[reSIProcate] UA "486 Busy Here" & ACK in case of SIP-Proxy

[Clemens Fischer (VI/SEA)]

Thnx for your answer,

you are right, as the ACK needs to be sent to the same address & port to which the original request was sent to --> this seems to be a problem with the SIP proxy I'm using. I should definitely try another SIP proxy ...

Regards
// Clemens


-----Original Message-----
From: Fischl jason [mailto:jason.fischl at gmail.com] 
Sent: Montag, 07. März 2005 17:30
To: Clemens Fischer (VI/SEA)
Cc: resiprocate-devel at list.sipfoundry.org
Subject: Re: [reSIProcate] UA "486 Busy Here" & ACK in case of SIP-Proxy

An ACK to a failure response is handled hop by hop rather than end to end. Look at section 17.1.1 (INVITE Client Transaction).  In your demo UA it sounds like the behavior is correct. Your demo UA will not see the ACK message in the application since the ACK will be received by the transaction layer of the stack.

Jason


On Mon, 7 Mar 2005 09:40:50 +0100, Clemens Fischer (VI/SEA) <clemens.fischer at ericsson.com> wrote:
>  
> Hi reSIProcate gurus,
>   
> I've implemented a small (demo)UA and found a problem I couldn't come 
> around. My problem is that if (via proxy) my UA e.g. calls itself, my 
> UA will respond with 486 (busy here) and the ACK is sent towards the 
> proxy (which will correctly skip the ACK). According to my knowledge 
> the ACK sould be addressed directly to my UA ..
>   
> UA (x.x.111.35) Proxy             UA (x.x.111.35) 
> | INVITE          |                | 
> | --------------> | Invite         | 
> |                 | -------------> | 
> |                 | 486            | 
> | 486             | <------------- | 
> | <-------------- |                | 
>   
> ["is"]
>   
> | ACK             |                | 
> | --------------> |                |
>  
> |                 | 486(retrans)   | 
> | 486(retrans)    | <------------- | 
> | <-------------- |                | 
> etc. 
>   
> ["should???"]
>   
>  
> | ACK             |                | 
> | -------------------------------> | 
>   
> I've alread scanned all threads but couldn't find any related issue(s) ... 
>