I've found a wickedly funny, irreverent and occasionally even (mildly) offensive Webcast on SOA, integration and ESBs for you.
I know -- that's a rare feat in the IT sector. But I promise -- "Does My Bus Look Big in This?" delivers.
ThoughtWorks' Chief Scientist Martin Fowler and Dr. Jim Webber, global head of architecture for ThoughtWorks, gave this presentation at Qcon. You may remember Webber as the advocate of Guerrilla SOA -- a concept he revisits in this 42-minute talk.
No doubt, you're thinking you don't have time for a 42-minute lunch, much less a Web video. But you'll want to make time for this if you're at all interested in or involved with integration or SOA.
Here are six reasons why:
Along the way, they manage to call SOA out as a "dog's breakfast" and compare integration middleware to a "fat bloak" with a "beer belly" with "man boobs." And of course, there's the gratuitous reference to Lord of the Rings.
The core of their case is that SOA sounds good, but in practice tends to look like the same old spaghetti-connection architecture we've always had -- and an ESB just puts a nice door on it. One of the arguments for SOA -- and against Guerrilla SOA -- is that by designing SOA as a whole, you're able to create a more strategic architecture. But Fowler contends that this more traditional approach to SOA can also be inflexible and down-right unrealistic.
He says a SOA expert once told him the key to building an SOA was to make sure when you built a service, you built it so it never needed to be changed. "Does that sound like a workable plan to you?" he asks the audience.
Well, when you put it that way, no, it doesn't. He points out that agility is based on three principles:
And yet, he said he's heard from enterprise architecture teams who are proud of the fact that they don't talk with business users.
Achieving agility requires tools that allow you to build systems multiple times a day, quickly find out if something goes wrong, and take a vigorous approach to automated testing, he noted. It sounds pretty impossible -- until you consider that we already have a network that's achieved all of this on a much larger scale than a mere enterprise.
"The thing about the Internet is it's the stupidest network on the planet," said Webber. "It's dumb. Really dumb, but it's global. And the dumbness of the Internet is a real, real win because it enables us to do innovative things at the edges."
They don't just talk theory, either. Fowler and Webber offer specific technology suggestions for what makes their vision work. They offer a critical challenge to much of the status-quo, vendor-dominated thinking around SOA these days. And even if you disagree with their argument about Web-based design and the common-sense Guerrilla SOA, you'll still be entertained along the way.
Thanks to Rickard -berg, who pointed the way to this great resource when he linked to it from his "Stuck in the Middle" blog.