Ronan Bradley's Roads to SOA« Some reasons why government likes SOA | Main
May 05, 2006WIll Web2.0 kill the SOA platform?
I am struck by the difference between the discussions and announcements around SOA and Web2.0 – don’t panic I am not going to get involved in the REST versus SOAP debate or the debate about which is more hyped or a vaguer term!
What has struck me is the way the big software players are talking up the concept of a SOA platform in a very traditional way: as a stack of vertically integrated software all of which is required to solve the SOA problem, and recently covered on ebizq.net here and here . Bizarrely RedBoss, the new boy in town which should surely be rooted in the new world rather than the old, is taking it even further back in time by suggesting an Operating System is also needed in the stack!
At the same time as all of this, the analyses of what have already happened on the Web and what is accelerating with Web2.0 shows that the very concept of a platform is changing. This is best summarized by Tim O’Reilly’s much referenced article on “What is Web2.0”. The point of relevance here is that the Web 2.0 battle is not between a platform and an application or even between competing but broadly similar platforms – it is between radically different platforms with radically different business models.
So why is this not happening in the enterprise? The old claim that integrating the stack removes the need to integrate multiple vendors’ products and hence reduces cost and risk must be much less true than it was. Open Source in particular has demonstrated that it is possible and relatively easy to stick together multiple products/projects.
Obviously there are many different business dynamics at play in the enterprise than across the consumer focused web. One dynamic is the need in the enterprise to limit the number of vendors (‘one throat to choke’) and incumbency is also important both from a skills and relationship point of view. However, this can’t justify the view that a single integrated stack is the right way to go, unless the stack is at least a good-enough solution. Furthermore, integration in particular is clearly not about tightly coupling homogeneous single-vendor supplied software: Its essence is spanning across platforms, applications and organizations.
The question is therefore will the changes coming from consumer-centric Web-world change the way the enterprise-centric SOA-world to the same extent?
We are already seeing the use of technologies related to Web2.0 (RSS, AJAX etc) growing in the enterprise. However, these are towards the edges of what SOA is attempting to do as they are about user interaction rather than system to system integration: This is important but not yet core.
What I do not believe we have yet seen is the appearance of a google equivalent (which could be an open source project) for SOA – a company which solves the problems in a radically different and much better way and hence dislodges the good enough solutions. Until that happens the big vendors will continue to push enterprise software the same way they have pushed it for 10 years and more.
Posted by rbradley in SOA Predictions • Web2.0
Trackback Pings
TrackBack URL for this entry:
http://www.ebizq.net/mt/mt-tb.cgi/238
It's an interesting view but not one I see shared in the applications space where the money battles are being fought. Most recently, Jeff Nolan at SAP torpedoed SOA as a dead concept. The web 1.0 EAI vendors such as WM/TIBCO see SOA as something they've done for years. Everyone else wants a position so they can get into the Magic Quadrant. Often they've no idea why.
The real problem lies in the term itself, its vagueness and its associations with blue sky projects that demonstatrate little more than debatable business value.
For goodness sake Service 'Oriented'....is that some sort of veiled sexual IT preference thing?
flipancy aside, I prefer Loek Bakker's analysis where he talks about services 'based'...the distinction makes it much to understand how the loosley coupled puzzle fits together.
It's always been my view that for SxA projects to stand any chance, they needed to be applied to specific, edge projects. Web 2.0 is basically saying: 'Yes, we're edge but we deliver value where it's needed and we do it now through a global community we can ll tap into.' And that is what users want - or so their money says.
That still begs questions around open APIs which is where a number of web 2.0 services start to unravel. So maybe we are faced with same old/same old integration. It's simply the focus is shifting, and with it, the integration spend $$.
Posted by: Dennis Howlett at May 6, 2006 06:09 AM
Post a comment
Subscribe to feed