IBM is betting big on the broad-based acceptance of Service Oriented Architectures (SOAs) as the best way to help business users integrate application functions and data across the wild patchwork of software platforms they deal with every day.
At the center of IBM’s SOA strategy is WebSphere, which will play an integral role in helping fuel that strategy. And at the center of the company’s WebSphere strategy is Bob Sutor, a long-time IBM executive now in charge of overseeing the present state and future direction of WebSphere Application Server and WebSphere Studio product lines.
Before guiding the strategic direction of WebSphere, Sutor was IBM’s director of Web services technology, responsible for driving the cross-IBM Web services initiative. Prior to that he led the IBM-wide strategy for the development and promotion of key e-business industry open standards, including XML and Web services.
Sutor sat down with ED SCANNELL to discuss his views on the benefits of SOAs to corporate users, what IBM has in mind for the next generation of WebSphere to help nudge users more quickly over to SOAs, and how that product — code-named Vela — also helps users make their way toward an on-demand environment.
What do business users tell you is the most attractive aspect of SOAs to them?
Sutor: What is driving an interest in them is that companies have incredibly heterogeneous environments. Today, if you are talking about servers, then that also means you are talking about things all the way down to wireless devices. You have to expect that as time goes on, because of what you already have in your own company, because of mergers and acquisitions, or because of new software you are developing yourself, you are simply going to have a range of hardware and software that you need to get your most important work done. And the way to get the most efficiency out of that sort of environment is to try to make it look more homogeneous. The Microsoft strategy is to make it all homogeneous by telling users to use Windows everywhere and you are done. Well that is not quite how the world is going. Certainly Microsoft is part of any IT situation, but even more so Linux, mainframes, and other platforms are going to continue to be part of that. We see SOAs as a way of abstracting away many of the underlying details of how you actually get the job done.
What is IBM doing to quicken the adoption of SOAs among users?
Sutor: First, you educate them about Web services, which is one of the most practical ways today to create services that can be readily plugged into a SOA. Web services can work perfectly well to extend existing applications and, therefore, allows them to reuse existing software assets. It is not an all-or-nothing type of proposition. It means you can get into the game little by little. And this is what we recommend people do. They can pick a pilot project, learn about Web services, and then they can understand it in the context of their environment. Once they start getting some ROI they can enlarge that investment from there.
So what is generally the biggest factor holding back SOAs?
Sutor: If you look at global IT — and IT is a trillion dollar industry — there are millions of applications deployed throughout the world. There are tens of millions of connections among applications and companies. So in one sense, it is the sheer complexity of understanding how it is that your company is interconnected inside the company and then connected outside of it. So if you look at (SOAs) as something you have to do all at once, it is scary. So you say, OK, I am going to be conservative, I will stand back until this is better baked. I think what is also holding it back is the realisation for people that this is really the second generation use of the Web. Another part is we are still at the very early slope of the technology adoption.
What I am seeing, however, is the adoption of Web services and SOAs by particular industries. By that I mean financial services industries and sub-industries such as banking and insurance, but also government, health. But they all have to start standardising the descriptions of the Web services they expect their companies to use. If you have lots of one-off services you will never get to a full efficient SOA. This is the standardisation of how you talk to services and how they talk back to you. Some of this work has started because, after all, XML is five years old. People have started to understand how to represent their information as XML. But they also have to take the additional step of saying, look we are not just pushing information around — in particular industries you have business processes. So what we are really talking about here is using SOAs to implement business processes, and thatis what you standardise.
From a technical standpoint what direction is IBM going in with its server-based applications, such as WebSphere, to prepare for SOAs?
Sutor: We are adopting principles such as model-driven development from Rational Software. We are also adopting the view of services in the way we think about how we develop WebSphere itself, and how we do systems management and Tivoli too. Looking at what we will have during the second half of next year we have this project called Vela. It is really what we consider to be the next generation of application servers and, in our particular case, it will serve as the universal foundation for software group products.
How will Vela change your approach to developing other server-based applications?
Sutor: In terms of our software portfolio, we need to componentise them so that we can mix and match pieces from Tivoli, some from Lotus, and some from WebSphere. By doing so you not only get the core functionality but you can extend it as necessary. It means that we will be able to create software offerings that more closely match the immediate needs of our customers, particularly those in vertical industries. It also means we can more flexibly deal with things such as system messaging and enterprise-quality service. In the Vela timeframe, we are going to be driving things such as new look and feel around usability and much more support for the business cases involving the use of these different technologies. But more than just componentisation, (Vela) is providing a foundation for an On Demand operating environment. It will fit in with what we are doing with grid, autonomic computing, and it will be extremely obvious at that time as to how it will all fit into SOAs and what we can deliver.