Assured: The Isode Blog

Isode's XMPP Server: M-Link

Anyone paying any attention to the whitepapers we've published and the blog posts we've written over the last few months will not be surprised to learn that we're developing an XMPP server of our own, the first version of which will be launched with our next major product release, 14.2.

The reasons for developing an instant messaging and presence server (and basing it on the XMPP Open Standard) are ones we've discussed before but in short - we've always been an Open Standards company and we believe that instant messaging and presence has applications across many of our solutions areas  as well as helping us communicate with the partners and customers on whom we rely.

M-Link moved into customer beta a little while ago and we offered customers three ways of helping us test the product:

1. Download M-Link, set up M-link on their own domain using our evaluation guide, and link to Isode staff who are using the copy of M-Link running at isode.com (as well as anyone else with an IM account at other XMPP services like GMail or jabber.org).
2. Request an IM account at a service that Isode are running on 'friends.isode.com' and use that account to communicate with Isode staff and other XMPP accounts.
3. Use their existing XMPP service (if they have one) to communicate with Isode staff using M-Link.

As well as testing the product, the documentation and interoperability, customers have also been able to use the beta as a way of improving communications with Isode, the beta test is enabling us to both look at the product (M-Link) and the function (IM as it impacts our business).

We've also been able to properly test some of the many XMPP clients available for the various operating systems and platforms we find ourselves using. We'll be posting our opinions of some of those clients in a series of posts running up to the launch of 14.2

SIP and XMPP - Convergence or Co-Existence?

SIP is the widely deployed Internet Standard Session Initiation Protocol (RFC 3621).  XMPP is the Internet Standard eXtensible Messaging and Presence Protocol. There are substantial architectural similarities between services of these types, and advanced clients need presence, messaging, and session initiation capabilities. Integrating these services makes some sense, and two key attempts have been made.

SIMPLE is a family of protocols that builds presence and messaging over SIP. This layering has resulted in a specification of substantial complexity and rather poor performance, which suggests that this layering is not a good choice.  These technical issues are a key factor as to why XMPP is winning over SIMPLE as the standard of choice for presence and instant messaging.

JINGLE is a session initiation protocol built over XMPP.  This layering works better, and it is clear that JINGLE will have an important role in providing session initiation between clients that support XMPP. It is a clean and effective approach to help XMPP clients support voice and video within a protocol and security framework that is already implemented.

However, JINGLE does not provide a complete replacement for SIP.

A SIP client implementation is a very straightforward mechanism to set up a phone call, and is supported on a large number of phones. If call connect is all that is needed,  SIP is ideal. JINGLE only makes sense if XMPP is needed for other reasons.  VOIP phones will continue to be SIP based, and VOIP deployments will be based around PABX systems that include SIP servers.

Another reason for this is that JINGLE does not provide a framework for phone type services such as call forwarding, putting calls on hold and voicemail integration. The underlying reason is that JINGLE is based on communication starting out with (XMPP) presence, which is a model of user to user communication, as distinct from the more traditional model underlying SIP of device to device communication.   

JINGLE use will grow, based on voice and video services in the context of XMPP clients. SIP use will continue to grow in the context of VOIP phone deployment. There will be co-existence, and gateways between SIP and XMPP/JINGLE.

JINGLE Implementation Growing

JINGLE is a protocol developed by Google and the XSF (XMPP Standards Foundation) to support VOIP (Voice over IP) Communication, and it is used in Google Talk.  We are seeing increased implementation of JINGLE (Wikipedia lists eleven implementations, including Google Talk) and we are aware of
more.

XMPP is the Internet Standard eXtensible Messaging and Presence Protocol. Many XMPP clients now offer a "call" button (typically with a green phone icon) for each buddy (the XMPP roster) that allows direct voice communication between XMPP users.   This integration is a useful capability. Voice communication goes directly client to client using the real time VOIP (Voice over IP) protocol.

In order for this to work, the VOIP connection needs to be established, and this needs a session initiation protocol, to communicate the desire to communicate and then negotiate the details of the VOIP connection to be used.  JINGLE, which reflects the ringing of a telephone, is the XMPP protocol used to achieve this.

JINGLE is a general purpose session initiation protocol, and can be used to negotiate connections for other services such as video, and can also set up multi-party conferences.

JINGLE is an important standard that seems certain to grow with XMPP and be used alongside XMPP, to support voice and video for XMPP Clients.   Voice calls from such clients can be gatewayed to other VOIP services, mobile and landline.  GTALK-TO-VOIP is an example of such a service.

The XMPP/JINGLE/VOIP combination of open standards, gives a clear platform for solutions to compete against Skype, the (proprietary) market leading software phone system.

Smartphone manufacturers neglect email at their peril

The Web Worker Daily blog reports on a survey from webcredible that suggests that 33% of mobile phone users feel that email is still the most needed mobile utility. The blog post goes on to suggest that this is one of the reasons that (in the US at least) the iPhone quickly attained the status of second best selling 'smartphone', after the BlackBerry.

Both RIM and Apple have, in their different ways, taken email seriously as an application. They have both done their best to provide for the user an email interface that works well on these small form factor devices and have reaped the rewards, something that many other manufacturers seem to have neglected to do.

But there's more to a good user experience than a nice interface. Ensuring that the appropriate standards are used would be another step in the right direction, so that the user does not suffer from the poor performance one would get implementing protocols designed for desktop operation on devices that have limited storage and relatively low bandwidth.

Smartphone device manufacturers who decide to combine a decent user interface with the appropriate open standards for handling and delivering email (such as the LEMONADE profile that we've talked about previously) still have a chance to stake out a commanding market position.

Microsoft, Yahoo, AOL and XMPP

As I write, market reaction to the Microsoft/Yahoo deal seems less than enthusiastic with both companies stocks down. The reaction from some technology bloggers seems equally negative, although a lot more entertaining:

"It’s like tying the Titanic to the iceberg. It’d keep you from sinking just long enough to freeze to death."

..and from the ever-entertaining Fake Steve Jobs:

"It's like taking the two guys who finished second and third in a 100-yard dash and tying their legs together and asking for a rematch, believing that now they'll run faster."

With our own commitment to the XMPP instant messaging and presence protocol and being of a mind that any successes for the XMPP protocol are good for the entire community, one question that this merger raises here at Isode is what does this mean for the adoption of open standards instant messaging now that two of the giants of proprietary IM may soon to be one entity (at least 'soon' by mega-corporate-merger standards)?

MS and Yahoo announced an interworking of their two proprietary IM systems back in 2005 but neither has shown much interest in opening up to the wider world. So on the face of it, not good for XMPP penetration amongst a huge user base of existing IM users.

Adding together the IM user bases of Microsoft and Yahoo brings them to within striking distance of AOL, who still maintain a lead in active and overall users (in the western world at least).

Last week there was quite a bit of buzz around AOL's experiments with XMPP, I wonder if the merger will finally push AOL fully into the XMPP camp, not just for technical reasons but also to regain the initiative in the IM space?