Most companies need to figure out how they can more effectively exploit the technologies they already have in place, rather than what new IT systems they need, according to Universite du Quebec and New Jersey Institute of Technology professor, Stephane Gagnon. He specialises in the management of IT and teaches both management and computer science. In a recent interview with IDG's Tommy Peterson, Gagnon discussed IT-business alignment and related topics.
We've heard a tremendous amount in the past few years about aligning technology with business goals. What progress do you think has been made?
Stephane Gagnon (SG): It's typical for fast-growing companies that getting the house in order and using IT to do so have been successful. The question that's more important is, what are they doing with that technology? ERP systems, for example, are barely leveraged. Name me [any] companies that use all the functionalities they pay for. The answer will be, unfortunately, very disappointing. And that relates to management issues - to what extent did they plan ahead to use the functions, to leverage the functionality of all the applications they bought? That's the alignment we're talking about. Being successful in deploying systems and re-engineering processes is OK. But fully leveraging the technology, that's where the success is still lacking. That's why the emerging technologies of service-oriented architectures (SOA) and business process management (BPM) come into play and fit into the solution now. They provide a chance to leverage further and in new ways all the functionality sitting there, doing nothing, and tie it up together in properly redesigned business processes. Those companies that have succeeded at deploying basic enterprise systems need to leverage them better now.
So it's not the vendors lagging behind user needs? Vendors can't do more than what the end users are really prepared to do. What products should vendors come up with?
SG: They're pretty much coming up with all we need so far. We have all the technology we need to accomplish, for example, a complete redesign of a multibillion-dollar company - that's not an issue. We have everything we need to run this company from the ground to the top and make it a real-time enterprise. The question is, how do we want to leverage that technology, and are we prepared to leverage it? Right now, however angry senior management may be with the IT departments of this world, and to what extent IT departments are in turn angry with vendors, I think it's merely to hide something. This is just a game to point blame. Whenever we point blame to vendors, it's a false problem. The actual problem is leveraging technology. We have plenty of technology. We have plenty of vendor competition as well. In the SOA space, for example, we have plenty of large vendors, plenty of smaller vendors and plenty of risky adopters, who are ready to adopt the stuff early. The trick is to do it right.
The heterogeneous environment is generally acknowledged as a fact of life. Why isn't there more focus on integration issues, since those problems loom as large as they ever have?
SG: Perhaps it's a problem with 'integration' as a word. Technically, 4-5 years ago we were still in the enterprise application integration (EAI) era. This was merely a message-queuing technology allowing the integration of systems with message-passing interfaces. Today, we see a new trend, leveraging component-level integration, with SOA especially, and BPM that rides on the SOA. The integration problems in companies have changed. It's no longer a matter of integrating applications and data. It's primarily now to redesign processes and leverage component services. The reason you don't hear so much about integration as you did in the past is because it's now being addressed by runtime platforms upfront; it's no longer a separate module like it was when you were buying MQSeries or Tibco. That may be the reason technically that we don't hear as much as before about integration projects. Seventy per cent of new IT investment projects involve integration problems and most likely leverage some integration technologies in some ways.
Have we turned this major corner, and are we doing SOA from now on - or at least until something else comes up to change that landscape?
SG: "That's what we do now" is always a risky phrase. I wouldn't say we should only implement such technologies now. Whenever it's time to develop a new application or to maintain an existing application and modernise it, and especially to redesign processes, in that case, yes, [SOA is the choice]. Web services as a component mechanism, as opposed to message queues or other component mechanisms, and using [Business Process Execution Language] as a standard to develop [and deploy] processes and using a runtime environment such as those we were listing a minute ago, these tool kits are certainly preferable to the ones we previously had. If we're talking about routine maintenance of applications, which still remains quite a bit of an IT investment, then obviously SOA and BPM are not the answer right now.
Why isn't adoption going as rapidly as some had predicted?
SG: We're on the foot of another trend where BPM suites and SOA runtime environments are going to pick up. The leading indicator would be the investment in [quality assurance]. What QA brings and what enterprise portfolio management brings is transparency of application infrastructure. It tells you specifically which portions actually perform well or are, unfortunately, the bottlenecks in processes. It then tells you precisely where to go to surgically modernise or modify applications, or re-engineer processes or entire organisations, and in the process tells you what the potential is of implementing BPM and SOA. This trend on QA and testing over the past two years is just the start of a real trend to SOA and BPM.