Informatica's Perspectives blog recently had a nice piece about data integration by one of its managing directors, Suganthi Shivkumar. It was a nice summary of the situation with enterprise data, challenging readers to imagine how better integration could be, the historical basis for data fragmentation, and the costs of these problems.
Shivkumar wrapped up with an explanation of how taking an enterprise-wide approach to data integration-particularly though the use of an Integration Competency Center-could change the course of enterprise information for the better.
Like I said-a nice piece, if you're completely new to the problem of enterprise information. For those of you who read a lot about data integration, its core message is a familiar refrain: Make integration an enterprise-wide task.
The readers' comments, however, were a bit more lively and thought-provoking.
"While I agree with everything written here, I submit that a fundamental component was omitted from the paper. While many organizations are establishing Integration Competency Centers (ICC), most are underfunded to accomplish what needs to be accomplished," challenged reader Charles Heiser. "Logically, it makes all the sense in the world to establish an ICC but when it comes time to open the wallet, few step forward."
Heiser explains that from his experience, squeezing value from an ICC takes time, and with annual budgeting, it's hard to justify the expense when you won't realize any benefits in that budget year. This acts as a disincentive for business units, he argues, who are being asked to fund the ICC and yet "may see little benefit to their bottom line."
He suggests ICCs will find more success with a blended model, where individual projects are pursued but include an overhead fee that goes to fund the ICC. He also suggests a very practical political alliance with Finance may prove helpful in funding an enterprise-wide ICC.
There's more than one way to approach the problem, of course. I've shared Rick Sherman's view that there's more than one type of ICC, and it's worth pointing out that the types of ICC vary substantially in terms of their funding requirements.
One thing is becoming increasingly clear, however: Continuing with status quo is a bad, bad idea.
David Linthicum recently drove home this point with two posts looking at how bad data can mess up SOA. Linthicum specifically chides SOA architects for neglecting the data.
"What's missing within most typical SOA projects is the focus on the data, and that is killing SOA," Linthicum writes. "The truth is that most failed SOA projects can be traced to the lack of a data level understanding, and while this is still an issue in this day and time is beyond me."
In a follow-up post this month, he elaborated on the problem, observing that there are three data-related problems SOA architects typically ignore or neglect:
Overall, he though this showed how often SOA architects are too focused on tools and technology, instead of-all together now-the whole architecture.
While I doubt SOA's the only thing impacted by this problem, I think his example nicely illustrates how bad it can be to ignore data and enterprise data integration. It may be hard to justify the cost of an ICC, but it's quickly becoming much harder to justify the cost of the status quo.