I just read a really interesting post that sparked me off on a train of thought that I just had to share.
Joe McKendrick writes an SOA blog for ZDNet (amongst other things) and often poses some interesting questions or disseminates some key nuggets of thought from the SOA blogosphere. One of his recent posts talks about how Pfizer have bridged between Service Oriented Architecture (SOA) and Master Data Management (MDM). I found this particularly interesting for two reasons.
- The Microsoft ESB Guidance Toolkit has just gone RTM and been published to MSDN. Marty Wasznicky (who led the team that built this) acknowledged at the Microsoft SOA and BP Conference that this Patterns and Practices guidance toolkit was put together following a successful ESB implementation that Microsoft produced using BizTalk Server at Pfizer. Clearly this is a key plank to the SOA that Pfizer is operating.
- Conchango led the way on MDM and it's usage in SOA a couple of years ago. My colleagues Mick Horne, Rob Grigg and Jamie Thompson were working on a large scale integration project in the Oil & Gas Industry. They faced three key problems. They needed to commodify what they were building so that this could be rolled out to other markets, plus they needed to standardise the process of getting data from multiple disparate data sources, and each of these data sources had their own data dictionaries and domain specific languages. Clearly a Service Oriented approach was going to tick the first two boxes, and MDM rode to the rescue for the third problem. They called it Service Oriented Business Intelligence, and coined yet another TLA - SOBI. Their work was published in the January 2006 issue of the Microsoft Architecture Journal and presented to the Microsoft Architects Insight Conference. Jamie published a series of posts giving loads of details into how they solved the problem. You can find a summary post here.
So what does this series of links tell us? That a good idea can be a long time coming?
Perhaps. The argument goes that IT, for all it's promise of "agility" "scalability" and every other -ility at the forefront of a CIO's agenda, is still just an industry like any other. So whilst Silverlight can be the darling of rotating text boxes in no time at all, no C level decision maker is going to stake the future of business application construction and data rationalisation for their organisation on SOA and MDM if it might just be the latest great idea that doesn't quite make it into the air.
Perhaps this news from Pfizer shows us that it's not only our Oil & Gas client (who were in a unique position of having no other choice but to adopt a SOA + MDM route) who think that SOBI (or SOA + MDM, or one version of the truth plus a single enterprise data dictionary plus composite applications built upon services, or whatever else you want to call it) is an idea that has delivered the -ilities that it promised, and has not only taken off, but delivered down to earth benefits.