The Web services grid

Pervasive computing will come, and when it does, e-business will be smooth.

In case you haven’t noticed, in the past few months I’ve tried to revamp this column. Whereas in the past I was content to point out an interesting social or political trend related to technology, I’m now narrowing the focus to the technology itself. And I don’t just talk about subtle aspects of technologies that we all use. Now I talk about tech we don’t yet use, but we soon will–i.e., emerging technologies. I have shied away from these topics in the past because I have often been burned by other tech publications, whose “next big thing” quickly made the next big thud. But our readers want more emerging stuff. So I will do my level best to discuss technologies that we will all be using in six months to a year.

In light of the new attitude, I’m embarrassed to admit that I have not been looking forward to this month’s topic–e-business. Frankly, there have not been a lot of good stories to tell in the e-business arena. E-business has become a dirty word these last couple of years. The hype surrounding can’t-miss trends such as enterprise application integration (EAI) and customer relationship management (CRM) has deflated faster than you can say “dot-bomb.” So I have been dreading this issue for months.

But recently a couple of emerging trends really got my juices flowing: Web services and grid computing. I have long considered them separate trends, but a bullet point from our cover story this month by Nelson King got me thinking about them as two parts of one whole. He claims that Web services are expected to play a major role in “massively parallel (grid) computing.” When I stopped to ponder this, it made sense because, while we have been trying to find business uses for grid computing, we never had the applications to do it–until now. The more I thought and read about it, the more excited I got about the possibilities.

When looking back at the failures in e-business technology over the last couple of years, the problems can be organized into two areas: a lack of software standards and a lack of hardware power. Web services offer modular open-standard software built on XML. Open standards will enable organizations to develop collaborative projects without hiring a legion of custom programmers to connect nonstandard systems. With grid computing, organizations that need the power of a supercomputer to mine the data in their shared databases will be able to do this without adding a supercomputer to their server rooms. Similar to what users do when they hook into [email protected] (the search for extra-terrestrial intelligence), unused processing cycles in all the PCs in organizations will satisfy the hardware needs of e-business ventures.

A few years ago, everyone thought EAI was the future of e-business, but it cost more to implement the systems than the business benefits justified. Often, by the time an EAI solution was coded and implemented, the venture had dissolved or changed so radically that the solution was worthless. In the age of mergers and acquisitions, the cost of custom coding technology to fit an e-business venture is too risky. Web services are supposed to change all this by giving programmers modules that can be plugged in here and there quickly and efficiently. If a venture changes, unplug one module and plug in another. Web services are ways of making the applications as nimble as the companies that need them.

The other inhibitor of EAI was speed. When you integrate two applications and all the data that they handle over a network, you’re bound to suffer with latency. The common approach to this problem is to throw a lot of hardware at it. Get bigger and faster servers; get gigabit Ethernet with all the hardware that goes with that; get faster Internet connections in the case of distributed ventures or partnerships. Few projects bring in enough revenue to overcome these costs. But since Web services run over common IP networks, clustering technologies including version 3 of the Globus Toolkit can take advantage of existing distributed resources to handle the hardware demands in data-intensive Web services ventures, and then some. Cheaper software plus cheaper hardware will enable organizations to cost–justify a lot more e-business than they currently can.

CRM is perhaps the most maligned e-business technology category. But the concept makes perfect sense: Use technology to better understand what customers need and to help cater your sales, marketing, and development efforts to those needs. But CRM is notorious for not delivering on these promises. Our CRM feature in this issue outlines a common CRM chain reaction. Vendors sell solutions to companies that don’t have the talent to implement them; companies underestimate the complexity of the data-gathering task; bad data combined with poorly implemented solutions leads to poor customer satisfaction. CRM was supposed to improve customer relations while reducing customer support and service staffs. Companies reduced their staffs before realizing that their CRM solutions were not working. Ergo, CRM often resulted in lost business rather than increased business.

The more I think about how Web services and grid computing can help redeem CRM, the more excited I get. One of the key problems with CRM is integrating lots of disparate data. Data can be distributed across sales, customer service, customer support, shipping, billing, collections, and so on. All these data repositories say different things, but combined, they provide a complete picture of a customer. Without all pieces of the CRM data puzzle, the applications provide incomplete analysis and the above scenario rears its ugly head.

Well, what if you could just develop Web services that integrate all this data and deliver an analysis to every point of customer access in the company? Just set up a peer-to-peer IP network between the players, run a Web services application that links the data, and use the massively parallel processors in the PCs to crunch the numbers for the Web services analytics application. Instant CRM. Instant return on investment.

When will this nexus of technologies turn e-business failures into successes? Certainly before we explore e-business in next June’s issue, if not before. In any event, I’m actually looking forward to the next time we cover e-business topics in depth. And I can now freely talk about e-business without worrying about the profanity police.

Related posts

Leave a Comment