P2P 'fat clients' not for Internet collaboration

In today's online world, most of us live by the philosophy "Have browser, will travel." Often we can travel as far as we like on the Internet with just this one thin client. And many of us prefer to travel light, resisting efforts to add new client applications to our desktops or to bury our browsers under applets, plug-ins and other new software layers.

So it's puzzling to see the re-emergence of the Internet "fat client" in the form of peer-to-peer environments. In all the attention given to peer-to-peer architectures, we overlook the fact that these environments often require each user to download, install, configure and update multimegabyte peer software packages. In itself, the peer software, which corresponds to an individual desktop, usually supports presentation, application logic and data-store functions.

For sure, there are great differences among today's Internet-centric peer-to-peer environments. Not all peer-to-peer products require hard-disk-hogging software downloads. Many of them rely to varying degrees on server-based processes such as search engines, directories and message routing. And there's no denying that peer-to-peer technology has its killer (albeit sometimes controversial) applications, most notably music file swapping and compute-cycle brokering.

But Internet-spanning collaboration environments are not well suited to the peer-to-peer architectural model, in spite of all the attention paid recently to Groove Networks Inc.'s flagship offering. It's asking too much for people in different organizations to standardize on one vendor-proprietary peer-to-peer collaboration environment such as Groove and to download and install that particular vendor's peer software. Also, what happens when users participate in multiple Internet-centric peer-to-peer-based collaboration environments? Do they need to download and install multiple vendor-proprietary peer software packages on their desktops and laptops? How can users access peer-to-peer environments from PDAs and other resource-constrained handheld clients?

In today's business environment, collaboration is very fluid. Teams rapidly form and dissolve across time, space and corporate boundaries. In a business-to-business environment, practically the only collaboration-relevant common denominators shared by all potential team members are access to the Web, Internet e-mail and the telephone. Where online "teamware" and other Web services are concerned, the browser is the preferred client. Peer-to-peer collaboration environments such as Groove cannot achieve much marketplace traction in the face of this unstoppable trend.

Groove's marketing machine has benefited greatly from the fact that company founder Ray Ozzie is also the inventor of Lotus Domino/Notes. With Ozzie involved, the IT world half-expects lightning to strike twice, as if he and his new firm have some magic touch that will propel the peer-to-peer model to ubiquity in the Internet collaboration market.

But that just isn't happening. Groove is one of the few Internet collaboration platform vendors to base its technical architecture on peer-to-peer. Most commercial Internet collaboration products rely on browser interfaces and place zero to minimal configuration requirements on the browser. The leading offerings in this area are browser-oriented products and services from vendors such as eRoom Technology Inc. and WebEx Communications Inc.

These browser-oriented products offer functionality similar to that of premises-based groupware environments such as Lotus Domino/Notes and Microsoft Exchange/Outlook. Browser-oriented services can support requirements for business-to-business communication, coordination and resource sharing. These Web-centric products support user roaming while providing full access to collaborative functionality. And to varying degrees, they can supplement the functionality of organizations' internal collaboration environments.

From a supportability standpoint, browser-oriented collaboration environments have many advantages over peer-to-peer products. The former are easier to configure, administer, personalize and support than peer-to-peer environments, which require software installations on every client.

If there's a clear niche for peer-to-peer collaboration environments such as Groove, it's in support of small, stable, self-contained workgroups of people who interact frequently. In other words, peer-to-peer, as its name implies, may be the appropriate collaboration environment for true peer groups, in which participation is by invitation and users trust each other enough to open up their file systems and desktops to mutual perusal.

But in all likelihood, stand-alone, peer-to-peer collaboration products won't survive long. They'll migrate architecturally toward browser-oriented environments or be absorbed into the leading, premises-based groupware environments -- Lotus Domino/Notes and Microsoft Exchange/Outlook -- which are themselves becoming more browser-oriented.

Thin is in -- thin clients, that is -- and will remain a dominant application architecture of the Web-computing era. Recent interest in peer-to-peer products and services won't alter that deep-seated trend.

This story, "P2P 'fat clients' not for Internet collaboration" was originally published by Network World.

Copyright © 2001 IDG Communications, Inc.

7 inconvenient truths about the hybrid work trend
Shop Tech Products at Amazon