Google and IBM pushing on Jabber: should we forget SIMPLE?
When GTalk has been launched I believed it will migrate entirely to SIP/SIMPLE. It was also announced in the Developer page:
Just to make an example, Gizmo Project uses SIP to make calls, and Jabber to handle the contact list and presence status.
Technorati Tags: Giacomo Vacca, Presence, Google, GTalk, IBM, Lotus Sametime, Cisco, Microsoft LCS
4. Do you plan to support other real-time communication protocols?After this announcement from Google(that simply declares compatibility between GTalk and Lotus Sametime 7.5), I'm less optimistic, and I look forward to see a harder work for the integration of SIP (voice) services with XMPP/Jabber (IM and presence) services... Also because both Cisco and Microsoft (with Live Communication Server) are pushing "their" SIP, vanishing at least partially the effort sustained during the standardization process.Google Talk supports XMPP with the beta release. We plan to support SIP in a future release. Additionally, we will evaluate other protocols as appropriate, to continue to deliver on our commitment to open communications.
Just to make an example, Gizmo Project uses SIP to make calls, and Jabber to handle the contact list and presence status.
Technorati Tags: Giacomo Vacca, Presence, Google, GTalk, IBM, Lotus Sametime, Cisco, Microsoft LCS
Labels: Cisco, Giacomo Vacca, Gizmo Project, Google, GTalk, IBM, Lotus Sametime, Microsoft LCS, Presence
3 Comments:
I think SIMPLE is not adequat for presence and IM.
It is really primitive compared to XMPP and hardly interoperable from one version to another.
SIP is good for voip however and much more advanced than Jingle (XMPP counterpart).
I think that's what Google means: They will supports SIP for VOIP interoperability, but never SIMPLE, which is not at all a good IM and presence standard.
--
Mickael Rémond
Process-one
Hi Michael,
thank you for your comment. I partially agree with you regarding the limitation of SIMPLE compared to XMPP for IM and presence.
My point is that it's a shame that session signaling (for voice, video, etc) is not converging to presence and instant messaging signaling with the same, standard, protocol.
This is due by a lack in SIP as a protocol, but there is also the responsibility of software vendors which produce SIP stacks that "should work" or with "minimal modifications".
If you know XLite, you can see an example of how SIMPLE does work.
Now Google and IBM are marking a direction, and probably SIMPLE will encounter even more resistance... and this is - in my opinion - a bad trend for the VoIP industry.
Best Regards,
Giacomo
Unfortunately Giacomo, IT is full of missed opportunities like this…
When the XMPP workgroup was created at the IETF, there was a tentative toward narrowing the gap between XMPP and SIMPLE that took the name UMPP (where U stands for “universal”). The idea at the time (end September 2003) was to use SIP to negotiate an XMPP session between end-points. I should say that this attempt failed because of the stubbornness of the SIMPLE wg participants who insisted to have SIMPLE play the presence part. That was a mistake, and we have the demonstration today: SIMPLE is simply not ready.
There are also a few details that did not really played in favor of SIMPLE. The first and rather important one being the geographical concentration of SIP/SIMPLE expertise. As SIP developments have been in vast majority outsourced by the bug telco vendors to India, SIMPLE expertise is naturally found in that country.
The second aspect is the complexity of SIMPLE which makes it difficult to grasp by the average open source developer outside those versed in “voice over…”. When one add the rather rigid framework provided by SIP/SIMPLE as a protocol and the vast differences in interpretations by different implementers, it makes it “elitist: compared to XMPP.
Having been in the business of designing converged systems supporting both protocols simultaneously, I can vouch that any application programmer can grasp how XMPP works, the same is not true from SIP/SIMPLE.
To your last point about XLite, I have a great respect for the development team at Countepath. eyebeam is probably the most “standard” SIMPLE client out there, and we worked closely at a time when testing for IMS demonstration. But they had to come up with new builds every week in order to accommodate the ever changing scope of XCAP, and the various lists management that have to be part of a “workable” SIMPLE solution. And while doing this, they had no time to automatize the status change to “on the phone” when a call was taking place…
As Ken Camp put it, simplicity will make the winning combination. This is not the first quality of SIMPLE. As a result, it has bared its mainstream adoption by web application developers. But, I don’t see it as a “bad trend”, just a refocus for a better outcome. That said, it will have a cost because of investments that have been made…
Jean-Louis
Post a Comment
Subscribe to Post Comments [Atom]
<< Home