CNET también está disponible en español.

Ir a español

Don't show this again

HolidayBuyer's Guide
Internet

Commentary: Peer-to-peer a business buzzword in 2001

Meta Group says that although Napster popularized peer-to-peer technology, businesses need to get to know the technology better before applying it.

Napster, which received a great deal of press in 2000, is one example of the allure and potential impact of peer-to-peer technology.
See news story:
Courts set rhythm for online music

Most attention has been focused on the music sharing and other intellectual property via peer-to-peer technologies over the Web. We also expect such technologies to play a role in future IT architectures. However, users should be careful of specific makers claims because significant technical challenges remain in some peer-to-peer areas.

The battles surrounding Napster initially point to a number of changes in publishing markets. Traditionally, the publisher was a choke point for linking content producers and consumers, and could enforce contracts. Because it is becoming increasingly easy for individuals to share intellectual property with one another in digital form, this model is beginning to break down.

We believe the music and book industries should move away from their traditional financial model of $15 per CD or hardback and think more in terms of 20 cents per recording or $5 per digital book download, making up the cash flow on volume. They may actually end up making more money using this method.

Peer-to-peer applications also raise significant copyright issues. Companies publishing music, books, and other intellectual content on a file-sharing model will want to be paid royalties for the rights to use material. They will increasingly look to digital rights management solutions to manage their rights to material published in peer-to-peer networks.

Looking beyond the impact on publishing of Internet music and file trading, peer-to-peer technology is also enjoying a rebirth as a potential business technology. While peer-to-peer is not a new idea--having first appeared almost 40 years ago in federated data sharing technology--it is showing signs of becoming a major theme in IT industry marketing.

Peer-to-peer takes several different forms. One-to-one models provide simple point-to-point sharing (such as video conferencing, file sharing transfers, or instant messaging). One-to-many models divide a task across multiple platforms (for instance, the virtual supercomputer, such as Beowulf). Many-to-many models involve users/endpoints sharing resources in a team or collaborative fashion (for example, collaborative gaming, chat, Groove Networks). Each of these peer-to-peer variants will be utilized with different levels of success and penetration within corporations.

One application of peer-to-peer technology that is often forgotten in discussion is machine-to-machine communications. This is an increasingly important feature in pervasive computing and some collaboration applications.

During the next few years, handhelds and information appliances will be configured to acquire and exchange information with a variety of individualized services, offered through intermediaries and from other devices owned and operated by individuals or organizations. This peer-to-peer approach to information aggregation and dissemination will allow devices to act on their owners' behalf, automating many mundane tasks (such as interacting between PIMs in a workgroup to exchange calendar information, coordinating free time available for appointments, forwarding documents to groups for review).

This type of capability is clearly beyond what typical peer-to-peer services, such as instant messaging, now offer, but we expect it to drive a significant portion of the market over the next two to three years and become a staple of the peer-to-peer market in four to five years.

Other applications have also begun to explore the potential of peer-to-peer technology. For example, some CAD systems use a form of the federated data technology to move complex CAD files as close as possible to the people who use them most. In this model, the system tracks requests for data sets. When it sees a pattern of requests for the same data from a user, it moves that data to the user's workstation or the server nearest to that person, along with the associated backup, access security and other management tools. It then informs the system that the "database of record" for that data has moved to a new address.

Although the peer-to-peer model appears compelling, it is fraught with problems, particularly for businesses.

Peer-to-peer does not work well in all applications and creates significant new management challenges. More importantly, it raises security issues, greatly increasing a system's potential vulnerability to computer virus attacks, for instance, as files are traded and systems bypass centralized security gates in favor of direct machine-to-machine interaction.

It can also change network traffic patterns significantly--generally not for the better, as more traffic will stay at the edges of the network and never make its way back through the core.

Peer-to-peer wreaks havoc on business networks. When clients suddenly become servers, traffic flows become more random and chaotic. Infrastructure developers need to understand the impact of peer-to-peer applications on the network before they permit them to be deployed.

We don't believe that a pure peer-to-peer model will become dominant in corporations.

META Group analysts Val Sribar, Steve Kleynhans, David Willis, Jessica Lajoinie, William Zachmann, Peter Burris, Jack Gold, David Cearley, and Dale Kutnick contributed to this article.

Visit Metagroup.com for more analysis of key IT and e-business issues.

Entire contents, Copyright © 2000 Meta Group, Inc. All rights reserved.