Re: stateful proxy

"Adam B. Roach" <Adam.Roach@Ericsson.com> Mon, 16 August 1999 16:48 UTC

Message-Id: <199908161647.LAA09832@b04a24.exu.ericsson.se>
Subject: Re: stateful proxy
To: akhuntia@ssd.usa.alcatel.com
Date: Mon, 16 Aug 1999 11:47:35 -0500
Cc: confctrl@ISI.EDU
In-Reply-To: <199908161415.JAA04256@sun179.ssd.usa.alcatel.com> from "Ashok Khuntia" at Aug 16, 99 09:15:37 am
From: "Adam B. Roach" <Adam.Roach@Ericsson.com>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Sender: owner-confctrl@zephyr.isi.edu
Precedence: bulk

>I got a question. Does a stateful proxy server keep CALL STATE 
>information or it keeps only the transaction state information? 
>If it keeps the call state information, where can I get more 
>information on that?

There is (understandably) a lot of confusion about this topic. As
defined in RFC2543, a stateful proxy keeps *transaction* state
information. (Read the first and last paragraphs of section 12.3 
carefully).

We've been calling the variety of proxy that keeps *call* state
information "call stateful proxies."

Exactly how much state you keep, and how long you keep it, is heavily
dependant on your particular application. For most of the applications
I have been examining (largely dealing with value added services), the
proxies will probably want to keep the same sort of information as a 
UAC/UAS, for the entire duration of the call.

To support this type of state in the proxy, there is a proposed extension
(draft-ietf-mmusic-sip-session-timer-02.txt); its purpose is to allow
proxies to free resources if the call terminates abnormally (e.g. without
the proxy being told).

--
Adam Roach, Ericsson Inc. |  Ph: +1 972 583 7594 | 1010 E. Arapaho, MS L-04
adam.roach@ericsson.com   | Fax: +1 972 669 0154 | Richardson, TX 75081 USA

stateful proxy Ashok Khuntia
Re: stateful proxy Adam B. Roach