate host from the one chosen the previous time the URI was resolved.

Finally, re-registrations which merely refresh an existing valid registration SHOULD be sent over the same flow as the original registration.

4.2.1 Registration by Other Instances

A User Agent MUST NOT include a reg-id header parameter in the Contact header field of a registration if the registering UA is not the same instance as the UA referred to by the target Contact header field. (This practice is occasionally used to install forwarding policy into registrars.)

Note that a UAC also MUST NOT include an instance-id or reg-id parameter in a request to unregister all Contacts (a single Contact header field value with the value of "*").

4.3 Sending Requests

When a UA is about to send a request, it first performs normal processing to select the next hop URI. The UA can use a variety of techniques to compute the route set and accordingly the next hop URI. Discussion of these techniques is outside the scope of this document but could include mechanisms specified in RFC 3608 (Willis, D. and B. Hoeneisen, “Session Initiation Protocol (SIP) Extension Header Field for Service Route Discovery During Registration,” October 2003.)[RFC3608] (Service Route) and [I-D.rosenberg-sip-route-construct] (Rosenberg, J., “Clarifying Construction of the Route Header Field in the Session Initiation Protocol (SIP),” July 2005.).

The UA performs normal DNS resolution on the next hop URI (as described in RFC 3263 (Rosenberg, J. and H. Schulzrinne, “Session Initiation Protocol (SIP): Locating SIP Servers,” June 2002.)[RFC3263]) to find a protocol, IP address, and port. For non-TLS protocols, if the UA has an existing flow to this IP address, and port with the correct protocol, then the UA MUST use the existing connection. For TLS protocols, there must also be a match between the host production in the next hop and one of the URIs contained in the subjectAltName in the peer certificate. If the UA cannot use one of the existing flows, then it SHOULD form a new flow by sending a datagram or opening a new connection to the next hop, as appropriate for the transport protocol.

4.4 Detecting Flow Failure

The UA needs to detect when a specific flow fails. The UA actively tries to detect failure by periodically sending keepalive messages using one of the techniques described in this section. If a flow has failed, the UA follows the procedures in Section 4.2 (Initial Registrations) to form a new flow to replace the failed one.

The time between keepalive requests when using UDP-based transports SHOULD be a random number between 24 and 29 seconds while for TCP-based transports it SHOULD be a random number between 95 and 120 seconds. These times MAY be configurable.

4.4.1 Keepalive with TCP KEEPALIVE

User Agents that are capable of generating per-connection TCP keepalives with timer values consistent with those in this section MAY use TCP keepalives instead of using STUN keepalives for TCP-based flows.

4.4.2 Keepalive with STUN

User Agents that form flows, check if the configured URI they are connecting to has a 'keepalive' URI parameter (defined in Section 12 (IANA Considerations)) with the value of 'stun'. If the parameter is present and the UA is not already performing keepalives using another supported mechanism, the UA needs to periodically perform keepalive checks by sending STUN (Rosenberg, J., “Simple Traversal of UDP Through Network Address Translators (NAT) (STUN),” July 2005.)[I-D.ietf-behave-rfc3489bis] Binding Requests over the flow as described in Section 8 (STUN Keepalive Processing).

If the XOR-MAPPED-ADDRESS in the STUN Binding Response changes, the UA MUST treat this event as a failure on the flow.

4.4.3 Flow Recovery

When a flow to a particular URI in the outbound-proxy-set fails, the UA needs to form a new flow to replace the old flow and replace any registrations that were previously sent over this flow. Each new registration MUST have the same reg-id as the registration it replaces. This is done in much the same way as forming a brand new flow as described in Section 4.2 (Initial Registrations); however, if there is a failure in forming this flow, the UA needs to wait a certain amount of time before retrying to form a flow to this particular next hop.

The amount of time to wait depends if the previous attempt at establishing a flow was successful. For the purposes of this section, a flow is considered successful if outbound registration succeeded and keepalives have not timed out for min-regtime seconds (default of 120 seconds) after a registration. For STUN-based keepalives, this means three successful STUN transactions over UDP or one successful STUN transaction over TCP. If a flow is established and is alive after this amount of time, the number of consecutive registration failures is set to zero. Each time a flow fails before two minutes, the number of consecutive registration failures is incremented by one. Note that a failure during the initial STUN validation does not count against the number of consecutive registration failures.

The number of seconds to wait is computed in the following way. If all of the flows to every URI in the outbound proxy set have failed, the base time is set to 30 seconds; otherwise, in the case where at least one of the flows has not failed, the base time is set to 90 seconds. The wait time is computed by taking two raised to the power of the number of consecutive registration failures for that URI, and multiplying this by the base time, up to a maximum of 1800 seconds.

wait-time = min( max-time, (base-time * (2 ^ consecutive-failures)))

These times MAY be configurable in the UA. The four times are:

For example, if the base time is 30 seconds, and there were three failures, then the wait time is min(1800,30*(2^3)) or 240 seconds. The delay time is computed by selecting a uniform random time between 50 and 100 percent of the wait time. The UA MUST wait for the value of the delay time before trying another registration to form a new flow for that URI.

To be explicitly clear on the boundary conditions: when the UA boots it immediately tries to register. If this fails and no registration on other flows succeed, the first retry happens somewhere between 30 and 60 seconds after the failure of the first registration request. If the number of consecutive-failures is large enough that the maximum of 1800 seconds is reached, the UA will keep trying indefinitely with a random time of 15 to 30 minutes (900 to 1800 seconds) between each attempt.



 TOC 

5. Edge Proxy Mechanisms

5.1 Processing Register Requests

When an Edge Proxy receives a registration request with a reg-id header parameter in the Contact header field, it typically needs to store a "flow token", containing information about the flow from the previous hop, in a Path header field value as described in RFC 3327 (Willis, D. and B. Hoeneisen, “Session Initiation Protocol (SIP) Extension Header Field for Registering Non-Adjacent Contacts,” December 2002.)[RFC3327]. The token MAY be placed in the userpart of the URI. If the edge proxy is the first SIP node after the UAC, it either MUST store a flow token in a Path header, or reject the request. In addition, the first node MUST include an 'ob' URI parameter in its Path header field value.

Each subsequent edge proxy examines the first Path header field value. If this URI does not contain an 'ob' parameter, the edge proxy MUST ignore the reg-id parameter and MUST NOT include an 'ob' parameter if it adds a Path header field value. If the first Path header field value contains an 'ob' parameter, this indicates that the first edge proxy performed outbound processing. In this case the edge proxy MUST store a flow token in a Path header, unless it has positive knowledge that the URI in previous Path header is reachable from any node on the public Internet, and that the next hop SIP node can reach any node on the public Internet. This insures that there is a reachable path from the authoritative proxy back to the User Agent. Regardless if the proxy includes a flow token, if it adds a Path header field value, it MUST include the 'ob' parameter in its Path URI.

5.2 Generating Flow Tokens

A trivial but impractical way to satisfy the flow token requirement in Section 5.1 (Processing Register Requests) involves storing a mapping between an incrementing counter and the connection information; however this would require the Edge Proxy to keep an impractical amount of state. It is unclear when this state could be removed and the approach would have problems if the proxy crashed and lost the value of the counter. Two stateless examples are provided below. A proxy can use any algorithm it wants as long as the flow token is unique to a flow, the flow can be recovered from the token, and the token can not be modified by attackers.

Algorithm 1:
The proxy generates a flow token for connection-oriented transports by concatenating the file descriptor (or equivalent) with the NTP time the connection was created, and base64 encoding the result. This results in an identifier approximately 16 octets long. The proxy generates a flow token for UDP by concatenating the file descriptor and the remote IP address and port, then base64 encoding the result. (No NTP time is needed for UDP.) This algorithm MUST NOT be used unless all messages between the Edge proxy and Registrar use a SIPS protected transport. If the SIPS level of integrity protection is not available, an attacker can hijack another user's calls.
Algorithm 2:
When the proxy boots it selects a 20-octet crypto random key called K that only the Edge Proxy knows. A byte array, called S, is formed that contains the following information about the flow the request was received on: an enumeration indicating the protocol, the local IP address and port, the remote IP address and port. The HMAC of S is computed using the key K and the HMAC-SHA1-80 algorithm, as defined in [RFC2104] (Krawczyk, H., Bellare, M., and R. Canetti, “HMAC: Keyed-Hashing for Message Authentication,” February 1997.). The concatenation of the HMAC and S are base64 encoded, as defined in [RFC3548] (Josefsson, S., “The Base16, Base32, and Base64 Data Encodings,” July 2003.), and used as the flow identifier. When using IPv4 addresses, this will result in a 32-octet identifier.

5.3 Forwarding Requests

When an Edge Proxy receives a request, it applies normal routing procedures with the following addition. If the Edge Proxy receives a request where the edge proxy is the host in the topmost Route header field value, and the Route header contains a flow token, the proxy compares the flow in the flow token with the source of the request. If these refer to the same flow, the Edge Proxy removes the Route header and continues processing the request. Otherwise, if the top-most Route header refers to the Edge Proxy and contains a valid flow identifier token created by this proxy, the proxy MUST remove the the Route header and forward the request over the flow that received the REGISTER request that caused the flow identifier token to be created. For connection-oriented transports, if the flow no longer exists the proxy SHOULD send a 430 Flow Failed response to the request.

The advantage to a stateless approach to managing the flow information is that there is no state on the Edge Proxy that requires clean up or that has to be synchronized with the registrar.

Proxies which used one of the two algorithms described in this document to form a flow token follow the procedures below to determine the correct flow.

Algorithm 1:
The proxy base64 decodes the user part of the Route header. For a TCP-based transport, if a connection specified by the file descriptor is present and the creation time of the file descriptor matches the creation time encoded in the Route header, the proxy forwards the request over that connection. For a UDP-based transport, the proxy forwards the request from the encoded file descriptor to the source IP address and port.
Algorithm 2:
To decode the flow token, take the flow identifier in the user portion of the URI and base64 decode it, then verify the HMAC is correct by recomputing the HMAC and checking it matches. If the HMAC is not correct, the proxy SHOULD send a 403 response. If the HMAC is correct then the proxy SHOULD forward the request on the flow that was specified by the information in the flow identifier. If this flow no longer exists, the proxy SHOULD send a 430 Flow Failed response to the request.

Note that this specification needs mid-dialog requests to be routed over the same flows as those stored in the Path vector from the initial registration, but techniques to ensure that mid-dialog requests are routed over an existing flow are not part of this specification. However, an approach such as having the Edge Proxy Record-Route with a flow token is one way to ensure that mid-dialog requests are routed over the correct flow.

5.4 Edge Proxy Keepalive Handling

All edge proxies compliant with this specification MUST implement support for the STUN NAT Keepalive usage on its SIP ports as described in Section 8 (STUN Keepalive Processing).



 TOC 

6. Registrar Mechanisms: Processing REGISTER Requests

This specification updates the definition of a binding in RFC 3261 (Rosenberg, J., Schulzrinne, H., Camarillo, G., Johnston, A., Peterson, J., Sparks, R., Handley, M., and E. Schooler, “SIP: Session Initiation Protocol,” June 2002.)[RFC3261] Section 10 and RFC 3327 (Willis, D. and B. Hoeneisen, “Session Initiation Protocol (SIP) Extension Header Field for Registering Non-Adjacent Contacts,” December 2002.)[RFC3327] Section 5.3.

When no +sip.instance media feature parameter is present in a Contact header field value in a REGISTER request, the corresponding binding is still between an AOR and the URI from that Contact header field value. When a +sip.instance media feature parameter is present in a Contact header field value in a REGISTER request, the corresponding binding is between an AOR and the combination of the instance-id (from the +sip.instance media feature parameter) and the value of reg-id parameter if it is present. For a binding with an instance-id, the registrar still stores the Contact header field value URI with the binding, but does not consider the Contact URI for comparison purposes. A Contact header field value with an instance-id but no reg-id is valid, but one with a reg-id but no instance-id is not. If the registrar processes a Contact header field value with a reg-id but no instance-id, it simply ignores the reg-id parameter. The registrar MUST be prepared to receive, simultaneously for the same AOR, some registrations that use instance-id and reg-id and some registrations that do not.

Registrars which implement this specification MUST support the Path header mechanism (Willis, D. and B. Hoeneisen, “Session Initiation Protocol (SIP) Extension Header Field for Registering Non-Adjacent Contacts,” December 2002.)[RFC3327].

In addition to the normal information stored in the binding record, some additional information needs to be stored for any registration that contains a reg-id header parameter in the Contact header field value. First the registrar examines all Path header field values, if any. If any of these does not have an 'ob' URI parameter, the registrar MUST ignore the reg-id parameter and continue processing the request as if it did not support this specification. Likewise if the REGISTER request visited an edge proxy, but no Path header field values are present, the registrar MUST ignore the reg-id parameter. Specifically, the registrar MUST use RFC 3261 Contact binding rules, and MUST NOT include the 'outbound' option-tag in its Supported header field.

If the UAC has a direct flow with the registrar, the registrar MUST store enough information to uniquely identify the network flow over which the request arrived. For common operating systems with TCP, this would typically just be the file descriptor and the time the file descriptor was opened. For common operating systems with UDP this would typically be the file descriptor for the local socket that received the request, the local interface, and the IP address and port number of the remote side that sent the request.

In addition, unless the registrar has positive knowledge that the topmost Path header URI is reachable from the authoritative proxy, it must store the flow information for the previous hop. The registrar MAY store this information by adding itself to the Path header field with an appropriate flow token.

The registrar MUST also store all the Contact header field information including the reg-id and instance-id parameters and SHOULD also store the time at which the binding was last updated. If a Path header field is present, RFC 3327 (Willis, D. and B. Hoeneisen, “Session Initiation Protocol (SIP) Extension Header Field for Registering Non-Adjacent Contacts,” December 2002.)[RFC3327] requires the registrar to store this information as well. If the registrar receives a re-registration, it MUST update any information that uniquely identifies the network flow over which the request arrived if that information has changed, and SHOULD update the time the binding was last updated.

The Registrar MUST include the 'outbound' option-tag (defined in Section (Contact Header Field)) in a Supported header field value in its responses to REGISTER requests for which it has performed outbound processing. The Registrar MAY be configured with local policy to reject any registrations that do not include the instance-id and reg-id. Note that the requirements in this section applies to both REGISTER requests received from an Edge Proxy as well as requests received directly from the UAC.

To be compliant with this specification, registrars which can receive SIP requests directly from a UAC without intervening edge proxies MUST implement support the STUN NAT Keepalive usage on its SIP ports as described in Section 8 (STUN Keepalive Processing).



 TOC 

7. Authoritative Proxy Mechansims: Forwarding Requests

When a proxy uses the location service to look up a registration binding and then proxies a request to a particular contact, it selects a contact to use normally, with a few additional rules:

The proxy uses normal forwarding rules looking at the next-hop target of the message and the value of any stored Path header field vector in the registration binding to decide how to forward the request and populate the Route header in the request. Additionally, when the proxy forwards a request to a binding that contains a reg-id, if the binding has a previous hop flow associated with it, the proxy MUST send the request over the same network flow that was saved with the binding. This means that for TCP, the request MUST be sent on the same TCP socket that received the REGISTER request. For UDP, the request MUST be sent from the same local IP address and port over which the registration was received, to the same IP address and port from which the REGISTER was received.

If a proxy or registrar receives information from the network that indicates that no future messages will be delivered on a specific flow, then the proxy MUST invalidate all the bindings in the target set that use that flow (regardless of AOR). Examples of this are a TCP socket closing or receiving a destination unreachable ICMP error on a UDP flow. Similarly, if a proxy closes a file descriptor, it MUST invalidate all the bindings in the target set with flows that use that file descriptor.



 TOC 

8. STUN Keepalive Processing

This section describes changes to the SIP transport layer that allow SIP and the STUN (Rosenberg, J., “Simple Traversal of UDP Through Network Address Translators (NAT) (STUN),” July 2005.)[I-D.ietf-behave-rfc3489bis] NAT Keepalive usage to be mixed over the same flow. The STUN messages are used to verify connectivity is still available over a flow and to provide periodic keepalives. Note that these STUN keepalives are always sent to the next SIP hop. STUN messages are not delivered end-to-end.

The only STUN messages required by this usage are Binding Requests, Binding Responses, and Error Responses. The UAC sends Binding Requests over the same UDP flow, TCP connection, or TLS channel used for sending SIP messages. These Binding Requests do not require any STUN attributes. The UAS responds to a valid Binding Request with a Binding Response which MUST include the XOR-MAPPED-ADDRESS attribute. After a successful STUN response is received over TCP or TLS over TCP, the underlying TCP connection is left in the active state.

If a server compliant to this section receives SIP requests on a given interface and port, it MUST also provide a limited version of a STUN server on the same interface and port as described in Section 12.3 of [I-D.ietf-behave-rfc3489bis] (Rosenberg, J., “Simple Traversal of UDP Through Network Address Translators (NAT) (STUN),” July 2005.). When STUN messages are sent with a SIP over TLS over TCP flow, the STUN messages are sent inside the TLS-protected channel.

It is easy to distinguish STUN and SIP packets sent over UDP, because the first octet of a STUN packet has a value of 0 or 1 while the first octet of a SIP message is never a 0 or 1. For TCP or TLS over TCP flows, determining if the first octet of the next message in a stream is SIP or STUN is still straightforward, however implementations need to be preared to receive STUN messages which cross a stream buffer boundary, and SIP and STUN messages which share the same stream buffer.

Because sending and receiving binary STUN data on the same ports used for SIP is a significant and non-backwards compatible change to RFC 3261, this section requires a number of checks before sending STUN messages to a SIP node. If a SIP node sends STUN requests (for example due to misconfiguration) despite these warnings, the node may be blacklisted for UDP traffic, or cause its TCP server to loose framing over its connection. For each target node (as determined by IP address, address family, and port number), the sender needs to determine if that destination is validated to support STUN, that it does not support STUN, or that it needs to be validated.

When a URI is created that refers to a SIP device that supports STUN as described in this section, the 'keepalive' URI parameter, as defined in Section 12 (IANA Considerations) SHOULD be added to the URI, with a value of 'stun'. This allows a UA to inspect the URI to decide if it should attempt to send STUN requests to this location.

A SIP node MUST NOT send STUN requests over a flow unless it has an explicit indication that the target next hop SIP server claims to support STUN. For example, automatic or manual configuration of an outbound-proxy-set which contains the keepalive=stun parameter is considered sufficient explicit indication. Note that UACs MUST NOT use an ambiguous configuration option such as "Work through NATs?" or "Do Keepalives?" to imply next hop STUN support. A SIP node MAY also probe the next hop using a SIP OPTIONS request to check for support of the 'sip-stun' option tag in a Supported header field.

Futhermore, even with explicit indication of next hop STUN support, a SIP node needs to validate support for STUN the first time it sends traffic to a specific unvalidated target destination. If an explicit probe indicates support for the 'sip-stun' option-tag, that destination is validated for STUN support. If an explicit probe does not indicate support for the 'sip-stun' option-tag, the target destination does not support STUN request, and the UAC MUST NOT send further STUN requests to this destination. A SIP node MAY send one STUN request and its retransmissions to an unvalidated destination. If a STUN request ever succeeds to a destination, that destination is thereafter validated for STUN support. If this initial STUN request does not result in a STUN response, the SIP node MUST NOT send additional STUN requests over this flow, unless and until a next-hop probe later validates the destination. In addition, the SIP node SHOULD remember unvalidated destination nodes that have been used within one hour and SHOULD NOT send additional STUN messages to any of these destinations. Note that until STUN support has been verified, an initial STUN failure over UDP is not considered a flow failure. For UDP flows, an unvalidated flow can still be reused for SIP traffic, however for unvalidated TCP or TLS over TCP flows, the connection over which STUN requests were sent MUST be closed.

Typically, a SIP node first sends a SIP request and waits to receive a final response (other than a 408 response) over a flow to a new target destination, before sending any STUN messages. When scheduled for the next NAT refresh, the SIP node sends a STUN request to the target. If none of the STUN requests succeed (result in a STUN success response), and the UAC has not already done so, the UAC sends an OPTIONS request to the next hop to verify support for the 'sip-stun' option-tag.

Once a destination is validated to support STUN messages, failure of a STUN request (including its retransmissions) is considered a failure of the underlying flow. For SIP over UDP flows, if the XOR-MAPPED-ADDRESS returned over the flow changes, this indicates that the underlying connectivity has changed, and is considered a flow failure. A 408 response to a next-hop OPTIONS probe is also considered a flow failure.

8.1 Explicit Probes

This section defines a new SIP option-tag called 'sip-stun'. Advertising this option-tag indicates that the server can receive SIP messages and STUN messages as part of the NAT Keepalive usage on the same port. Clients that want to probe a SIP server to determine support for STUN, can send an OPTIONS request to the next hop by setting the Max-Forwards header field to 0. The OPTIONS response will contain a Supported header field with a list of the server's supported option-tags.

A UAC SHOULD NOT include the 'sip-stun' option-tag in a Proxy-Require header. This is because a request with this header will fail in some topologies where the first proxy support sip-stun, but a subsequent proxy does not. Note that RFC 3261 does not allow proxies to remove option-tags from a Proxy-Require header field.

8.2 Use with Sigcomp

When STUN is used together with SigComp (Price, R., Bormann, C., Christoffersson, J., Hannu, H., Liu, Z., and J. Rosenberg, “Signaling Compression (SigComp),” January 2003.)[RFC3320] compressed SIP messages over the same flow, how the STUN messages are sent depends on the transport protocol. For UDP flows, the STUN messages are simply sent uncompressed, "outside" of SigComp. This is supported by multiplexing STUN messages with SigComp messages by checking the two topmost bits of the message. These bits are always one for SigComp, or zero for STUN.

All SigComp messages contain a prefix (the five most-significant bits of the first byte are set to one) that does not occur in UTF-8 encoded text messages [RFC-2279], so for applications which use this encoding (or ASCII encoding) it is possible to multiplex uncompressed application messages and SigComp messages on the same UDP port.

The most significant two bits of every STUN message are both zeroes. This, combined with the magic cookie, aids in differentiating STUN packets from other protocols when STUN is multiplexed with other protocols on the same port.

For TCP-based flows, SigComp requires that all messages are processed by the SigComp compressor to facilitate framing. For these transports, STUN messages are sent encapsulated in the SigComp "well-known shim header" as described in Appendix A of RFC 3320 (Price, R., Bormann, C., Christoffersson, J., Hannu, H., Liu, Z., and J. Rosenberg, “Signaling Compression (SigComp),” January 2003.)[RFC3320].



 TOC 

9. Example Message Flow

The following call flow shows a basic registration and an incoming call. At some point, the flow to the Primary proxy is lost. An incoming INVITE tries to reach the Callee through the Primary flow, but receives an ICMP Unreachable message. The Caller retries using the Secondary Edge Proxy, which uses a separate flow. Later, after the Primary reboots, The Callee discovers the flow failure and reestablishes a new flow to the Primary.


                [-----example.com domain -------------------]
Caller           Secondary             Primary            Callee
  |                 |                  |     (1) REGISTER |
  |                 |                  |<-----------------|
  |                 |                  |(2) 200 OK        |
  |                 |                  |----------------->|
  |                 |                  |     (3) REGISTER |
  |                 |<------------------------------------|
  |                 |(4) 200 OK        |                  |
  |                 |------------------------------------>|
  |                 |                  |                  |
  |                 |           CRASH  X                  |
  |(5) INVITE       |                  |                  |
  |----------------------------------->|                  |
  |(6) ICMP Unreachable                |                  |
  |<-----------------------------------|                  |
  |(7) INVITE       |                  |                  |
  |---------------->|                  |                  |
  |                 |(8) INVITE        |                  |
  |                 |------------------------------------>|
  |                 |(9) 200 OK        |                  |
  |                 |<------------------------------------|
  |(10) 200 OK      |                  |                  |
  |<----------------|                  |                  |
  |(11) ACK         |                  |                  |
  |---------------->|                  |                  |
  |                 |(12) ACK          |                  |
  |                 |------------------------------------>|
  |                 |                  |                  |
  |                 |          REBOOT  |                  |
  |                 |                  |(13) REGISTER     |
  |                 |                  |<-----------------|
  |                 |                  |(14) 200 OK       |
  |                 |                  |----------------->|
  |                 |                  |                  |
  |(15) BYE         |                  |                  |
  |---------------->|                  |                  |
  |                 | (16) BYE         |                  |
  |                 |------------------------------------>|
  |                 |                  |      (17) 200 OK |
  |                 |<------------------------------------|
  |     (18) 200 OK |                  |                  |
  |<----------------|                  |                  |
  |                 |                  |                  |

This call flow assumes that the Callee has been configured with a proxy set that consists of "sip:pri.example.com;lr;keepalive=stun" and "sip:sec.example.com;lr;keepalive=stun". The Callee REGISTER in message (1) looks like:


REGISTER sip:example.com SIP/2.0
Via: SIP/2.0/UDP 192.0.2.1;branch=z9hG4bKnashds7
Max-Forwards: 70
From: Callee <sip:callee@example.com>;tag=7F94778B653B
To: Callee <sip:callee@example.com>
Call-ID: 16CB75F21C70
CSeq: 1 REGISTER
Supported: path
Route: <sip:pri.example.com;lr;keepalive=stun>
Contact: <sip:callee@192.0.2.1>
  ;+sip.instance="<urn:uuid:0C67446E-F1A1-11D9-94D3-000A95A0E128>"
  ;reg-id=1
Content-Length: 0

In the message, note that the Route is set and the Contact header field value contains the instance-id and reg-id. The response to the REGISTER in message (2) would look like:


SIP/2.0 200 OK
Via: SIP/2.0/UDP 192.0.2.1;branch=z9hG4bKnashds7
From: Callee <sip:callee@example.com>;tag=7F94778B653B
To: Callee <sip:callee@example.com>;tag=6AF99445E44A
Call-ID: 16CB75F21C70
CSeq: 1 REGISTER
Supported: outbound
Contact: <sip:callee@192.0.2.1>
  ;+sip.instance="<urn:uuid:0C67446E-F1A1-11D9-94D3-000A95A0E128>"
  ;reg-id=1
  ;expires=3600
Content-Length: 0

The second registration in message 3 and 4 are similar other than the Call-ID has changed, the reg-id is 2, and the route is set to the secondary instead of the primary. They look like:


REGISTER sip:example.com SIP/2.0
Via: SIP/2.0/UDP 192.0.2.1;branch=z9hG4bKnqr9bym
Max-Forwards: 70
From: Callee <sip:callee@example.com>;tag=755285EABDE2
To: Callee <sip:callee@example.com>
Call-ID: E05133BD26DD
CSeq: 1 REGISTER
Supported: path
Route: <sip:sec.example.com;lr;keepalive=stun>
Contact: <sip:callee@192.0.2.1>
  ;+sip.instance="<urn:uuid:0C67446E-F1A1-11D9-94D3-000A95A0E128>"
  ;reg-id=2
Content-Length: 0
SIP/2.0 200 OK
Via: SIP/2.0/UDP 192.0.2.1;branch=z9hG4bKnqr9bym
From: Callee <sip:callee@example.com>;tag=755285EABDE2
To: Callee <sip:callee@example.com>;tag=49A9AD0B3F6A
Call-ID: E05133BD26DD
Supported: outbound
CSeq: 1 REGISTER
Contact: <sip:callee@192.0.2.1>
  ;+sip.instance="<urn:uuid:0C67446E-F1A1-11D9-94D3-000A95A0E128>"
  ;reg-id=1
  ;expires=3600
Contact: <sip:callee@192.0.2.1>
  ;+sip.instance="<urn:uuid:0C67446E-F1A1-11D9-94D3-000A95A0E128>"
  ;reg-id=2
  ;expires=3600
Content-Length: 0

The messages in the call flow are very normal. The only interesting thing to note is that the INVITE in message 8 contains a Record-Route header for the Secondary proxy, with its flow token.

Record-Route:
 <sip:PQPbqQE+Ynf+tzRPD27lU6uxkjQ8LLUG@sec.example.com;lr;user=flow>

The registrations in message 13 and 14 are the same as message 1 and 2 other than the Call-ID and tags have changed. Because these messages will contain the same instance-id and reg-id as those in 1 and 2, this flow will partially supersede that for messages 1 and 2 and will be tried first by Primary.



 TOC 

10. Grammar

This specification defines new Contact header field parameters, reg-id and +sip.instance. The grammar includes the definitions from RFC 3261 (Rosenberg, J., Schulzrinne, H., Camarillo, G., Johnston, A., Peterson, J., Sparks, R., Handley, M., and E. Schooler, “SIP: Session Initiation Protocol,” June 2002.)[RFC3261] and includes the definition of uric from RFC 2396 (Berners-Lee, T., Fielding, R., and L. Masinter, “Uniform Resource Identifiers (URI): Generic Syntax,” August 1998.)[RFC2396].

Note: The "=/" syntax used in this ABNF indicates an extension of the production on the left hand side.

The ABNF[RFC4234] (Crocker, D. and P. Overell, “Augmented BNF for Syntax Specifications: ABNF,” October 2005.) is:

 contact-params =/ c-p-reg / c-p-instance

 c-p-reg        = "reg-id" EQUAL 1*DIGIT ; 1 to 2**31

 c-p-instance   =  "+sip.instance" EQUAL
                   LDQUOT "<" instance-val ">" RDQUOT

 instance-val   = *uric ; defined in RFC 2396

The value of the reg-id MUST NOT be 0 and MUST be less than 2**31.



 TOC 

11. Definition of 430 Flow Failed response code

This specification defines a new SIP response code '430 Flow Failed'. This response code is used by an Edge Proxy to indicate to the Authoritative Proxy that a specific flow to a UA instance has failed. Other flows to the same instance may still succeed. The Authoritative Proxy SHOULD attempt to forward to another target (flow) with the same instance-id and AOR.



 TOC 

12. IANA Considerations

12.1 Contact Header Field

This specification defines a new Contact header field parameter called reg-id in the "Header Field Parameters and Parameter Values" sub-registry as per the registry created by [RFC3968] (Camarillo, G., “The Internet Assigned Number Authority (IANA) Header Field Parameter Registry for the Session Initiation Protocol (SIP),” December 2004.). The required information is:

 Header Field                  Parameter Name   Predefined  Reference
                                                  Values
 ____________________________________________________________________
 Contact                       reg-id               Yes    [RFC AAAA]

 [NOTE TO RFC Editor: Please replace AAAA with
                      the RFC number of this specification.]

12.2 SIP/SIPS URI Parameters

This specification arguments the "SIP/SIPS URI Parameters" sub-registry as per the registry created by [RFC3969] (Camarillo, G., “The Internet Assigned Number Authority (IANA) Uniform Resource Identifier (URI) Parameter Registry for the Session Initiation Protocol (SIP),” December 2004.). The required information is:

    Parameter Name  Predefined Values  Reference
    ____________________________________________
    keepalive        stun               [RFC AAAA]
    ob                                  [RFC AAAA]

    [NOTE TO RFC Editor: Please replace AAAA with
                         the RFC number of this specification.]

12.3 SIP Option Tag

This specification registers two new SIP option tags, as per the guidelines in Section 27.1 of RFC 3261.

Name:
outbound
Description:
This option-tag is used to identify Registrars which support extensions for Client Initiated Connections. A Registrar places this option-tag in a Supported header to communicate the Registrar's support for this extension to the registering User Agent.

Name:
sip-stun
Description:
This option-tag is used to identify SIP servers which can receive STUN requests described in the STUN NAT Keepalive usage on the same ports they use to receive SIP messages.

12.4 Response Code

This section registers a new SIP Response Code, as per the guidelines in Section 27.1 of RFC 3261.

Code:
430
Default Reason Phrase:
Flow Failed
Reference:
This document

12.5 Media Feature Tag

This section registers a new media feature tag, per the procedures defined in RFC 2506 (Holtman, K., Mutz, A., and T. Hardie, “Media Feature Tag Registration Procedure,” March 1999.)[RFC2506]. The tag is placed into the sip tree, which is defined in RFC 3840 (Rosenberg, J., Schulzrinne, H., and P. Kyzivat, “Indicating User Agent Capabilities in the Session Initiation Protocol (SIP),” August 2004.)[RFC3840].

Media feature tag name: sip.instance

ASN.1 Identifier: New assignment by IANA.

Summary of the media feature indicated by this tag: This feature tag contains a string containing a URN that indicates a unique identifier associated with the UA instance registering the Contact.

Values appropriate for use with this feature tag: String.

The feature tag is intended primarily for use in the following applications, protocols, services, or negotiation mechanisms: This feature tag is most useful in a communications application, for describing the capabilities of a device, such as a phone or PDA.

Examples of typical use: Routing a call to a specific device.

Related standards or documents: RFC XXXX

[[Note to IANA: Please replace XXXX with the RFC number of this specification.]]

Security Considerations: This media feature tag can be used in ways which affect application behaviors. For example, the SIP caller preferences extension (Rosenberg, J., Schulzrinne, H., and P. Kyzivat, “Caller Preferences for the Session Initiation Protocol (SIP),” August 2004.)[RFC3841] allows for call routing decisions to be based on the values of these parameters. Therefore, if an attacker can modify the values of this tag, they may be able to affect the behavior of applications. As a result, applications which utilize this media feature tag SHOULD provide a means for ensuring its integrity. Similarly, this feature tag should only be trusted as valid when it comes from the user or user agent described by the tag. As a result, protocols for conveying this feature tag SHOULD provide a mechanism for guaranteeing authenticity.



 TOC 

13. Security Considerations

One of the key security concerns in this work is making sure that an attacker cannot hijack the sessions of a valid user and cause all calls destined to that user to be sent to the attacker.

The simple case is when there are no edge proxies. In this case, the only time an entry can be added to the routing for a given AOR is when the registration succeeds. SIP already protects against attackers being able to successfully register, and this scheme relies on that security. Some implementers have considered the idea of just saving the instance-id without relating it to the AOR with which it registered. This idea will not work because an attacker's UA can impersonate a valid user's instance-id and hijack that user's calls.

The more complex case involves one or more edge proxies. When a UA sends a REGISTER request through an Edge Proxy on to the registrar, the Edge Proxy inserts a Path header field value. If the registration is successfully authenticated, the registrar stores the value of the Path header field. Later when the registrar forwards a request destined for the UA, it copies the stored value of the Path header field into the Route header field of the request and forwards the request to the Edge Proxy.

The only time an Edge Proxy will route over a particular flow is when it has received a Route header that has the flow identifier information that it has created. An incoming request would have gotten this information from the registrar. The registrar will only save this information for a given AOR if the registration for the AOR has been successful; and the registration will only be successful if the UA can correctly authenticate. Even if an attacker has spoofed some bad information in the Path header sent to the registrar, the attacker will not be able to get the registrar to accept this information for an AOR that does not belong to the attacker. The registrar will not hand out this bad information to others, and others will not be misled into contacting the attacker.



 TOC 

14. Requirements

This specification was developed to meet the following requirements:

  1. Must be able to detect that a UA supports these mechanisms.
  2. Support UAs behind NATs.
  3. Support TLS to a UA without a stable DNS name or IP address.
  4. Detect failure of a connection and be able to correct for this.
  5. Support many UAs simultaneously rebooting.
  6. Support a NAT rebooting or resetting.
  7. Minimize initial startup load on a proxy.
  8. Support architectures with edge proxies.



 TOC 

15. Changes

Note to RFC Editor: Please remove this whole section.

15.1 Changes from 04 Version

Moved STUN to a separate section. Reference this section from within the relevant sections in the rest of the document.

Add language clarifying that UA MUST NOT send STUN without an explicit indication the server supports STUN.

Add language describing that UA MUST stop sending STUN if it appears the server does not support it.

Defined a 'sip-stun' option tag. UAs can optionally probe servers for it with OPTIONS. Clarified that UAs SHOULD NOT put this in a Proxy-Require. Explain that the first-hop MUST support this option-tag.

Clarify that SIP/STUN in TLS is on the "inside". STUN used with Sigcomp-compressed SIP is "outside" the compression layer for UDP, but wrapped inside the well-known shim header for TCP-based transports.

Clarify how to decide what a consecutive registration timer is. Flow must be up for some time (default 120 seconds) otherwise previous registration is not considered successful.

Change UAC MUST-->SHOULD register a flow for each member of outbound-proxy-set.

Reworded registrar and proxy in some places (introduce the term "Authoritative Proxy").

Loosened restrictions on always storing a complete Path vector back to the registrar/authoritative proxy if a previous hop in the path vector is reachable.

Added comment about reregistration typically happening over same flow as original registration.

Changed 410 Gone to new response code 430 Flow Failed. Was going to change this to 480 Temporarily Unavailable. Unfortunately this would mean that the authoritative proxy deletes all flows of phones who use 480 for Do Not Disturb. Oops!

Restored sanity by restoring text which explains that registrations with the same reg-id replace the old registration.

Added text about the 'ob' parameter which is used in Path header field URIs to make sure that the previous proxy that added a Path understood outbound processing. The registrar doesn't include Supported: outbound unless it could actually do outbound processing (ex: any Path headers have to have the 'ob' parameter).

Added some text describing what a registration means when there is an instance-id, but no reg-id.

15.2 Changes from 03 Version

Added non-normative text motivating STUN vs. SIP PING, OPTIONS, and Double CRLF. Added discussion about why TCP Keepalives are not always available.

Explained more clearly that outbound-proxy-set can be "configured" using any current or future, manual or automatic configuration/discovery mechanism.

Added a sentence which prevents an Edge Proxy from forwarding back over the flow over which the request is received if the request happens to contain a flow token for that flow. This was an oversight.

Updated example message flow to show a failover example using a new dialog-creating request instead of a mid-dialog request. The old scenario was leftover from before the outbound/gruu reorganization.

Fixed tags, Call-IDs, and branch parameters in the example messages.

Made the ABNF use the "=/" production extension mechanism recommended by Bill Fenner.

Added a table in an appendix expanding the default flow recovery timers.

Incorporated numerous clarifications and rewordings for better comprehension.

Fixed many typos and spelling misteaks.

15.3 Changes from 02 Version

Removed Double CRLF Keepalive

Changed ;sip-stun syntax to ;keepalive=stun

Fixed incorrect text about TCP keepalives.

15.4 Changes from 01 Version

Moved definition of instance-id from GRUU[I-D.ietf-sip-gruu] (Rosenberg, J., “Obtaining and Using Globally Routable User Agent (UA) URIs (GRUU) in the Session Initiation Protocol (SIP),” July 2005.) draft to this draft.

Added tentative text about Double CRLF Keepalive

Removed pin-route stuff

Changed the name of "flow-id" to "reg-id"

Reorganized document flow

Described the use of STUN as a proper STUN usage

Added 'outbound' option-tag to detect if registrar supports outbound

15.5 Changes from 00 Version

Moved TCP keepalive to be STUN.

Allowed SUBSCRIBE to create flow mappings. Added pin-route option tags to support this.

Added text about updating dialog state on each usage after a connection failure.



 TOC 

16. Acknowledgments

Jonathan Rosenberg provided many comments and useful text. Dave Oran came up with the idea of using the most recent registration first in the proxy. Alan Hawrylyshen co-authored the draft that formed the initial text of this specification. Additionally, many of the concepts here originated at a connection reuse meeting at IETF 60 that included the authors, Jon Peterson, Jonathan Rosenberg, Alan Hawrylyshen, and Paul Kyzivat. The TCP design team consisting of Chris Boulton, Scott Lawrence, Rajnish Jain, Vijay K. Gurbani, and Ganesh Jayadevan provided input and text. Nils Ohlmeier provided many fixes and initial implementation experience. In addition, thanks to the following folks for useful comments: Francois Audet, Flemming Andreasen, Mike Hammer, Dan Wing, Srivatsa Srinivasan, Dale Worely, Juha Heinanen, Eric Rescorla, Lyndsay Campbell, and Erkki Koivusalo.



 TOC 

Appendix A. Default Flow Registration Backoff Times

The base-time used for the flow re-registration backoff times described in Section 4.4.3 (Flow Recovery) are configurable. If the base-time-all-fail value is set to the default of 30 seconds and the base-time-not-failed value is set to the default of 90 seconds, the following table shows the resulting delay values.

# of reg failures all flows unusable >1 non-failed flow
0 0 secs 0 secs
1 30-60 secs 90-180 secs
2 1-2 mins 3-6 mins
3 2-4 mins 6-12 mins
4 4-8 mins 12-24 mins
5 8-16 mins 15-30 mins
6 or more 15-30 mins 15-30 mins


 TOC 

17. References



 TOC 

17.1 Normative References

[I-D.ietf-behave-rfc3489bis] Rosenberg, J., “Simple Traversal of UDP Through Network Address Translators (NAT) (STUN),” draft-ietf-behave-rfc3489bis-02 (work in progress), July 2005.
[RFC2119] Bradner, S., “Key words for use in RFCs to Indicate Requirement Levels,” BCP 14, RFC 2119, March 1997 (TXT, HTML, XML).
[RFC2141] Moats, R., “URN Syntax,” RFC 2141, May 1997 (TXT, HTML, XML).
[RFC2396] Berners-Lee, T., Fielding, R., and L. Masinter, “Uniform Resource Identifiers (URI): Generic Syntax,” RFC 2396, August 1998 (TXT, HTML, XML).
[RFC2506] Holtman, K., Mutz, A., and T. Hardie, “Media Feature Tag Registration Procedure,” BCP 31, RFC 2506, March 1999.
[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.
[RFC3320] Price, R., Bormann, C., Christoffersson, J., Hannu, H., Liu, Z., and J. Rosenberg, “Signaling Compression (SigComp),” RFC 3320, January 2003.
[RFC3327] Willis, D. and B. Hoeneisen, “Session Initiation Protocol (SIP) Extension Header Field for Registering Non-Adjacent Contacts,” RFC 3327, December 2002.
[RFC3840] Rosenberg, J., Schulzrinne, H., and P. Kyzivat, “Indicating User Agent Capabilities in the Session Initiation Protocol (SIP),” RFC 3840, August 2004.
[RFC3841] Rosenberg, J., Schulzrinne, H., and P. Kyzivat, “Caller Preferences for the Session Initiation Protocol (SIP),” RFC 3841, August 2004.
[RFC3968] Camarillo, G., “The Internet Assigned Number Authority (IANA) Header Field Parameter Registry for the Session Initiation Protocol (SIP),” BCP 98, RFC 3968, December 2004.
[RFC3969] Camarillo, G., “The Internet Assigned Number Authority (IANA) Uniform Resource Identifier (URI) Parameter Registry for the Session Initiation Protocol (SIP),” BCP 99, RFC 3969, December 2004.
[RFC4122] Leach, P., Mealling, M., and R. Salz, “A Universally Unique IDentifier (UUID) URN Namespace,” RFC 4122, July 2005.
[RFC4234] Crocker, D. and P. Overell, “Augmented BNF for Syntax Specifications: ABNF,” RFC 4234, October 2005.


 TOC 

17.2 Informative References

[I-D.ietf-sip-gruu] Rosenberg, J., “Obtaining and Using Globally Routable User Agent (UA) URIs (GRUU) in the Session Initiation Protocol (SIP),” draft-ietf-sip-gruu-04 (work in progress), July 2005.
[I-D.ietf-sipping-config-framework] Petrie, D., “A Framework for Session Initiation Protocol User Agent Profile Delivery,” draft-ietf-sipping-config-framework-08 (work in progress), Mar 2006.
[I-D.rosenberg-sip-route-construct] Rosenberg, J., “Clarifying Construction of the Route Header Field in the Session Initiation Protocol (SIP),” draft-rosenberg-sip-route-construct-00 (work in progress), July 2005.
[RFC2104] Krawczyk, H., Bellare, M., and R. Canetti, “HMAC: Keyed-Hashing for Message Authentication,” RFC 2104, February 1997.
[RFC3188] Hakala, J., “Using National Bibliography Numbers as Uniform Resource Names,” RFC 3188, October 2001.
[RFC3548] Josefsson, S., “The Base16, Base32, and Base64 Data Encodings,” RFC 3548, July 2003.
[RFC3608] Willis, D. and B. Hoeneisen, “Session Initiation Protocol (SIP) Extension Header Field for Service Route Discovery During Registration,” RFC 3608, October 2003.


 TOC 

Authors' Addresses

  Cullen Jennings (editor)
  Cisco Systems
  170 West Tasman Drive
  Mailstop SJC-21/2
  San Jose, CA 95134
  USA
Phone:  +1 408 902-3341
Email:  fluffy@cisco.com
  
  Rohan Mahy (editor)
  Plantronics
  345 Encincal St
  Santa Cruz, CA 95060
  USA
Email:  rohan@ekabal.com


 TOC 

Intellectual Property Statement

Disclaimer of Validity

Copyright Statement

Acknowledgment