|
SOA: The shiny new toy
by Scott Gnau
When I was a kid, I collected Matchbox Racers. You know the kind—little, bright colors, the ones you could send whizzing across the floor to terrorize the dog.
My favorite was a blue Ford GT with a white stripe, the same one that was on the side of the Matchbox carrying case. I couldn't tell you how many hours I spent playing with it.
I have a confession to make, though: as much as I loved that GT, I was always a sucker for the shiny new car.
Maybe it's human nature. We're all suckers for the shiny new toy, whether it's the latest home electronics gizmo or the shiny new business tool. Like, say, service-oriented architecture (SOA).
We can all agree SOA is important. Finally, we're able to create applications that integrate well and are easier to support across
an enterprise, which means you can deploy more applications to more users more efficiently and at a lower cost. And that's great.
So what?
If you're not giving your end users any
more time in their day, will more applications really help them? What's important
is not only doing application integration to serve end users but application integration
to automate business processes, and that's
the whole promise of SOA.
But there's a catch. If your underlying data infrastructure is not built solidly, you could actually be delivering answers that are detrimental. Real-time enterprise without a true enterprise data warehouse underpinning
will just enable you to make suboptimal decisions faster. Is that good for your business?
There's more, of course. Service-oriented architecture requires a completely new way of thinking for data warehouse professionals, like moving from FORTRAN to object-oriented coding. It's a different skill set, so organizations are going to have to wake up and
staff the data warehouse team to that skill set; otherwise they risk getting backed into the position of applying new technology on
a fundamentally old architecture.
And it's not just the IT staff that has to think differently. A lot of times a shiny new toy like SOA will come along and people
will implement it purely on a technology benefit basis. Wrong. What you really have
to do is decide what your business needs
to be successful and set a vision. Define
an architecture that allows your business to execute on that vision and only then choose your technology.
At Teradata, our core competency is the data integration layer, a key part of that overarching architecture. We've also built technology and tools to streamline integration with the application tools that you may have already deployed in your enterprise.
Sure, keep your eye on those shiny new toys, but it's the solidly built favorites that
will continue to surprise you with what they can do. It's just like those Matchbox cars.
No matter what else I got, I always came
back to that Ford GT—its original design and
materials still outlasted anything I ever added to my collection. T
Scott Gnau is vice president and general manager in research and development
at Teradata.
Photo courtesy of San Diego Union-Tribune.
© Teradata Magazine-June 2006
back to top
|
Quick Links
Print
Send to colleague
PDF
E-mail us
