Complexity Threatens to Take the "Byte" out of E-Business. (Tech Topics)

When it comes to taking novel technologies from the boardroom to the portal, identifying the obstacles are intuitive enough. Let's take web services, the subject of this month's special supplement. Potential issues can crop up from 1. the unknown; 2. incomplete standards. (Such as field-related standards within a specialty area, for intance mortgage syntax); 3. an emerging base of vendors too new to have much of an industry reputation (and so hard to choose from); 4. the challenge of finding the right pilot to determine the scope of utility. And so on. But mature web-based systems pose their own challenges, those of ownership. Moreover, they aren't as bug-free as their years in the market might suggest. As the glare of the media spotlight and, perhaps, organizational attention, drifts to the next new thing, the challenge with the "old IT stuff" is keeping perspective and continuing to ask tough questions. Is CORBA (common object request broker architecture) cutting it? Are Java beans, Websphere servers, and Linux boxes or newer generations of core processing technology breaking down silos as intended? Are customers are truly delighted or merely silently disapproving due to flaws in your bank's website design? Interesting questions one and all. Because in the gradual evolution of technology, the new replaces the old--sort of. The result of this can be unintended complexity that makes management of IT tougher than it needs to be, affecting costs and performance. And, it's a phenomenon that ought to be factored into new application development on the web. Robert Shear, chief technology officer, at Grey Stone Solutions, Woburn, Mass., points to the tendency of technology to be accretionary at most organizations. Part of the problem, of course, has been the radical change of the technology resulting in a cluttering "cache" of museum pieces lying in wait at most institutions. "From 1975 to 1985, online systems and structured programming were the norm," Shear explains. "From 1985-1995, relational databases, client-server architecture, and object-oriented computing were introduced and began to get somewhat established. From 1995 to the present you've had an increased emphasis on networked computing (i.e., distributed computing) and internet-enabled computing," Shear adds. Deployment and operation of said technology, of course, was and is easier said than done. "Hardware advances come first, then applications tend to catch up," says Shear, "and that lag creates a lot of the challenges an organization faces." Layered like sediment in a fossil, one system gets superimposed on another, but unlike a fossil, old layers don't just flatten out and die. They add complexity--compounded each time an application is added to the enterprise. Complexity also adds to the "dial-tone" (i.e., maintenance) cost for organizations and boosts the developmental expense. This is especially true where the web is concerned, because (whatever the specific application or objective) the requirement is typically one of data aggregation. This requires either messaging to retrieve data at new back-end systems or-in the case of web services-passing off data "forms" to get work done in a distributed way. Promises and challenges If you're a student of computer history or just keep back issues of trade publications like this one, then you know that integration was supposed to have gone the way of the typhoid fever-only it hasn't. Enterprise application interface (EAT) integration methods, for instance, can consolidate and simplify the computing environment, but the projects also tend to be expensive (though speedier and cheaper than hardwiring "end points one at a time). The cost problem is significant enough that web services vendors, warts and all, be used over the next five to ten years to do the job instead, says Ron Schmelzer, a senior analyst with ZapThink, Waltham, Mass. …