Re: [Sip-implementors] Is session-expires be sent in 2xx response even though UAS doesnt support session timer?

2007-12-13 Thread Raj Jain
> But that is just politeness. The UAC is still *allowed* to send > reinvites and updates whenever it wants. Agreed. I can only HOPE (can we get that keyword added in RFC 2119 :-) that UAC implementers will understand the importance of negotiation as opposed to just coding to what is "allowed". R

Re: [Sip-implementors] Is session-expires be sent in 2xx response even though UAS doesnt support session timer?

2007-12-13 Thread Paul Kyzivat
I guess if you *want* to negotiate this then you can use this mechanism to do it. But the result is that you won't negotiate to do none. I guess that is ok if you intend to do it anyway. But that is just politeness. The UAC is still *allowed* to send reinvites and updates whenever it wants.

Re: [Sip-implementors] Is session-expires be sent in 2xx response even though UAS doesnt support session timer?

2007-12-13 Thread Raj Jain
That's a good example. I guess the fundamental issue w/ doing end-to-end things such as refreshes w/o negotiating is that it's fundamentally one-sided. Let's say a UA (for whatever reason) decides to do session refreshes at every minute and as a result swamps the proxies and UAS. If this was done

Re: [Sip-implementors] Is session-expires be sent in 2xx response even though UAS doesnt support session timer?

2007-12-13 Thread Paul Kyzivat
If you really really intend to refresh the session in any case, and want to negotiate the interval, then of course you can indeed send the session-expires in the initial invite. But if nobody but you wants to do it then you will then be bound by your own offer. There are some more sophisticated

Re: [Sip-implementors] Is session-expires be sent in 2xx response even though UAS doesnt support session timer?

2007-12-13 Thread Raj Jain
> The only value of including Session-Expires would be to put a *maximum* > on the refresh interval. The logic for doing that is as questionable as > requesting it at all. That's interesting. Let's say the UAC would like the session be refreshed no longer than every 5 minutes. If it doesn't includ

Re: [Sip-implementors] Is session-expires be sent in 2xx response even though UAS doesnt support session timer?

2007-12-13 Thread Paul Kyzivat
Raj Jain wrote: > Paul, > > I agree w/ what you said about the actual purpose of session-timers. > However, I'm wondering about your following statement: > >> So, while it is *legal* for a UAC to include the Session-Expires there >> is no reason to do so. > > If the UAC does not include Sessio

Re: [Sip-implementors] Is session-expires be sent in 2xx response even though UAS doesnt support session timer?

2007-12-13 Thread Raj Jain
Paul, I agree w/ what you said about the actual purpose of session-timers. However, I'm wondering about your following statement: > So, while it is *legal* for a UAC to include the Session-Expires there > is no reason to do so. If the UAC does not include Session-Expires: and Min-SE: in its INVI

Re: [Sip-implementors] Is session-expires be sent in 2xx response even though UAS doesnt support session timer?

2007-12-12 Thread Paul Kyzivat
It seems I have to say this every few weeks... *WHY* is the UAC requesting a session timer??? The real purpose of session timer is so that a record-routed proxy can cause one of the UAs to send periodic messages - confirming for the proxy that the session is still active. (A proxy isn't allowed

[Sip-implementors] Is session-expires be sent in 2xx response even though UAS doesnt support session timer?

2007-12-12 Thread praveen dandin
Hi all, I have a query regarding session timer functionality. UAC sends the INVITE message with Session-Expires header. The UAS does not support session timer fuctionality. Now possible behaviour of UAS are: 1) UAS copies the value of Session-Expires header value into 2xx and sends the respon