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

List:       sip
Subject:    Re: [SIP] How Expire field works  in this case ?
From:       Henning Schulzrinne <schulzrinne () cs ! columbia ! edu>
Date:       2000-06-24 17:04:45
[Download RAW message or body]



> rafi wrote:
> 
> Hi,
> 
>                  Consider a call set up between  UAC and
>      UAS  passing through 'n' number of proxies. When
>     UAC  generated INVITE message it used   Expire
>     field with time  't'.  All the intermediate proxies  upon
> receiving
>     INVITE message decided  to use time 't' in order to limit their
>     search. When  INVITE finally arrives at UAS (assuming user is
>    presently unavailable),   it decides to wait for time duration 't'.
>     Now, before timer of duration 't' really expires  at UAS, timer at
> proxy (first proxy
>     in the call setup) expires first resulting in call being released.
> Thus call
>     MAY be released  much earlier than call waiting time promised at
> UAS.
>     Thus, how  should  UAS use time 't' present in INVITE request when
>     user is  currently not available?.
> 

The timer is not suitable for microsecond timing. If forwarding your
call takes more than a second, something is seriously broken (and your
call throughput is going to be pretty low). Thus, the window where the
call gets terminated by the first proxy rather than the UAS is going to
be measured in units of milliseconds. In addition, even if the first
proxy cancels the call, the total search time is still the expiration
time, so this is correct behavior. The phone will stop ringing either
way. (This can happen in any event if a proxy has a shorter
pre-configured patience than the UAS.)



> 
>        Second question is how does UA handles the
> case of INVITE crossing each other for a given pair of
> user ?.  Is this decision is left to user of SIP stack ?

This question has been discussed several times on the list, at great
length. Please take a look at the archives.

> 
> 
> Thank You
> Rafi Assadi H.M.
> Silicon Automation Systems
> Bangalore,  INDIA
> 
>


_______________________________________________
SIP mailing list
SIP@lists.bell-labs.com
http://lists.bell-labs.com/mailman/listinfo/sip

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