"You don't need a two-ton car to move your 150-pound butt on crowded city streets." -Dean Kamen
I just read an interview with Dean Kamen --inventor of the Segway and robochair -- in which he talks about his out-of-the-box philosophy toward innovation:
"I get up in the morning and work really hard at really difficult problems, because I think sometimes taking a different approach to big problems is likely to solve them -- in fact, far more likely to solve them -- than all the approaches that haven't worked up to date. If it's still a problem now it means that conventional wisdom hasn't worked."
Kamen's approach to big problems has been assembling small-scale, relatively cheap technology in new ways to address these big problems. Which made me wonder how he would approach the problems of integration of complex, spaghetti-code-laden systems across enterprises.
One thing is certain -- Kamen would look for ways to make integration and compontization lighter, simpler, faster, and smarter.
Web services initially offered such refreshing simplicity. But those were the days when they were mainly concentrated on Web server apps. Can we keep this simplicity as we move deeper into the enterprise?
Perhaps Kamen's approach would include more reliance on devices -- a la IBM Datapower or Reactivity. The solution as he would see it would be to develop miniaturized black boxes that would handle all aspects of SOA, from security to load management. And perhaps even provide a way to store and deploy services themselves separate from the server infrastructure.
Or, architecturally, Kamen would likely say that enterprises have become too huge, unwieldy, and unmanageable. He would be a proponent of breaking processes -- or even businesses themselves -- down into manageable components, perhaps to a fine-grained level.
Conventional wisdom says that big problems such as integration require big solutions. Kamen may beg to differ. The Segway hasn't exactly been a runaway bestseller, but it does offer a compelling alternative for our traffic-choked, gridlocked cities. How would Kamen look at the SOA challenge? Is the conventional wisdom -- throw more big investments at the problems -- working? Or should we seek ways to be lighter and faster?