While attempting to solve an integration problem at Sun Microsystems, Jim Green realized the existing technologies he had at his disposal were inadequate. After completing several projects, he also realized the same code was going to be written repeatedly. Clearly, a product to help with integration was needed. So, in 1995, Active Software was formed. What started with just nine people is now a public company, worth more than $1 billion, and a leading player in the booming market for enterprise integration technologies. eAI Journal’s editor-in-chief, Tony M. Brown, recently spent time visiting with Jim Green, founder and CEO of Active Software.

eAI Journal: What was Active Software trying to achieve back in 1995?

Jim Green: We didn’t know exactly what to call what we were building initially, but we were building one of the first message brokers, a name coined by GartnerGroup in 1996. So there were a couple of years where we could only describe rather than name what we’d built. We were the first company to put Java tools on middleware. That was considered pioneering back then. We were one of the first companies to build adapters. Back then, middleware just had interfaces. A fundamental change occurred in that people used to build their own applications, and so you could provide a middleware layer — an Object Request Broker (ORB) or a Transaction Processing (TP) monitor — and people would incorporate that into their programs. Increasingly through the 1990s, IT departments started purchasing applications, which meant they were no longer doing development and could no longer incorporate anything into applications. So, now you had to tie these applications together by putting some standalone product in between the packaged applications. That created a new market.

eAIJ: How has the market evolved since then?

Green: People used to be interested in providing tight coupling between their applications. As they became more aggressive with their programs, they started talking about multiple applications communicating simultaneously. Fundamentally, they were trying to implement some sort of business process. The process determined the number of application connections, not the technology. The technology had to be much more adaptable, and that’s where hub-and-spoke architecture came from, and where publish and subscribe started to take off, and where the idea of information flowing around the network in real-time began. Many people don’t realize that, throughout the 1980s and early 1990s, the predominant model of communication was not passing information. There’s been a paradigm shift away from building distributed applications to moving information between applications. That was fundamental. We were pioneering that with great uncertainty as to whether the industry would agree. Today, it’s accepted as a matter of fact. Later, around 1998 or so, the level of discourse was lifted to the point where people were talking about implementing entire business processes, and so the idea of process automation came into play.

eAIJ: How is e-business driving the market?

Green: What we’re seeing today is this idea of breaking down the definition of an enterprise. The boundaries of a company have now dissipated. The constituency for IT used to be company employees; now it’s employees, customers, partners, and everybody who interacts with the company. Many IT projects are now interested in providing better communications to all the people who surround the corporation. And that’s caused a shift from Enterprise Application Integration (EAI) into the more general scope of e-business, which provides more effective real-time communications between applications within and beyond the enterprise.

eAIJ: What common problems are your customers trying to solve?

Green: Sometimes, it’s an internal IT project to make the company more efficient and, increasingly, it’s a movement toward an e-commerce- or e-business-type arrangement. We see many more companies being reactive — trying to react to the movement of the Web from just a Web server marketing presence to more of an interactive activity with their customers or suppliers or partners. Many people are just trying to keep up; nobody wants to be left behind. Speed is everything. There’s one overwhelming thing that people want — and that’s speed of implementation. They also want ease of use. If you lower the skill levels and structure the product appropriately for the project team, it facilitates people being productive and making great progress in a short time. We also hear a lot about total life cycle cost. People want to move quickly, yet reduce their maintenance costs and their total life cycle costs. So it’s important that systems have an extensible architecture. It needs to be flexible enough to support changes over time — so maintaining overall lifecycle costs and ensure that can maintain the system and observe how it really works in operation.

eAIJ: Integration projects can be notoriously complex. How do you go about deploying a solution?

Green: We have a clean model. We have the notion of an integration process that consists of multiple components. We capture the components with well-defined boundaries and we can reuse the components for multiple processes. If you have a particular business rule or algorithm, you can now capture it as a component and move that component various places. A bunch of information there collectively implements your system. The information includes the business logic that resides in the components and business events or information. That might be in application-specific formats or in XML format. The information also includes security parameters, configuration of the adapters, network topology, the brokers, and administration tools. With our product, when you get that working in the Research and Development (R&D) laboratory, you can press a button and all that information is collected into what we call an interaction. An interaction is nothing more than a packaged integration process that can be transported from one location to another.