IBM Ups the Enterprise Ante (Continued)
Approaches to Enterprise Integration
Q: What do you see as the most pressing problem that IT customers are confronting today?
Swainson: Customers want to be able to preserve their existing set of assets and reuse them, and they also want to be able to build new sets of assets and integrate them with things they already have. Not a lot of people are writing new CRM systems or new back-end systems from scratch. But a great number of people are trying to take those existing systems and put them together in ways that lead to business efficiency. This whole notion of looking at your business systems as a set of business processes that could get optimized and enhanced is becoming the major focal point for business IT functions. Four or five years ago the focal point was to build or buy new functionality, but now it is about integrating what you've got and maybe doing some incremental new development to enhance the current environment.
Q: Is that mainly a function of the economic conditions, or is there something deeper there?
Swainson: Well, clearly the economic conditions contribute to this. People are looking for a surer return on investment. Most businesses are not in an expansionary mode; they're in a shore-up-the-foundations mode. In that environment, it makes sense to fix what you have done for the last 20 years rather than embark on new challenges. It's partly due to the fact that IT had been in an expansionary cycle, and people did a lot of new system development without doing the necessary integration back into their current systems. So, you're seeing people come around and clean up the mess, if you will. This isn't unique to IT. If you look at the business cycles that we typically go through, you see a period of contraction when people try to consolidate their gains.
Q: Discuss the many types of integration needs and technologies and provide a context for understanding how they relate.
Swainson: It became clear in the last several years that the superficial integration of applications wasn't achieving the business goals that people were setting for themselves. For example, in the financial-services environment, you might have achieved some degree of efficiency by building a demand-deposit system. But if you looked at the whole process from a customer perspective, substantial gaps continued between the different silos of automation and nothing allowed you to traverse the gaps.
The thing that changed our perspective on that was the Internet, specifically the need to allow customer self-service into these business applications. Prior to that, what we had were clerks and skilled people who sat on top of these systems and bridged them manually, sometimes writing things down on pieces of paper as they went from system to system to system. Well, that became untenable in a world of online self-service banking. Every industry has its analog of this; I'm using banking as an example.
What we saw were people starting to think about turning these processes on their sides and viewing them from the customer's perspective. How do I create a customer-loan process, customer-approval process, or customer-inquiry process and turn that into a set of steps in one process and not a set of monolithic processes?
Q: How does that map to IBM products and technologies?
The technologies that we have—MQSeries and WebSphere—were good for building the base of that, and we made a couple of key acquisitions—of CrossWorlds Software and Holosofx—that allowed us to accelerate our pace. These allow us to go to customers with a broad-based offering that lets them think about transforming business processes in business terms, not technology terms. With the slowing of the economy and customers' desire to see more value from IT investments, our business results in this area have been nothing short of spectacular. Our business integration revenues were up by 60 percent last year. And we think this is going to persist. There's a huge opportunity to displace existing people costs with automation.
Q: You've named a number of integration products, and in fact IBM has a whole portfolio of technologies in this area. What's the framework for relating your various integration technologies, and how can you better communicate that to the market?
Swainson: The Eskimos have 20 words for "snow," and we have only one word for "integration," but we think there are at least three meta-ideas here. One is something called application integration, which is the use of message-oriented middleware to connect things. It's actually the oldest of the techniques and it has the advantage of being relatively simple. Tools such as MQSeries play in this nicely. It's simple, reliable, and rugged, and it can form the foundation for higher levels of integration. For many customers who simply want to string a couple of applications together into a process, this is typically where they will start.
The second style of integration revolves around the notion of building a new application model to integrate existing applications. People typically use an application server such as WebSphere to do this, taking advantage of high-performance connectors such as JCA (J2EE Connector Architecture). What you're doing is building a new application model that fits over the top of existing application models, creating proxies for those inside the new model, and then using the new model to represent the application to the outside world. This is useful in applications such as banking solutions where you might have multiple back ends storing multiple views of something that you want to consolidate into a single view and present to the customer.
The third big style of integration that we see is process integration, where you're trying to take a set of business processes and create a flow across those processes with conditional logic that allows you to determine the next step depending on the outcome of the last step. That's useful in an enterprise view where you're trying to look at a supply chain from end to end. The supply chain will necessarily include general-ledger systems, ordering systems, and fulfillment systems, and they all need to live in harmony. So you need something to orchestrate or choreograph the flow over the top.
So we think those three styles of integration will be the way that people look at the overall integration space. People don't start at any one point necessarily. It is not atypical for customers to start at the application-integration level, but more and more we're seeing people start from the business-process side with a need to fix an inefficient supply chain or an inefficient demand-delivery system. I don't want to be in the business of telling customers how to fix their problems. I want to be in the business of having a solution, no matter what choice they end up making, because I'm not sure any one of these is wrong—it's simply appropriate in different circumstances of time, money, skills, and legacy investments.
Back to top
|