« JackBe: nimble AJAX clients for SOA? | Main | Sound SOA advice from the 1960s »

Joe McKendrick’s summary at the end of August of opinions about whether reuse is achievable with SOA or even if it is the real benefit of SOA seems to have ignited an increasingly widespread debate. Reuse is now being discussed in CIO blogs such as here , slammed by Miko Matsumura in “Die SOA Reuse Die”, and defended in my own Illuminatus Research blog among other places, with Joe stirring the fires again at his own blog .

So why the big debate? First off, it is not surprising that at this stage in the SOA adoption curve, failures (or at least lack of success) will be trumpeted more loudly than success. However, there have been successes with SOA reuse for instance at Wachovia as I previously commented on.

Even if we accept these set-backs, I do not believe that we should simply abandon reuse as a goal – frankly without reuse, aren’t we giving up on the benefit of SOA which is most tangible? And while it is tempting to move away from the tangible and adopt the higher level abstract goals of “business agility” or “better infrastructure”, they are hard to measure without being tied back to reuse and hence hard to justify a business case with. As my Illuminatus Research colleague, Steve Craggs puts it:

“a lot of these are grand, strategic benefits can be somewhat difficult to measure, and this is a problem to those people trying to build business cases to justify SOA investment. I think this brings us back to reuse as being a major pragmatic driver of SOA - that is, a driver that actually drives real SOA investment.”

Which brings me to my own three step solution for driving up reuse levels (yes – it is high level and simplified but anyway):

1. Get the technology right – make sure your service definitions are designed with reuse in mind (Gary So of Webmethods has recently covered this subject . Get the technology back-up in place to ensure that services are easy to reuse from registries to repositories to governance to wikis.

2. Get the organization right – SOA isn’t just about technology and in particular it is clear that relying on the technology alone won’t deliver reuse. Some commentators seem to throw the towel in and say that conservative organizations won’t be able to make the change, I am not sure so. The primary change I would recommend is to identify clearly service owners. This was the approach taken in Wachovia and other places that have had SOA reuse success. It is also an approach which is inline with even conservative management theory: clear responsibility, clear measurement.

3. Pay up: Incent with bonuses both the service owner and the service consumers to meet reuse targets. If you incent both ends of the transaction, things start to happen. Yes, there are risks of abuse that must be policed against but the rewards are greater than those risks. I have written a free-to-download piece on this and the second step on at the new Illuminatus Research site http://www.illuminatusresearch.com/.

Unfortunately, as technologists we are comfortable with item 1, willing to discuss item 2 primarily in the context of the IT department (SOA competency centres and service librarians are good ideas, simply not sufficient) and seem shy to address 3 at all!

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

Trackback Pings

TrackBack URL for this entry:
http://www.ebizq.net/mt/mt-tb.cgi/790