We are pleased to announce that we have launched support for Jingle
XEP-166 and XEP-167 for Google Talk calls to and from Gmail, iGoogle,
and Orkut. We have also added the same level of support to libjingle
(http://code.google.com/p/libjingle), which is used by many native
clients. From this point on, it will be our primary signalling
protocol, and the old protocol will only remain for backwards
compatibility. We also plan to soon update Google Talk on Android to
speak Jingle, but we do not plan on updating the Google Talk Windows
We suggest all clients that interop with Google Talk to switch to
using Jingle rather than the old protocol. We will remain backwards
compatible with legacy clients by continuing to speak the old protocol
as well. If you wish to continue working with legacy clients, such as
the Google Talk application for Windows, you may also wish to continue
speaking the old protocol. But the future is Jingle, and the old
protocol will eventually go away.
Finally, we are still working on implementing XEP-176 (ICE-UDP). In
the meantime, you'll need to use our draft-06 version of ICE, which is
implemented both in libjingle and in libnice, two open source
I hope that this will be a support to the Jingle community and futher
our efforts to have open standards for voice and video communication.
First off, I'm super proud of these guys finishing off this work. It was a long time coming.
As for why we chose XMPP over SIP: The team had made the choice to base the core IM product on XMPP before I joined. That decision was based on the ease of implementation and interoperability. XMPP for IM is much simpler than SIP/SIMPLE. The easiest answer here is that it was the most logical fit to the product we had built at the time.
My personal opinion of SIP is based on a 6 year old snapshot of the technology. Things have undoubtedly changed in the meantime. Based on that snapshot, I wasn't a fan at the time.
- It is an amazingly complex and confusing technology. It is worth noting that SIP is the longest IETF RFC out there. Meanwhile, the core XMPP spec is easy to understand and implement. It is also an IETF standard. Extensions are well modularized and documented.
- The security is hacked on after the fact and, at the time, was unevenly implemented. In basic SIP, one unauthenticated UDP packet makes a phone ring. XMPP security is based on TLS and DNS and is a required part of the core.
- Federation was a patchy mess and showed SIP's telecom roots. You pretty much had to get the lawyers involved to do anything. Federation in XMPP is a DNS SRV lookup.
The world has moved forward since then so hopefully this provides some context for the decision at the time.
And now this news indicates that it appears to be AbandonWare.
Which is a real pity.