Original URL: http://www.theregister.co.uk/2010/11/11/ws_i_battle_for_the_web/

Microsoft and IBM web-control war finally silenced

Google and Facebook join a fresh fight

By Gavin Clarke

Posted in Developer, 11th November 2010 12:58 GMT

One era in web politics has passed with the closure of the Web Services Interoperability (WS-I) organization just as another era has opened.

The WS-I said Wednesday after nearly 10 years it's winding down operations and merging its assets and maintenance of its specs to OASIS.

The group said the release of WS-I member approved final materials for Basic Profile (BP) 1.2 and 2.0, and Reliable Secure Profile (RSP) 1.0 fulfills its last milestone as an organization. "By publishing the final three profiles, WS-I marks the completion of its work," the group said.

Most will hardly have heard of the WS-I, yet the operation saw the last pitched battle by enterprise IT vendors to retain control a web quickly slipping out from their control.

WS-I was led by IBM and Microsoft but dressed up as an industry group. It was built so that IBM and Microsoft could retain control of their desktop and server platforms while ensuring these platforms' grew in an increasingly open world. SOAP, WSDL, UDDI and XML were the bridges between these worlds, and IBM and Microsoft positioned themselves as the architects of what those bridges should look like.

IBM and Microsoft in the early 2000s devised a roadmap of WS-* specs covering how they believed apps and data should talk and be exchanged on the web. The pair then set about filling in that roadmap. IBM and Microsoft would announce a spec with other, chosen vendors specializing in a given space - such as WS-Security with Verisign or WS-BPEL with BEA Systems, SAP and Siebel - and then submit it to a standards group, invariably OASIS. WS-I was not a standard group but provided profiles and tools to let people test their web-services implementation.

The IBM and Microsoft roadmap defined a modular and interoperable family of WS* specs, operating as a framework and at a component level.

The then still great Sun Microsystems had no place in this vision. Microsoft didn't care for Java and it saw no reason why anybody shouldn't be happy with .NET. For IBM, it was the dawn of a pissing match with Sun over who should become the de-facto leader of Java.

The WS-I was announced in February 2002 with an opening line up of 53 middleware, integration, tools and ERP vendors big and small, plus some major consumers of enterprise technology. Yes, this was truly a broad group united with a single goal. Just one thing was missing. Sun - the company that created Java, which was being used by millions of developers.

Why? It emerged Bill Gates, then Microsoft chairman, and said he could live with the WS-I if: "We have the positioning clearly in our favor. In particular, Sun not being one of the movers/announcers/founding members." At that time, Microsoft was not a great joiner of things that drove interoperability so participating in the WS-I was a major departure.

Gates' company wanted the WS-I to be constructed in such as way that it was as unpalatable to Sun as possible. That suited IBM as it tried to assert control over Java, and so IBM and Microsoft tried to carve up the web and exclude the only other company with enough of a vested interest in what was happening and the sheer size to stop them. Sun only clawed its way back in by submitting to elections to the WS-I board a year later in 2003 with tiny WebMethods. Yes, that was embarrassing. It was meant to be.

The WS-* specs from IBM and Microsoft suited the enterprisey world of those in the WS-I clubhouse. WS-* and the WS-I paved the way for the Service Oriented Architectures (SOAs) bubble, a cacophony of hype about a set of systems that could never be delivered but paid the wages of consultants and enterprise vendors, and involved some kind of choreography wrapped in a portal.

The world was not interested in heavy building blocs of specs and a few technologies ordained by a handful of vendors with a vested interest in ensuring their place at the table. Developers outside the WS-I picked up on HTTP, REST and JSON that allowed loosely coupled communications. These kick open the door on the consumer web of dynamic and scripted services and today's cloud frenzy. Inevitably they also started feeding back into the WS-I vendors' software.

As it closed down Tuesday, the WS-I remained frozen in that war between big systems vendors to re-gain control over the web. The leaders include Microsoft, IBM, Oracle and SAP. There was no Google, Amazon, Twitter or Facebook. Probably never even heard of it. WS-I has not only not moved with the times, the times have gone in a completely different direction.

The WS-I became irrelevant for other another reason: its members couldn't stop buying each other or succumbing to others - Microsoft bought Groove Networks, Hewlett-Packard Compaq, SAP Business Objects and Sybase, and Oracle claimed BEA, Oynx and - oh, yeah, Sun. A legion of enterprise application and data integration dots were also wiped off the board by other enterprise borgs - Iona Technologies, CapeClear and WebMethods.

Don't think the passing of the WS-I marks the end of big tech companies' attempts to control the web - the very crucible in which the WS-I was born. Far from it.

The war of who controls the internet, and ultimately you - and your data - is as hot as it ever was back in 2002 as the old guard jockied to rig the game and stay relevant as their world changed. Only today, it's the new guard - the web giants, Google, Facebook and Apple - rather than enterprise players that are doing the jockeying. Sure Microsoft and Oracle are players, but they aren't in the driving seat this time. The fight, meanwhile, has drifted from the dusty halls of groups like the WS-I and into the front lines and the trenches on the community as Google and Facebook battle for access to each others' users' email information.

Today it's no longer about who controls the process that's used to write the paper spec that we all then supposedly dance to - as it was in 2002. Now, it's over where the APIs are built and how far and how fast those APIs can penetrate our world in meaningful numbers that the web giants can claim. ®