08/21/08 :: [SOA] IT's Silver Bullet [permalink]
No surprise I really like the two latest addition to the REST / WS-* debate (or WOA / SOA). Eric Roch asks whether WOA is simpler than SOA and Arnon discusses when to use (or not to use) REST.
I really like these posts because they convey a very important
message that goes well beyond the silly WOA/SOA debate. What Eric and Arnon tell us is that there is no silver bullet, there are
only problems and solutions. The only silver bullet there is (as
Eric points out), is
"think very hard about what you are about to do", and that's
technology neutral as we say. Let's get some perspective
about what happens when you don't think really hard about what you do.
The world is changing as we speak. Yes, I know, yawn... In the coming years, I predict there will be a massive rate of company failures (meaning being bought by their competitors). The reason? simply because, across all industries, a lot of companies cannot compete anymore, the fittest will kill and absorb the weakest. My best guess is that 50% of the companies that are operating today will be acquired within 5 years. This is just a question of margin and IT's inability to deliver best in class business models.
What's happening? IT does not matter anymore... What does it have to do with a company's margin or business model?
Today's weaker players believed what
Nick Carr
(and other Pundits like John
Seely-Brown) told them
and drunk the cool-aid of the "utility computing" myth. They stopped
investing in their IT organization, or more exactly, they stopped
thinking about their IT organization as both an offensive and
a defensive competitive weapon. A lot
of IT organizations became lazy and preferred playing technology
roulette or plain politics rather
than achieving excellence. They hired lousy hand waving CIOs to "manage" the utility,
they failed to train their workers, they failed to upgrade their
systems, keep up with the latest technologies and instead went into an
"agile" mode where everything of value can be done and must be done in
less than X weeks (where X is rarely more than one) with hardly anyone
or anything, let alone any thinking.
Thank you Nick, you have designed one of the greatest Trojan horse in the history of the US economy.
The vendors did not help much, they somehow accelerated this phenomenon (involuntarily), they were busy understanding what to do with the Web and deliver half-baked, ever evolving, always expensive, but of course, wonderful "suites". Actually, the only thing I like about REST/WOA is that it is baked, never evolving, free, unfortunately it is not "wonderful" (i.e. it is no silver bullet)
So today, where are we? I said it, I'll say it again, IT is dying. Some people benefit from that. If you have a consulting or SI business, you are likely to have more business. But be careful for what you wished for: an economy with a weak IT, with a "utility" IT will inevitably lead to weaker companies in the global economy, unable to compete (and invest).
How do you compete? Contrary to what Mark Masterson has written, business process automation has been incredibly successful. How many paper-based processes there are left? How much visibility do we have in business operations today? Business Process Automation is not just about replacing people with servers, it is also about making the work of these people meaningful and let the task of "keeping state" to the computer (which humans can't do very well): a Sales person can now focus 100% of his or her time on selling because he or she does not have to mess with fulfillment or support. What has not been successful is BPML or BPEL as a technology to deliver business process agility or create a process-centric programming model. Yeah, that has failed miserably (I digress). Business Process Automation is directly related to margin, the more efficient this automation, the more visibility this automation provides, the better the margin, the better an enterprise can compete locally and globally.
IT should not matter (which is very different from Nick's IT does not matter). IT should deliver what the business needs, at a cost compatible with the business model and within reasonable time frames. Companies should compete on their business models, not on their inability for their IT organization to delivery the business model. This is what I mean by IT should not matter. There is no silver bullet for that, if you continue down the path of extreme simplification rather excellence (which does not prevent you to keep things as simple as possible) you are doomed to fail because you will built systems that can't be changed with elements that can't be reused. Any adjustment to your business model will cause excruciating pain.
What Nick Carr did not tell you is that, in this day and age, if your margin is trailing industry leaders by a few percent it almost always means a death sentence. You have about 12 to 18 months to react once you reached this point. And only your IT organization can help you out: it is not a question of Off-shoring or SaaStifying entire processes and solutions, it is a question of "right-sourcing" (i.e. performing tasks and operate assets where the cost is lowest): only IT can orchestrate that, no ISV, no SI, no cloud or rain can do that job. Understand that there is no silver bullet (Web 2.0, WS-*, WOA, Social Network, Cloud Computing... ) other than excellence.