Data Integration Remains a Major IT Headache
Study shows that data integration is still costly and requires a lot of manual coding.
By now, we all know AT&T wants to buy out T-Mobile, which, as far as I can tell, wants to be bought. In 12 months or so, we should know if the deal goes through. If it does, the merger will make AT&T the largest wireless carrier in the U.S. with 130 million customers.
Of course, integration will be a huge issue if that deal goes through. So recently, Brett Sheppard, the executive director at Zettaforce and former senior analyst with Gartner, examined how T-Mobile handles integration across "a federated, disparate infrastructure."
Even if you have no interest in the AT&T deal, it's a chance to see how a large telecommunications company deals with data silos, integration and legacy applications. I must say, I'm surprised at the level of detail included in this O'Reilly Radar article. Sheppard doesn't just talk generically about data integration, but lists specific legacy applications, specific "Shadow IT" applications and data warehouses, and includes an evaluation of specific solutions T-Mobile uses to address problems. It's so detailed that it's a bit tedious to read, but the graphics of the infrastructure go a long way to clarify how everything relates.
Sheppard talked with T-Mobile Integration Architecture Manager Sean Hickey, who shared the infrastructure required to bring together and manage all this data and manage it. The key, it seems, is a virtualized data integration layer that includes:
Its data infrastructure also divides information into six virtual data zones, each tied to a business area like "customer," "customer experience" or "supply chain." Sheppard writes:
From a data architecture perspective, the data zone approach helps pinpoint where there are complex systems to maintain, shadow IT, redundant feeds, differences in data definitions or incompatible data. This approach also helps highlight where business rules are embedded all over, leading to duplicate or inconsistent business rules, versus more centralized rule management.
The article also looks at how this integration layer impacts regional marketing, churn management and customer care. Like I said, it's more detailed than you'd typically expect from a tech journal article, with enough detail to satisfy the curiosity of most architects and technology leaders. T-Mobile has 3.6 million customers as of the second quarter of 2011, when it valued its service revenues at $4.6 billion, and, frankly, it's unusual to see that much detail online about the infrastructure of a large technology-driven company.
Full disclosure: I have a minor interest in AT&T acquiring T-Mobile. Until this year, I bought minutes as I needed from T-Mobile, spending a whopping $25 at T-Mobile every two months for two cell phones. Yep, you read that right. And the coverage was perfect - no problems, no static, just excellent. But one minor mid-life crisis and an addiction to Angry Birds later, and I'm stuck with a two-year plan, a $140-a-month bill from AT&T and coverage that, frankly, frustrates me. So, yes, I wouldn't mind the deal going through so I could have better cell phone reception.