William Ruh recently became Vice President of Software Sciences & Analytics, at GE Global Research. Ruh has more than 25 years of industry experience in enterprise application integration and object technology. He served as Executive Vice President and Chief Technology Officer of Concept Five Technologies Inc., and as the Chief Technology Officer and Senior Vice President of Software AG Inc. He is the author of four books, including Enterprise Integration: The Essential Guide to Integration Solutions, published by Addison Wesley.
What are the big changes from an EAI perspective since you wrote Enterprise Integration?
RUH: I think we have learned a lot. We moved from the idea that being object oriented would solve integration problems to a state where we realize we need a mechanism to integrate. Everyone had different ways of doing it initially. Now we have come back, driven by Web technology and the Java world and open source, and have seen that those technologies have now been embedded back into these common development environments. Now services are running embedded so it is not just driving back-end business apps but front-end operational services.
Is there too much governance?
RUH: When enterprise architecture and then EAI came out there was a lot of feeling that you needed to have strong governance and standards. That was because the building of adapters was expensive. There wasn’t a standard way to integrate. Since then, a lot of legacy systems have added in standard interfaces, specifically Web-based services, and everyone supports that now. We have moved to where a lot of this is available off the shelf, even for a legacy system that you custom wrote. We have abstracted those applications from each other. We don’t have to build them anymore, and the Web services have become more enriched. Now businesses have opened things like their HR systems and business systems, and the pieces can all be reused.
So, governance is necessary but it is different from 2005. Now it is not about arguing over the Web service. The arguments are about what a contract should look like. People are looking for enriched capabilities, especially security. We are more focused on centralized governance but we find we don’t have to be so proscribed.
In a way it sounds like you are saying SOA has sort of disappeared...
RUH: I’m not sure SOA will go away. The service-oriented architecture is more than just integration. We won’t come up with a new architectural model any time soon because this one is working well. I think we have reached the point where the world has solidified, the standards are in place, and we know where to draw the line. This form of architecture has a lot of legs. The next decade will be about building out services.
What do you mean by that?
RUH: I believe this is the decade of services. We have had a movement now where the consumer Internet really drove things and allowed SOA to mature and the technology standards to mature. It forced businesses to open themselves up and participate. If you didn’t you were out of business.
Now I think we are moving into a world where the consumer Internet will continue but industry is going to participate in new ways like the smart grid. Services will be embedded in older industries that will transform them. The only question is how fast. For example in healthcare, with digital records, it is possible to develop new services for doctors, nurses and patients. Every industry will start to see more services embedded within hardware and sensors, and the ability to utilize those and offer more services to the customer base.
On a more granular level, what are the most serious or common SOA mistakes?
RUH: You have to be attuned to the fact that the only constant is change. Technology is always advancing. Architects who can balance out the building of services and remain ready to integrate what they have with something new, they will be the winners. Folks who look at things as if they will last 50 years and aren’t focused on change will encounter real problems.
Change occurs fast, and the ability to use these things in an agile way is important. Some people focus on building specific capability. They do it well and there is some ability to reuse, but that wasn’t their focus, so I see a lot of restarts in those cases where people have to go back and do it again.