Re: [Sip-implementors] Retry intervals

2008-12-31 Thread Dale Worley
On Wed, 2008-12-31 at 15:54 -0800, Maxim Sobolev wrote: > I believe you are using the wrong tool to implement the required > functionality. You should really use application level timeouts to do > failover (i.e. cancel call in one UAC and make a new UAC instance to > alternative destination) ins

Re: [Sip-implementors] Retry intervals

2008-12-31 Thread Maxim Sobolev
Daler Worley wrote: > How should we handle retry intervals for resending SIP requests which > have received no response? RFC 3261 prescribes doubling the retry > interval each time, starting with an interval (T1) which is an estimate > of the round-trip time of the network, and ending with 32 time

Re: [Sip-implementors] BYE in Confirmed/Moratorium state (RFC 5407)

2008-12-31 Thread Maxim Sobolev
Dale Worley wrote: > On Mon, 2008-12-29 at 15:35 -0800, Maxim Sobolev wrote: >> Dale Worley wrote: >>> Looking at the figure, it appears that Confirmed/Moratorium is allowed >>> to receive and act on BYE. This seems to be an error to me, but even if >> I don't think it's an error. RFC3261 also all

[Sip-implementors] Retry intervals

2008-12-31 Thread Dale Worley
How should we handle retry intervals for resending SIP requests which have received no response? RFC 3261 prescribes doubling the retry interval each time, starting with an interval (T1) which is an estimate of the round-trip time of the network, and ending with 32 times that, for 6 retries total.

Re: [Sip-implementors] How to determine where to send RTP packet in multi-proxy SIP network

2008-12-31 Thread Dale Worley
On Tue, 2008-12-30 at 22:37 +0800, Serbang, Nabam (Nabam) wrote: > So how u1 knows where to send RTP packet. Usually it will send to > address present in Contact header. U1 will *never* use the IP address in the Contact header to determine where to send RTP packets, it will use the address in the

Re: [Sip-implementors] Chaos in Dialog subscription "standar"

2008-12-31 Thread Dale Worley
On Wed, 2008-12-31 at 00:34 +0100, IƱaki Baz Castillo wrote: > But what I see is the fact that each vendor implements "dialog" subscription > in a propietary way. For example, when configuring a Li phone for dialog > subscription it requires to choose between subscription server type > (Ast*

Re: [Sip-implementors] BYE in Confirmed/Moratorium state (RFC 5407)

2008-12-31 Thread Dale Worley
On Mon, 2008-12-29 at 15:35 -0800, Maxim Sobolev wrote: > Dale Worley wrote: > > Looking at the figure, it appears that Confirmed/Moratorium is allowed > > to receive and act on BYE. This seems to be an error to me, but even if > > I don't think it's an error. RFC3261 also allows BYE on early dia

Re: [Sip-implementors] Chaos in Dialog subscription "standar"

2008-12-31 Thread Christian Stredicke
"buttons" are about documents. In principle you could send it via HTTP/AJAX (which is IMHO a real alternative) or you can send it in a instant message. We don't like SUBSCRIBE because then there are so many dialogs on the user agent. But yes, of course you could also use a SUBSCRIBE dialog to ca

Re: [Sip-implementors] Chaos in Dialog subscription "standar"

2008-12-31 Thread Daniel-Constantin Mierla
Hello, On 12/31/08 12:09, Christian Stredicke wrote: > It is even worse. There are features that are simply not possible with dialog > state. For example, think about a simple DND button that you want to program > on your phone or just a speed dial. > > I guess dialog state was not designed for

Re: [Sip-implementors] Chaos in Dialog subscription "standar"

2008-12-31 Thread Christian Stredicke
It is even worse. There are features that are simply not possible with dialog state. For example, think about a simple DND button that you want to program on your phone or just a speed dial. I guess dialog state was not designed for the use on phones with buttons and LED. I regret and apologiz