The Register ®

Biting the hand that feeds IT

The Register » Software » Developer »

Original URL: http://www.theregister.co.uk/2005/12/05/ibm_soa/

IBM mainlines on SOA

By Philip Howard, Bloor Research
Published Monday 5th December 2005 23:06 GMT

Comment This is the first of several articles following on from IBM’s recent Software Group annual analyst conference. Typically at these events there is some sort of theme that runs through all the presentations but this is usually as much notional as a reflection of reality. When it comes to individual product or product group presentations the theme conveniently gets lost.

Not so this year. This year not only was it clear what the theme was but all of the various IBM product groups – Lotus (Workplace), Tivoli, WebSphere, Information Management (DB2 et al) and Rational – were an integral part of the message. And the message of course is SOA (service oriented architecture).

The first thing to note is that while IBM talked more or less non-stop for two days about SOA it barely mentioned web services. The point being, of course, that you can build an SOA without web services. It is unlikely that you would but it is conceptually possible. But even if you do use web services they are not likely to be 100 per cent pervasive. And, even if they were then that is still only a small part of the SOA story. So, CICS 3.1 (which supports the generation of web services for CICS transactions) was barely mentioned, while the construction of web services in Rational was not discussed much either.

Actually, there were a number of processes demonstrated that were actually generating web services but that wasn’t really the point of the presentations. For example, building composite applications using the Workplace portal (by end-users) or the WebSphere Portal (by IT) was demonstrated to illustrate how easy these environments are to use, rather than the generation of web services per se.

Similarly, the demonstration of Information Server (to be released next year and discussed in a subsequent article) in which you could simply lasso an ETL (extract, transform and load) process and publish it as a service operation by means of a wizard was more aimed at illustrating the functionality of this product than at the fact that you were actually creating a web service.

No, web services was not the issue but SOA on a much broader basis: the governance of the SOA environment (both in development and on an on-going basis), the lifecycle of SOA, the business process management that underpins SOA, the way that master data management supports SOA, the importance of integration to SOA, the need for open standards to drive SOA, the monitoring of SOA and assuring that it meets service level agreements, and so on and so forth.

Once you start to appreciate the relevance of all of these tasks to SOA, you start to appreciate why IBM has 25 times more SOA customers to-date than its nearest rival (according to IBM’s figures). There is simply a breadth of requirement that is needed that no-one else can match.

That doesn’t, of course, mean that the company will continue to dominate the market to this extent: it will come back to the question of whether you want a best-of-breed solution (because, undoubtedly, there will be suppliers that can offer particular point solutions that are superior to IBM’s) or a single vendor that can supply a full breadth of capability. There is enough demand for the latter approach that IBM is likely to continue to be the market leader for SOA for the foreseeable future. ®

Copyright © 2005, IT-Analysis.com (http://www.it-analysis.com)

© Copyright 2008