Informatica's Ash Parikh is a man on a mission: He thinks it's time companies gave data first consideration when moving to a service-oriented architecture.
This isn't surprising, since Parikh works for Informatica, a company that specializes in data integration and quality tools. But before you wax cynical about that, let me just add that Parikh once told me it was this very proposition of exploring how data works with SOA that attracted him to Informatica in the first place.
The problem, as Parikh explains in a blog post this week, is that companies focus too much on application architecture with SOA, assuming the data integration issue would somehow resolve itself as they moved to an SOA.
Unfortunately for them, it didn't, says Parikh, who says he recently took some time to explore the issue further, talking to analysts, experts and practitioners in the integration space.
He's come back with tales of woe from real companies who he said fell victim to two SOA myths:
- SOA will guarantee business agility.
- The SOA technology stack doesn't need to include a data integration layer.
He gives a real-world example of a financial company that assumed its ESB and EAI (enterprise application integration), along with their SOA implementation, would be able to handle the data issues. Instead, they wound up with what Parikh terms an IT fiasco:
"Their back-end data sources processed customer information at varying latencies such as batch, trickle-feed and real-time. Their customer information across these heterogeneous data sources was incomplete, inconsistent, inaccurate and not in the desired format. Further, there was no seamless way to prove who touched the information and when the information was updated to support their requirements for regulatory compliance."
The ESB, EAI and even hand-coding failed to solve the data problems, he adds.
"As you can see, it is rather unfortunate that data integration often becomes the 'blind spot' in an SOA implementation. Without a complementary service-oriented approach at the data layer, SOA will more often than not, continue to fail."
He includes a link to a previous post about data services, which includes a graph showing how this concept works. You can also check out my April interview with Parikh, "Drilling down on the Concept of Data Services," if you'd like to learn more.