[prev in list] [next in list] [prev in thread] [next in thread] 

List:       sip-implementors
Subject:    Re: [Sip-implementors] retransmission detection and stateless
From:       Roman Shpount <roman () atelo ! com>
Date:       2005-06-29 23:23:26
Message-ID: s1bt42c32d6fu47a () atelo ! com
[Download RAW message or body]

Jeroen van Bemmel wrote: 
> Hi all, 
> 
> While testing my SIP code I encountered the following sequence of events: 
> 1. UAC sends INVITE 
> 2. UAS creates a server transaction and sends 100, 180[tag] which 
> creates a dialog, 200[tag], ST is destroyed again 
> 3. Before the 100 reaches the UAC it retransmits its INVITE, which 
> arrives after the ST is destroyed 
> 4. The INVITE does not match any existing transaction, so it is 
> presented to the UAS core 
> 5. The UAS core forwards it to the application which again responds with 
> 100, 180[tag] but when the dialog is created it is found to already exist 
[...] 
> Anyone encountered something similar before? Does anyone see any 
> problems with such stateless to-tags? 

This is simular to the problem I encountered last year. See the foolowing for \
discussion:

http://lists.cs.columbia.edu/pipermail/sip-implementors/2004-August/007030.html
(or to get all the related messages search for "Problem with an intial SIP INVITE \
servertransaction" in \
http://lists.cs.columbia.edu/pipermail/sip-implementors/2004-August)

http://www1.ietf.org/mail-archive/web/sip/current/msg10326.html

The implementation I suggested was in operation for about a year now and seems to \
work well. ___________________________________
Roman Shpount, VP of Technology
aTelo, Inc. -- www.atelo.com


[prev in list] [next in list] [prev in thread] [next in thread]