Supporting Multiple Path Routing in the Session Initiation Protocol (SIP)
draft-worley-redundancy-response-04

Status of this Memo

This Internet-Draft is submitted to IETF in full conformance with the provisions of BCP 78 and 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 September 8, 2009.

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 in effect on the date of publication of this document (http://trustee.ietf.org/license-info). Please review these documents carefully, as they describe your rights and restrictions with respect to this document.

Abstract

An increasing number of SIP architectures implement multiple path routing (MPR), which is the providing of more than one path for a call to reach a destination user agent (UA). A typical example is a redundant pair of gateways from a SIP system to the PSTN. A call from the SIP system to the PSTN can pass through either gateway to ultimately reach the destination telephone. In order to gain the benefits of redundancy in case one of the gateways fails or reaches capacity, a proxy forks INVITEs serially to both gateways.

Unfortunately, if the call passes through one gateway but fails at the destination phone (e.g., ring-no-answer), the proxy will then fork the call to the other gateway, because the proxy has no way to know that the call failed at the destination phone rather than at the first gateway. The second fork will fail in the same way at the same destination phone. This annoys both the caller (because the call takes twice as long as it should before failing) and anyone within earshot of the destination phone. Similar failures plague any other SIP architecture where a request can reach a destination through multiple paths.

To gain the benefits of MPR without suffering from this problem, the proxy which forks a request onto the redundant paths needs to be able to determine if a fork that failed reached the destination UA and was rejected by the UA (and so an alternate path should not be tried), or if the fork failed before reaching the UA (and so an alternate path should be attempted). This document is to begin a discussion of strategies for making this determination.

Table of Contents

1.  Background
2.  Proposed Solutions
3.  Related Work
4.  Security Considerations
5.  Revision History
    5.1.  draft-worley-redundancy-response-00
    5.2.  Changes from draft-worley-redundancy-response-00 to draft-worley-redundancy-response-01
    5.3.  Changes from draft-worley-redundancy-response-01 to draft-worley-redundancy-response-02
    5.4.  Changes from draft-worley-redundancy-response-02 to draft-worley-redundancy-response-03
    5.5.  Changes from draft-worley-redundancy-response-03 to draft-worley-redundancy-response-04
6.  References
    6.1.  Normative References
    6.2.  Informative References
§  Author's Address

1.  Background

An increasing number of Session Initiation Protocol (SIP) [rfc3261] (Rosenberg, J., Schulzrinne, H., Camarillo, G., Johnston, A., Peterson, J., Sparks, R., Handley, M., and E. Schooler, “SIP: Session Initiation Protocol,” June 2002.) system architectures implement multiple path routing (MPR), the feature of providing more than one path for a call to reach a destination user agent (UA). (MPR is also called path redundancy (PR) or alternate path retry (APR).) Typical situations are:

• multiple gateway devices that connect a SIP network to the PSTN, such that if one gateway is occupied to capacity, calls should be routed to the next gateway
• a PSTN gateway device as fallback when SIP connection over the public Internet fails
• sending a request to multiple services that determine how to reach a destination, with an order of precedence among the services as to which service is to be used (e.g., connecting to an ENUM contact, falling back to a PSTN gateway)

From a protocol point of view, a proxy is presented with the situation where a request (typically an INVITE) should be serially forked to more than one SIP URI, and from an application-layer point of view, all of the URIs are expected to ultimately reach the same user agent (UA). Of course, if one fork of the request succeeds at the destination UA, the proxy should not attempt any further forks.

If the forked request failed and did not reach the destination UA, then in order implement MPR, the proxy must attempt the next fork. But if the request did reach the destination UA, and the UA returned a failure response, sending the request to the same UA via a different path is unlikely to yield success, and may even degrade the user experience. For example, if the first request was not accepted by the attending human (ring-no-answer), sending a second request to the UA by a different path, which will cause the UA to alert again, is not the desired behavior.

The purpose of this document is to open the discussion of methods by which the proxy can implement the desired behavior.

Two subordinate problems are contained in the main problem: One is how the proxy becomes aware that the forking of a request is a MPR situation. Another is the question of when two destinations are to be considered to be "the same", and so constitute an MPR set.

2.  Proposed Solutions

This section discusses proposed solutions to the problem.

One possible solution is to split the failure response codes into two groups, one of which is only generated by proxies, the other is only generated by UAs. Then the response code from a failed fork should identify whether the call reached a destination UA or not. Most current failure response codes fall into one or the other of these two classes, but several of them can be generated by either proxies or user agents. In addition, the rules for choosing the "best" response from among the set of responses from a set of forks must be adjusted to give preference to UA-generated responses, which causes difficulties in some situations where the UAC might be able to amend its request in ways that increase its chances of success.

Other possible solutions to this problem are solicited.

3.  Related Work

This section lists related work.

ITU-T Recommendation Q.850 (telephone call failure cause codes)[Q.850] (ITU-T, “Usage of cause and location in the Digital Subscriber Signalling System No. 1 and the Signalling System No. 7 ISDN User Part,” May 1998.)

The Q.850 failure cause codes used in SS7 provide a "location" field that distinguishes the network (caller's local network, callee's local network, transit network) in which the failure occurred. In addition the cause codes are segregated in such a way that it is clear whether the call reached the destination telephone before failing.

RFC 3326: The Reason Header Field for the Session Initiation Protocol (SIP)[rfc3326] (Schulzrinne, H., Oran, D., and G. Camarillo, “The Reason Header Field for the Session Initiation Protocol (SIP),” December 2002.)

The Reason header can carry a Q.850 failure cause code. However, there is no provision for carrying the Q.850 location value. Given that the cause codes are well-segregated regarding network errors vs. endpoint errors, the cause code is probably sufficient to regulate MPR.

The SIP SPECIFY Method[sreeram‑specify] (Kanumuri, S. and S. Sarkar, “The SIP SPECIFY Method,” April 2007.)

Abstract: This document proposes an extension to the Session Initiation Protocol (SIP). This extension adds the SPECIFY method to the SIP protocol. The intent of the SPECIFY method is to allow conveying SIP entity's (Proxy, User Agent, Redirect Server) intention of asynchronous service change (i.e. to be taken out of service or has just return to service) with and with out a backup SIP entity within the same administrative domain.

No Service To This Number Reject Code[schwartz‑nsr] (Schwartz, D., “No Service To This Number Reject Code,” November 2008.)

Abstract: This specification discusses a SIP response error code ambiguity that may exist due to semantic differences between SIP [RFC3261] and TEL [RFC3966] URIs.

4.  Security Considerations

Alternate path retry presents no security considerations that are known to the author beyond what is present in non-MPR SIP system architectures.

5.  Revision History

5.1.  draft-worley-redundancy-response-00

First version.

5.2.  Changes from draft-worley-redundancy-response-00 to draft-worley-redundancy-response-01

Add note that the new variant of 500 would be like "reorder" in the PSTN.

Add note that dealing with 402 can be postponed to when 402 is standardized.

Add section proposing how to split response codes.

5.3.  Changes from draft-worley-redundancy-response-01 to draft-worley-redundancy-response-02

Narrow the scope of the document to be a problem description.

5.4.  Changes from draft-worley-redundancy-response-02 to draft-worley-redundancy-response-03

Changes were minimal.

5.5.  Changes from draft-worley-redundancy-response-03 to draft-worley-redundancy-response-04

Added "Related Work" section.

Updated contact information.

Revised Abstract.

6.  References

6.1. Normative References

6.2. Informative References

Author's Address