Every year IBM holds its Impact event devoted to SOA, a method of sharing application resources that may have lost its luster among analysts, but continues on the long road to adoption in many large enterprises. Steve Mills, senior vice president and group executive at the IBM Software Group, is seldom shy about offering his opinions about SOA, industry developments, or his competitors.
[ For a quick explanation of the ideas and benefits behind SOA, check out InfoWorld's three-minute InfoClipz on the subject. ]
At the IBM Impact event last week, I got some unfiltered responses from Mills about all of the above topics, including Oracle's purchase of Sun Microsystems (that question came last). What follows is an edited version of that Q&A session:
Eric Knorr: They say SOA is dead. So why are we even talking about it?
Steve Mills: It's a "rumors of my demise are greatly exaggerated" kind of thing. People say these kinds of things to be provocative. It's a silly conversation, like, "We had a dot-com bust, so the Internet is dead." No, the Internet has been with us for decades. It's not the kind of thing that dies once it's established.
Knorr: I've interviewed you about SOA every year for three years. What's different this year? How would you say customer SOA engagements are changing?
Mills: The level of sophistication on the part of the clients is dramatically greater. It's evolutionary -- 2009 is not some monster step function different than 2008. But going back to a 2004 timeframe, you had a lot of people looking for a way to get started. More recently it's about what's been working well.
Knorr: Governance problems stemming from organizational barriers have always been the greatest barrier to SOA adoption. Has that changed?
Mills: Tough economic times carry the benefit of forcing people to face up to governance issues they may have avoided before.
Knorr: A lot of the SOA work that was done in financial services is simply gone. Lehman Brothers, for example, had some of the best-laid SOA plans I've ever seen. What happens to that intellectual capital? Weren't financial services leading SOA?
Mills: No, no. We never felt that those customers had some unique corner on leading the effort. In fact I have always said for years now that the most prolific customers have been insurance companies -- dramatically more prolific than the financial institutions.
Knorr: Really? Any particular cases?
Mills: All of them. Universally, consistently -- whether it's State Farm, whether it's Aetna, Standard Life -- across the board. The reason why is that if you decompose what their business processes do, you can see the patterns. And they tend to be less oriented toward competing with each other from within. At many financial services firms, you compete with your colleagues for bonus money, so you have competitors outside and you have competitors inside. And that's a lot of the reason they've operated like they have in fragmented mode.
Knorr: In this economic environment, are you seeing more demand for lighter-way solutions that deliver at least some of the value of SOA?
Mills: Integration on the glass through portal technology has been a very popular way for government agencies to collaborate with each other for constituency-based services without changing their back-end applications, which they all covet and control.
Knorr: On to cloud computing. IBM has provided the infrastructure for many software-as-a-service ventures. You also provide outsourced datacenter capacity on a one-off basis for large customers. I'm sort of surprised that IBM isn't offering something like Amazon Web Services. Have you ruled that out?
Mills: No, we haven't ruled it out, but some of these business models are challenging to figure out how you make money. Infrastructure on demand is an odd notion, and you also get what you pay for. I have stuff running on Amazon. I have a wonderful contract with Amazon: It costs me nothing, and they guarantee and commit nothing. It's totally symmetrical. I have no quality of service, no security, no privacy, no recovery, and, by the way, if I ever stop running there, they actually will not guarantee that they will return anything of mine to me. It's a 100-percent zero-guarantee environment. A wonderful thing. And they don't charge me anything.
Now, if you're a major bank, how do you feel about that?
Knorr: We only have a couple of minutes left, so I have to ask you this: What do you think about Oracle buying Sun? Now they can bundle hardware and software like you do, just like the CloudBurst Appliance that you just announced.
Mills: Well, they can do that now.
Knorr: They've done that in their alliance with HP.
Mills: The problem with Sun is that for the past decade it's had no road map for Sparc, and sales have suffered as a result. Oracle's not a hardware company, so it doesn't have the wherewithal to deliver a road map for either design or fabrication. So Sun is exactly where it started. It doesn't gain anything as a hardware company by being acquired by a non-hardware company. And if you look at Sun's revenue, it's a hardware-based business.
So Oracle now has something that is not in any way compatible with what they do everyday. And they promised the street $1.5 billion in profit to boot. So what you're going to do is jack up all the prices in classical Oracle fashion and cut a lot of people in classical Oracle fashion. And all of that's going to lead to more customer disruption and likely create a lot of opportunity. For every action there's a reaction.