« Diversity seems to remain the core feature of an ESB | Main | Governments and SOA »

It has been clear since the end of last year that 2006 was going to be an exciting time for SOA adoption: a time when SOA would start proving itself or risk becoming last year’s cool idea. On the positive side, it is going to be a year when vendors’ offerings will start matching the promises of 2005 (as can be seen with recent announcements from Sonic Software, IONA and IBM among others). But it is also going to be a year when the stories of SOA missteps will echo louder than the SOA successes as some adopters learn the hard way and the press looks for a fresh story from SOA.

I have been thinking about what needs to happen for SOA to move through this dangerous period and here is a quick list – feel free to add your own as comments:

Develop SOA patterns
The value of patterns is well understood and in the integration/EAI space there has been some excellent publications such as Gregor Hohpe and Bobby Woolf’s Enterprise Integration Patterns. As SOA deals with integration, some SOA patterns should be simple updatings of the integration patterns. Of course others will not. I have seen various comments about the need for patterns and even a web-site with some initial suggestions but can’t find anything that feels any way complete. If any reader can point me at anything more concrete, please feel free to comment or email me.

Identify best-chance use cases
The next step up from patterns are use cases which allow practitioners to identify what type of project will yield the best results quickly. I know from my experience that there are types of projects which are good candidates for those high visibility first step into SOA initiatives and others which should be avoided (perhaps they are too small to matter or to entwined in political issues to be successful whatever the technology).

While the information is out there, the problem is always getting them into the public domain and inevitably vendor use cases highlight how great the product was over the pattern.

Focus on the standards that work and get to work
We live in a world always interested in the new new thing – and even more so in the technology industry. Bottom line is that the technology we have is good enough to get started with. SOA should be flexible enough to evolve as the standards evolve. Lets not wait for the WS-mumble to be standardized, released and adopted.

Figure out how to convince the business that SOA is a way to save money and increase efficiency, not burning budgets on this year’s new thing
As yet another survey shows that even IT chiefs are sceptical of SOA seeing it as hyped and a marketing term. The telling quote is from Melvin James of Diagonal (a Systems Integrator and SAP partner): “There is a worry at board level that SOA is about new technology and new spend rather than leveraging what organisations already have and improving process”.

As SOA champions we need to be clear on the real core benefits of SOA: a way to increase efficiency and reduce cost through leveraging existing investment. In fact these are precisely what board level management is looking for, but as Mr James also points out, too often it all gets lost in competing vendor pitches and technology standards.

Posted by rbradley in SOA Predictions • SOA concepts |Digg This|Add to del.icio.us

Trackback Pings

TrackBack URL for this entry:
http://www.ebizq.net/mt/mt-tb.cgi/179