Building an SOA without the hype

One cannot help but notice the buzz around the concept of the service oriented architecture (SOA). With so much industry talk on the topic, it is a wonder how users ever managed before. Read: Getting more than integration from SOA.

And anyone without an SOA strategy may as well give up and go home, because everyone else has got one.

The thing is, SOA is not a new concept. IT started out as a service to business; network providers offer a service to connect computers together; the chargeback scheme some IT directors use to charge business departments for IT; and the whole concept of utility-based computing relies on the premise of a service oriented approach to IT. Moreover, operating systems, databases and infrastructure middleware provide software architecture which deliver services that applications can use. Sounds very much like an SOA.

What has changed is that the IT industry has recognised the need for standard procedures and protocols to define communications between applications and service levels. Arguably, the suppliers should be applauded for agreeing to work together in industry alliances to define the required standards. But it has been the industry which sold users proprietary systems that could not readily communicate with each other. And now it is selling a solution to the problem it created.

Ignore the hype. There is no need to use the latest and greatest protocols to support an SOA. As Mark Saldanha, head of IT at Great Ormond Street Hospital has found, it is possible to develop a service oriented architecture without having to use web services.

A quest for openness

There can few more damning critiques of IT project failure than that presented by senior diplomat Norman Ling to the Foreign and Commonwealth Office last year. Foreign Office IT project pushes staff to 'wits end'.

Looking at the lessons learnt from problems with the FCO's Prism project to replace 30 legacy systems with a single Oracle ERP package, Ling said the troubled implementation highlighted the shortage of project and programme management skills at the FCO. He also urged greater transparency and independent scrutiny of future IT implementations.

Given the importance of Prism and the number of basic errors made in its project management, it is particularly disappointing that the FCO tried to keep private Ling's report. It makes the determined efforts of the Foreign Affairs Committee to bring the report into the light of day all the more laudable.