Monday, April 26, 2010

A meditation on SIP, SOA and Software

Today, the blog is thinking about SIP.

This was triggered by one of my customers turning to me yesterday and saying,

"SIP, isn't it just another of the IT industry's three letter acronyms for marketing? Just like SOA really, but even less likely to change things".

Now the blog has looked at SOA before (see posts like: "SOA - bringing CRM, telephony and business together? part 1") and the blog is a strong believer that SOA is a very major change in how IT is done. The downturn has perhaps slowed down the rate of SOA adoption, but nearly all the customers I work with are considering SOA approaches to at least some part of their IT environment.

For the communications industry, I think there is now little doubt of the impact of SIP. I suspect that SIP is going to bring with it a radical series of changes. To a certain extent (and to stretch an analogy), while SOA is the consequence of what client/sever did to the mainframe, SIP is the consequence of what IP has done to networking.

While I wouldn't defend this analogy beyond a certain point, I do think it highlights one interesting truth. The consequences of IP were to open up standards (much as client/server blew apart the vertically integrated architecture of the mainframe) and SIP just takes that to the next level. The impact of SIP may also be as disruptive as those changes were to some of today's business models.

The big thing for me about SIP is that it removes some of the last remaining geographic restrictions on call control. SIP trunking removes the last part of the TDM world, namely that lines had to terminate somewhere and there had to call control near it. To be sure, in the SIP world there still is a need for physical lines, but many of the physical dependencies on call management have gone.

For the call centre industry, this raises interesting possibilities. For example, Avaya have started to show how the use of a SIP session manager might allow them to virtualise ACDs without the application layer approach of Genesys or the network management approach of ICM. For Cisco, the rise of SIP represents a significant opportunity as services at the network layer (such as security) become increasingly important when using such a lightweight protocol. Also, SIP permits video as easily as voice, something that Cisco sees very much as the future. For other vendors, who haven't yet become so comfortable with IP, the rise of SIP represents a fundamental challenge.

Of course, an industry change tends to bring in new entrants and this is where it gets really interesting. I see SIP as ensuring that the future of the voice industry lies with software. That is a view some of the software firms share and is why so many have entered the voice market. I blogged this time last year on Microsoft and Google (See "The future of contact centre - Google, Salesforce, Skype & Microsoft"), but that was primarily from a CRM perspective.


Following VoiceCon this year (which I covered in this blog post), it's clear that so far Microsoft has the most advanced plans for voice of the software vendors. The enabler for this is SIP and OCS 14 leverages a very significant portion of its capability from what SIP enables. It's the capabilities of SIP that that provide OCS with its more interesting capabilities around presence, video and voice integration.

SIP has triggered a very interesting three-way fight. Previously separate areas (voice, data and desktop) are now different aspects of the same question. SIP brings into conflict the legacy voice vendors (with their communications expertise, such as Avaya), the network vendors (who have deep IP protocol vendors, such as Cisco) and the desktop/ software vendors (who understand presence and the desktop, such as Microsoft). It will be very interesting to see who can win this collision of different architectural layers.

No comments: