Anyone who has lived through more than a couple of IT paradigm shifts can recognize a familiar cycle of hope, hype, backlash and clean-up. We are in the middle of another one, a Wild West rodeo of unbridled vendor-fueled enthusiasm over component-based computing that builds process-oriented solution functionality using standard Web services. The "application" is effectively deconstructed, decomposed, made to traverse multiple legacy code bases and platforms to re-invent itself as a composite. At risk is the established approach of building ERP systems as totalizing and monolithic environments. The cattlemen and the ranchers are at it again, battling for space, resources, and mindshare.

The hope implicit in the new technology is the promise of delivering agility and business functionality much more quickly and without the cost of assembling a complete application. The hype is that this somehow means that today's monolithic ERP solutions will slowly heave themselves into a tar pit of obsolescence and obscurity (to mix both metaphors and animal species). And the backlash will come once proponents of SOA-based composite applications "discover" the need for governance, security, compliance, and a methodology that ensures that the integrity and performance expectations are consistently delivered. (Fences and branding irons may come in handy.)

And after the backlash comes the clean-up. I was at a major IT conference earlier this year, where the CEO of a major CMDB (Configuration Management Database) company was openly and confidently ecstatic about what he saw to be impending chaos. He felt that this movement toward component-based composite applications would unleash a "Pandora's box" of legion and rogue Web service oriented applications. A poorly conceived mash-up is not a pretty sight, and he imagined that his company would be profitably engaged to disentangle.

He may be right, but perhaps it is not too late to save hope from inevitable dismay. There is no doubt that yesterday's totalized enterprise resource planning systems, or ERPs, are not addressing the needs of a rapidly expanding global economy and a volatile and fickle but highly computer-literate workforce and consumer base.
No one who has gone through a massive and lengthy ERP deployment for HR, GL, or CRM would want to do it twice in one lifetime. Monolithic computer applications, which only a software programmer can change, promise to do everything but end up solidifying or creating silos within business departments. While they scale, handle thousands of users, and are for the most part secure and reliable, we now find that this kind of business computing is also an obstacle to growth.
Without doubt, being consumer-oriented and process-centric would seem to be easier to do with components rather than fixed enterprise applications that are difficult to update and change. Using a combination of SOA (service oriented architecture) and BPM (business process management), many see the opportunity to make a sudden and revolutionary leap past ERP. Even enterprise application heavyweights like Oracle and SAP are placing bets, investing heavily in Fusion and NetWeaver to make sure that they have a place in the new world as well as the old.