Distributed computing appeals to e-business

Back before the rise of client-server computing, one of the holy grails of computer science was distributed massively parallel computing.

Back before the rise of client-server computing, one of the holy grails of computer science was distributed massively parallel computing.

Under this architecture, multiple types of computers -- regardless of operating system -- would be able to work on the same tasks over a network connection.

Although it became possible to do this in various laboratory settings such as MIT, practical commercial applications for distributed massively parallel computing solutions were in short supply.

In fact, about the only people interested in this approach were academics who couldn't find enough grant money to afford time on a supercomputer. As a result, distributed massively parallel computing became known as the poor man's approach to supercomputing.

But if we fast-forward to the present day, it appears that the rise of e-business is beginning to make the concept of distributed massively parallel computing attractive once again.

In considering the future of e-business, we can easily picture a need for hundreds of servers to work in concert as digital marketplaces become increasingly complex. After all, a digital marketplace not only has to process transactions; it also has to coordinate information from all participating systems so that all the buyers and sellers in the system have access to all relevant information in real time.

We're still a few years away from having the foundation in place to support that type of Adam Smith nirvana for digital capitalism, but companies are at least starting to move in that direction.

Microsoft, for example, has outlined its .Net strategy, envisioning loosely coupled servers connected over the internet. And Hewlett-Packard has created the e-services architecture that leverages its core e-speak integration engine.

More importantly, some companies have actually begun to deliver products that incorporate distributed massively parallel computing technologies.

One of those companies is Plumtree Software, a developer of corporate portal software. In release 4.0 of that offering, Plumtree added a massively parallel portal engine that processes requests for data from multiple servers running Plumtree's portal software in parallel.

This is significant because it cuts to the core of e-business. Instead of requiring a user to request information from multiple individual systems one at a time, a user can simultaneously collect data from all the systems at the same time.

For example, a customer participating in a digital exchange could instantly see data on a specific set of products from multiple vendors instead of having to search each individual vendor site to compare features and pricing.

Of course, this example presupposes that every system is running Plumtree's software. That may be a viable approach for companies such as Proctor & Gamble, which as a Plumtree investor and customer, is using this software to help create its own business-to-business e-commerce network.

But for this concept to become the de facto way of doing business for everybody, we need this rocket science technology to move into the mainstream. And the only way that is going to happen is if the major operating system providers move to embrace it and can come up with standards to support this capability across multiple platforms.

Until that happens, the future advancement of e-business architectures is likely to be driven more by small independent software vendors such as Plumtree than by household names such as Microsoft, HP, Oracle, Sun Microsystems or IBM.

And given the speed at which those companies usually respond to changing customer requirements, that is probably a good thing.

Michael Vizard is the editor in chief at InfoWorld.

Join the newsletter!


Sign up to gain exclusive access to email subscriptions, event invitations, competitions, giveaways, and much more.

Membership is free, and your security and privacy remain protected. View our privacy policy before signing up.

Error: Please check your email address.
Show Comments