Read part two.Looking back at the past three years of SOA, what has surprised you?
SOA has shifted from being pure technology to be a combination of business and IT. We saw that shift starting to happen back in 2006. That's when we had a discussion at the senior level with Sam Palmisano, the [IBM] CEO, about the shift coming in the market with the business side coming forward. That's when it was decided that they would bring me over to the consulting side, but I kept my SOA responsibilities. So now I could drive it not only from the technology side, but also focus of the business platform.
I think what's surprising is how widespread SOA is. Usually with technology it starts with one industry, such as financial services, which has more money to invest. Or it starts in one particular geography, and usually if it's technology oriented it tends to emanate from the U.S. But in this case every industry we look at, every geography we look at, has jumped all over SOA. I think the reason is that for the first time we started talking about technology in terms of business services in a way that the line of business understands. They understand that if they combine services in new and innovative ways they can drive new business processes, lower costs, create new products and ultimately improve the business.
That has happened and the technology is finally there with open standards for Web services. The other thing that happened is the Web caused the globalization phenomenon. So it's almost like the perfect storm. Technology is finally there and open standards-based. There's market globalization and businesses understand they have to change to stay competitive no matter what industry they're in.
So it's been surprising how quickly SOA spread across every industry. It accelerated on the business side much faster than we originally thought it would. It's gone from something that was perceived in the early days as just technology that the line of business didn't want to talk about. Now, they want to talk to us. What still needs to happen in the next two or three years?
I think skills is the number one issue for a lot of clients. Skills in terms of the technology and how to build SOA. SOA is really an architectural style, an architectural pattern. Therefore you have to compose your services in a different way. The architectural principles are different from the old fashioned monolithic application. Therefore there's a new set of skills required. To build SOA you have to understand the decomposition of services. You have to decompose them at the right level. On the business side, most business don't understand their processes. They understand what they do and why, and they understand the input and the output. But they don't know what goes on in all of the intricacies of their business processes and how they all interrelate. Now they're starting to have to decompose and really understand their business processes at a much deeper level than ever before. And all that takes skills. For IBM global services that's a huge opportunity. I can bring in industry expertise and process expertise to help the client. Most of our SOA engagements are not the client saying come on in and do this for me. Almost all my SOA engagements are come on in and help me and work with me in a partnership so I can accelerate the adoption of SOA both from a business and an IT perspective. Do you find when you go into those engagements that IT doesn't have near enough visibility, and if so what needs to change?
If you look at how we traditionally measure IT, we used to look at service levels, the uptime of the network, what's the uptime of the application. We talked about it in terms of the infrastructure. The next step is all around what's happening with the business process and business demand. IT starts to look at how quickly can it do something. What's IT productivity? Can IT generate more revenue? Can IT take costs down? These are starting to become business metrics versus IT metrics. You don't measure IT by availability anymore. We hear a lot that people think SOA is about getting the right technology, such as an enterprise services bus (ESB), since moving to global services do you find yourself having a different perspective on technology?
We always tell clients start with what problem you are trying to solve. Don't try to solve the integration between business services. If all we were trying to do is connect and reuse services, you need an enterprise service bus. That's kind of the core technology. But if what you are trying to do is transform your business processes, putting in an ESB, you'll have an elegant infrastructure, but you won't transform your business.
When talking to CIOs I always use this analogy. I'm a car nut. I could buy a Ferrari, a great car, and I could put you in that Ferrari race car and you wouldn't win any Formula One races. I could take Michael Schumacher, one of the greatest race car drivers in the world, and put him in a Hyundai and he wouldn't win any Formula One races either. You need the combination of the right driver and the right car. In SOA, the car is the technology and the driver is the business. You can have an elegant infrastructure and not get any business value.
Don't focus on the technology first, focus on the business value. If you decompose your services and then need to connect them, then you have to have an ESB. But an ESB is a commodity. It's like an engine in a car. Yeah, I've got an engine in my car, but I don't buy the engine separate. If a client says all they want to do is implement an ESB, I consider them to be at the very early stage and there isn't that much business value there. Focus on the problem you're trying to solve, not just the technology. I tell people, technology is an enabler, and it's a great enabler.
In part 2 tomorrow, LeBlanc discusses issues of scalability and reliability and why some SOA implementations are hitting roadblocks in adoption.