Re: [sipcore] SIP Timer H and Timer B must have the same value ???

Adam Roach <adam@nostrum.com> Tue, 25 January 2011 14:12 UTC

Message-ID: <4D3EDAEA.7050403@nostrum.com>
Date: Tue, 25 Jan 2011 08:15:06 -0600
From: Adam Roach <adam@nostrum.com>
User-Agent: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.6; en-US; rv:1.9.2.13) Gecko/20101207 Thunderbird/3.1.7
MIME-Version: 1.0
To: mohamed.ati@orange-ftgroup.com
References: <F45661E8FBC74F4EB7E1E0386B562A7501D7FF76@ftrdmel0.rd.francetelecom.fr>
In-Reply-To: <F45661E8FBC74F4EB7E1E0386B562A7501D7FF76@ftrdmel0.rd.francetelecom.fr>
Content-Type: multipart/alternative; boundary="------------030101080501070909050606"
Received-SPF: pass (nostrum.com: 99.152.144.32 is authenticated by a trusted mechanism)
Cc: sipcore@ietf.org
Subject: Re: [sipcore] SIP Timer H and Timer B must have the same value ???
Precedence: list

[as participant]

On 1/25/11 06:39, Jan 25, mohamed.ati@orange-ftgroup.com wrote:
>
> I have a probleme understanding the following sentence RFC 3261
> regarding H and B Timers in section 17.2.1.
>
> "/Timer H determines when the server transaction abandons
> retransmitting the response. //Its value is chosen to equal Timer B,
> the amount of time a client transaction will continue to retry sending
> a request./"//
>
> Is this means that Timer H (in the server) and Timer B (in the client)
> must have the same value at both sides of a SIP hop ?
>
> If yes, why is this necessary and what are the impacts of not
> respecting this rule ? In fact, in some times this rule could not be
> satisfied: timers H and B are multiple of Timer T1 and this later may
> different at both sides of the SIP hop which is possible as the SIP
> hop node may have different types of connections at both sides (e.g.
> radio / Fiber). See figure below.
>

Actually, it's never a good idea to have differing values for T1 between
a client and a server. Robert Sparks has put together a good writeup of
the related issues here:

http://blog.tekelec.com/blog/bid/13564/Changing-SIP-s-Transaction-Timers-Part-1-of-2
http://blog.tekelec.com/blog/bid/13566/Changing-SIP-s-Transaction-Timers-Part-2-of-2

[as chair]

In the future, please address this kind of question to the
sip-implementors mailing list. See
https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors

/a

Re: [sipcore] SIP Timer H and Timer B must have t… mohamed.ati
Re: [sipcore] SIP Timer H and Timer B must have t… Adam Roach