« The mainframe: not sexy, but still core to SOA | Main | Will Web2.0 kill the SOA platform? »

I will have to admit to having just figured out how to publish comments that people have been making on this blog, which gives me the perfect opportunity to respond to some of them...

In particular, both James Taylor and Bob McIlree commented on my piece on government and SOA suggest the problem was not technical and was around management and project management. Bob goes further to say that

Indeed this is mere wand-waving and managerial mumbling of "magic words" in the hope that something, _anything_ might prove useful.

Before going any further and appearing to be using broad generalizations to condemn a huge diversity of people and organizations, I should of course caveat all of this by pointing out that some government agencies have been successful in both vision and delivery for a long time, and more are catching up fast and showing vision and capability ahead of many commercial organizations. However, inevitably the disasters make for better news.

To get back to the comments, I would absolutely agree that the problem is not exclusively about the approach taken, nor about the specific type of product used. However, ‘fixing’ the management approach will tend to go hand in hand with moving towards a SOA approach.

Fixing the management approach to a degree is about recognizing the scale and nature of the problem. The old approach of outsourcing strategic initiatives in isolation may have worked when these were typically around rolling out a new application which existed in isolation from any other. However more and more of the political agenda is about efficiency and agility through better organizational integration (nothing to do with technology necessarily). This agenda can be driven by the need to control escalating costs in a government controlled health sector or the need to detect and respond to security incidents or natural disasters. Better organizational integration is constrained by many operational constraints related to the sheer number of agencies involved (the many hospital groups, community health workers, family doctors, etc in the health example, the multiple police and emergency services at town, county, state and federal level in the second example) and the diversity of their systems as well as legal constraints in particular around privacy and security.

This political agenda translates into, among other things, an IT agenda which now must also be primarily about integration of IT assets as that underpins the organizational integration. Which is where SOA comes in: SOA can be used as a government wide architectural approach to integration which can inform each integration project that is undertaken.

The attractiveness is obvious:

* SOA is about integrating independent entities which matches the government domain with its many independent or semi-independent agencies involved, each with agendas and legacies. It allows each agency to make their own technology decisions and incorporate their choices within the overarching architecture.

* Furthermore, SOA is also about distribution rather the centralization of functionality. This matches both technological limitations associated with building mega-hubs and more importantly the political constraints as most citizens are suspicious of government agencies attempting to centralize control and power too much.

All of which has meant that governments world-wide have become some of the most enthusiastic adopters of SOA and is already being used strategically by a number of agencies that I am aware of.

Posted by rbradley in Government • Market trends |Digg This|Add to del.icio.us

Trackback Pings

TrackBack URL for this entry:
http://www.ebizq.net/mt/mt-tb.cgi/224

Posted by: Dennis Howlett at May 1, 2006 01:11 AM