IBM's John Swainson has a long and wordy job title that masks the importance of his role in the world's largest IT company.

As general manager of the application and integration middleware division of IBM Software Group, Swainson is the man responsible for WebSphere, Big Blue's set of application development and integration tools.

Middleware is one of the most fiercely competitive sectors in the market, with heavyweights such as Microsoft, Oracle and Sun Microsystems and specialists such as BEA trying to establish themselves as users' first choice for developing new business applications.

Swainson talked exclusively to Computing about the challenges for the WebSphere product range.

The definition of WebSphere has broadened over time in response to customer requirements. The original WebSphere was a J2EE application server and that remains the core of our product family. We always understood the value of WebSphere was giving customers transactional systems that could be integrated with their legacy systems. We weren't necessarily talking about 'greenfield' systems or building standalone solutions, which is what a lot of our competitors are doing. We always had the view it all had to connect together. That made our sales process a little longer, and took customers a bit longer to get stuff going. But then it started to build.

The second important thing that happened, around 2000, was we realised customers needed more than simply a product, an application server or a set of tools. What they needed was an integration platform, a way of building applications that could integrate with their existing environments. So we incorporated the MQ family of products, rebranded as part of WebSphere. At the beginning of 2002 we acquired Crossworlds because they had very strong process modelling and execution technology based upon Java. So we started to broaden the platform into business integration.

We also realised a lot of customers were building their own portals, writing user interfaces on the application server with J2EE, but they didn't have a set of tools or a framework that let them do that. We realised that was an opportunity and a customer requirement to create a portal offering in WebSphere.

What we have today is a family with three major elements - a capability to build applications, to integrate applications, and display those applications through portals. That three-legged stool has given us a very strong competitive position versus some other guys who had one or two of those dimensions to their products.

There might be others who would argue on this, and I am biased, but it would be fair to say WebSphere is at the core of IBM's on-demand strategy.

On-demand is a business strategy not a technical strategy, aimed at making businesses adaptive and flexible and able to structure themselves in a way that gives competitive advantage. Why doesn't everyone do this today? The problem is today's business systems have been written around different models, typically organisational, and almost every system reflects the state of an organisation at a point in time. That tends to set in concrete the processes the business uses. The notion of being able to step above those and integrate the processes and take things and reuse them in a more flexible fashion becomes really core to a business becoming more flexible, and the technology we use to do that is WebSphere.

The essence of WebSphere is the notion of integrating new and old things and giving people a combined view, something that looks new but is composed of things that are old.

We view them as formidable competitors, but we do have this complex relationship with them that goes back many years.

We're competitors, IBM is their largest reseller, we're suppliers to each others, we're technology collaborators. There's probably more complexity in the relationship than between any two companies in any industry. It's enormously challenging to keep straight and to not let the fact we are fighting each other every day bleed over into the other parts. Customers force us to do that because Microsoft and IBM are both a reality in customers. It's not about our problems, it's about customer's problems. They need us to work together to solve their problems. They need Windows-based systems to work with WebSphere-based systems.

The whole web services initiative started as a triumvirate with IBM, Microsoft and Ariba - who dropped by the wayside - and we were getting this strong message from customers that we had to do more. Web services emerged out of that. Out of that competition, ultimately the customer wins.

We're going to try to do that as naturally and seamlessly as we possibly can. As we add some of these new technologies it is extending the programming model in a coherent way so customers can take advantage of these capabilities.

So when you build a WebSphere-based application, one of the menu options becomes: 'Do you want to create a web services interface for this?' WebSphere Studio has already captured enough information to automatically generate the XML needed for the interface definition. It feels like part of the whole process of building applications. The same is true with grids.

If web services defines interfaces, grid defines how applications sit on top of computers. Web services is at the front end, grids is at the back. What we try to do is make sure we abstract that, so people don't need to understand a different set of commands or interfaces. We want it to be consistent and coherent with what they are already doing so they can take advantage of these things.

We're trying to not make technology into the focal point. It's not about the latest whizz-bang thing, it's about helping people to achieve business value and build applications faster and run those applications more efficiently. It's not about the mechanics of writing an application or administering a system.

We get so enamoured in our industry with the glory of the technology that we forget that technology doesn't matter. It's about what the customer does with it and how he achieves better business results. We're trying to introduce new inventions in a way that is evolutionary for our customers.

You will see web services continue to evolve and the current standards will be completed. You'll see more of the evolution of grid computing as it starts to become an essential part of WebSphere, and becomes part of the fabric of how you deploy applications.