[Standards-JIG] Jingle vs. Zoep

dirk.griffioen@voipster.com dgriffioen at voipster.com
Thu Feb 9 16:49:44 UTC 2006


Jean-Louis Seguineau wrote:

>I don't feel that bridging XMPP/SIP through a gateway is that of an issue
>for the most common features of SIP (call or message sessions) and SIMPLE
>(presence) It requires a little more work than what is describe in Peter's
>draft ;)
>
>In comparing the values of these two protocols, we needs to take into
>account the very different contexts in which they have been designed.
>Because it was originally designed to establish voice sessions between
>peers, SIP goes to great length in discovering a remote party through
>sophisticated DNS queries, it allows a vast array of transports to be used,
>and therefore accommodate for the shortcomings of each of them. I possess
>well defined mechanisms to cater with network issues or user agent issues
>through a vast array of error messages, and well specified time based state
>machines.
>
>On the XMPP side, the context is simpler, as we are using TCP on the network
>layer, and we rely on presence to assess the availability of a remote party.
>This is somewhat simpler, as XMPP does not need to encompass all possible
>combination that SIP has to deal with. To make it short, in XMPP if two user
>agents see each other presence, we have gone a long way into being able to
>establish a media session. 
>
>The real need at this point is to actually negotiate the media session. And
>this is not really the realm of SIP, but rather of SDP. Which is what Jingle
>is attempting (I'm simplifying on purpose here) As XMPP is presence based,
>it becomes very easy for a server to 'fork' calls to different user agents
>based on priorities and/or RAP and achieve a result similar to that of SIP.
>In the end, XMPP already provides the signaling transport. And as nobody
>disagree on the use of RTP, we just need the media negotiation to provide a
>decent media session establishment though XMPP. An in my opinion, Jingle
>could do a better job at it than encapsulating SIP within XMPP (not to
>mention the inefficiency of the entire payload). Why, because it took two
>RFCs to the SIP workgroup to finalize the media call establishment. And this
>work has been done and documented. And XMPP only needs a part of it and
>learn from the past mistakes the SIP workgroup made. In that field, we can
>build a clean sheet system without having being concerned with a legacy
>(which SIP has to do). 
>
>For XMPP Jingle has a much greater potential than anything that was
>conceived for a different context. It was the case of TINS, and IMHO, I do
>not think transporting SIP inside XMPP solves anything but a short term
>practical problem: at the time Jingle did not existed and it was a sensible
>way to go to create a VOIP XMPP client.
>
>My $0.02
>
>Jean-Louis
>
>P.S. I agree with Peter that the two protocols are better off in their own
>world and bridged through gateways. After all networks are all but a bunch
>of gateways to bridge communication gaps...
>
>  
>
Thanks Jean-Louis!

Regarding your remark on the gateway, there is some comment on that my 
response to Akito, can I please direct you there?

I do agree that both protocols came from a different background,  and 
secondly that Jingle could learn from the SIP workgroup in what to 
avoid; but already there is a comment that Jingle is getting to big and 
is 'overstepping its boundaries'. Are you not afraid this will happen 
again? And we will end up with a second 250 pages+ protocol description 
with possible implementation hardship.

(I is not that I am very fond of SIP or anything, but it is out there, 
well-defined and well-used.)

Regards, Dirk

>-----Original Message-----
>
>Message: 1
>Date: Wed, 08 Feb 2006 14:51:53 -0700
>From: Peter Saint-Andre <stpeter at jabber.org>
>Subject: Re: [Standards-JIG] Jingle vs. Zoep
>To: Jabber protocol discussion list <standards-jig at jabber.org>
>Message-ID: <43EA67F9.4090306 at jabber.org>
>Content-Type: text/plain; charset="iso-8859-1"
>
>-----BEGIN PGP SIGNED MESSAGE-----
>Hash: SHA1
>
>dirk.griffioen at voipster.com wrote:
>
>  
>
>>I do feel there is some overlap
>>between Jingle and Zoep, so what I would be interested in is in
>>determining where and how big this overlap is. Jingle is without doubt
>>very well thought out, but there is value in SIP too, if only for
>>legacy, connectivity or wider acclaim: now XMPP does SIP too .. which
>>maybe brings in parties formerly not aware of or not interested in
>>XMPP's possibilities.
>>    
>>
>
>It's amusing to see people arguing for the benefits of SIP on this list
>("I realize XMPP rocks, but don't forget, there is value in SIP too"!).
>I'm not being critical at all, but it's not something we see every day
>here. :-)
>
>I agree with you that we need to determine what the overlap is between
>Jingle and Zoep. Perhaps it would help for you to define the use cases
>that Zoep is addressing. Is this something that two XMPP clients would
>use to set up a voice chat or other multimedia session? Is it something
>that an XMPP client would use only when communicating with a SIP gateway?
>
>There are many ways that SIP and XMPP can interoperate. One way is
>through gateways. I've defined a spec for that at the IM and presence
>level (draft-saintandre-xmpp-simple) and I envision writing something
>like that for multimedia sessions as well once the Jingle spec is a
>little more stable. The gateway approach may seem ugly, but it is a
>workable model (we've been doing gateways in the Jabber world since
>1999). Better, I think, to let each protocol go its own way and interop
>through gateways than to mix the two in ways that no one ever intended.
>
>But I'm open to argument. :-)
>
>Peter
>
>
>
>  
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.jabber.org/pipermail/standards/attachments/20060209/37ae7630/attachment.html>


More information about the Standards mailing list