Samuel - take a look at the recently submitted:-

http://www.ietf.org/internet-drafts/draft-gurbani-sipping-connection-gui
delines-00.txt

This references a proposed keep-alive mechanism that will be appearing
in the archives after IETF 60.

Chris.


-----Original Message-----
From: Samuel Osorio Calvo [mailto:[EMAIL PROTECTED] 
Sent: 22 July 2004 09:27
To: [EMAIL PROTECTED]; [EMAIL PROTECTED]
Subject: RE: [Sip-implementors] keep-alive mechanism

Thanks a lot!!! 

I'll take a look at ICE/STUN option. 

Just if you remeber, can you please point me out where Jonathan's
comments regarding keep-alive comments are (date?)?

Is there any method which is closer to SIP standard?? or just
recommended mechanism??? I've been looking in RFC/drafts and see no
comments.

Regards,

Samuel
Unclassified


>>> <[EMAIL PROTECTED]> 07/22/04 06:12AM >>>

Hi Samuel,

Comments inclined.

HTH,

Regards,
Sreeram

-----Original Message-----
From: [EMAIL PROTECTED] 
[mailto:[EMAIL PROTECTED] On Behalf Of Samuel
Osorio Calvo
Sent: Tuesday, July 20, 2004 3:32 PM
To: [EMAIL PROTECTED] 
Subject: [Sip-implementors] keep-alive mechanism


Hi all!

Since SIP does not provide keep-alive mechanism, what is the best
approach to detect end-point failure within a reasonable time?

A) Session timers negotiation is useful for stateful proxies but not
for
UA-UA since it would take a long time to detect that an end-point has
crashed.
It is recommended, as far as I know, to use lack of RTP or RTCP data
to
detect end-point failure. But this may not be appropriate for some
scenarios, like for instance where end-points are muted by a third
element, or in large conferences with high RTCP retransmission rate
(not
to mention some real applications without RTCP part). How to deal with
ICMP errors (UDP) or TCP connection shutdown cases? Are they the best
approach to detect end-point failure (it will , though, only allow
detecting whenever sending a message)? Is it then appropriate to
developn an out-band mechanism? Or using SIP INFO or OPTIONS methods?

[Sreeram] 1.using OPTIONS is a good option.
                  When you observe that signaling path is working fine
and still media is down, to re-pair
                  send re-invite or update.
                   Using re-invite if you observe that there is no
change in the media description, and still media path is not
established
you can try with different path.
               3 (As jonathan pointed out some time back in the list).
                 another option is using ICE, which is using STUN to
do
connectivity checks

B) Imagine a centralised conference where focus and mixer are not in
the
same host. How does participants know that the focus has crashed?
Moreover, how to deal with the case of end-points and mixer working
but
no focus running? Shall the ens-points detect it as fast as possible
and
close the RTP streams?

[Sreeram] After  media path establishment ,if the signaling path is
crashed and the session timer did not expire on the either side
               than "how can the UA's know that the signaling path is
gone..!!!
               So same explanation as in case 1 should be used to
detect
this.
               Once it knows that the signaling path is gone!! It
should
send a bye and try to close the session.
               (or) send a invite or update to repair the connection

There is only a signalling relationship betweent the focus and
participants. Thus: Using the RTP/RTCP approach, does end-points send
a
BYE to the focus when they detect that the mixer has crashed??

Thank you. Comments are higly appreciated,

Samuel.
Unclassified
_______________________________________________
Sip-implementors mailing list
[EMAIL PROTECTED] 
http://lists.cs.columbia.edu/mailman/listinfo/sip-implementors 




Confidentiality Notice

The information contained in this electronic message and any
attachments to this message are intended
for the exclusive use of the addressee(s) and may contain confidential
or privileged information. If
you are not the intended recipient, please notify the sender at Wipro
or [EMAIL PROTECTED] immediately
and destroy all copies of this message and any attachments.
_______________________________________________
Sip-implementors mailing list
[EMAIL PROTECTED]
http://lists.cs.columbia.edu/mailman/listinfo/sip-implementors


This message has been scanned for viruses by MailControl - www.mailcontrol.com

_______________________________________________
Sip-implementors mailing list
[EMAIL PROTECTED]
http://lists.cs.columbia.edu/mailman/listinfo/sip-implementors

Reply via email to