Commentary: Sun's Jxta is a Joy ride

Sun Microsystems' Project Jxta is an interesting attempt to create a set of totally platform- and language-independent protocols for peer-to-peer networking, amounting almost to an operating system for the Web.

As it currently exists,

	See news story: Sun's Jxta hopes: D?J?va vu all over again?

Jxta is much too immature to be used as a development platform--for instance, it is missing both security and a directory of resources--so companies should not download this code and turn it over to their development groups. However, Jxta is exactly the kind of technology that corporate IT research groups should be experimenting with.

We believe that Jxta--unveiled in a Webcast this week--is easier to understand as a personal project of Sun co-founder and chief scientist Bill Joy than as something related to current Sun business initiatives. Joy himself acknowledges that Jxta is based on a vision he has had for 25 years. And it certainly does not fit into present Sun marketing directions.

For instance, Sun trumpets Java as the sole development platform that its customers should use, while Jxta is platform-independent. Although the initial Jxta implementations are in Java, Sun has promised implementations in C. In addition, Jxta is to evolve as an open-source project (via Jxta.org), and Sun has until now not participated actively in the open-source movement.

However, Joy's responsibility at Sun is to develop interesting new technologies, regardless of whether they fit with Sun's present business direction. In any case, Jxta's immaturity means it shouldn't be considered a product, and thus does not really conflict with Sun's current Java emphasis.

When Java first came out as a research project, it too was missing some key pieces such as security, and it was tangential to Sun's direction at the time. Jxta should be considered a very early infrastructure direction. Sun and others have ample time to fill in the gaps.

Jxta what is P2P good for, anyway?
Security and a directory system for peer-to-peer (P2P) computing are solvable technical problems. The larger question is what people will use Jxta for once it is developed. Inside the firewall, no one has developed a killer application for P2P. The few Meta Group clients who have bought licenses for P2P technology have left those licenses sitting on the shelf, because they have no idea what to do with them. Outside the firewall, you can point to Napster and Gnutella but little else at present.

However, we see a latent desire among many consumers for P2P applications, even if they are unsure of exactly what those applications would do. On the Web, Internet service providers such as America Online would like to develop P2P solutions for their subscribers to use. The current absence of an obvious "killer" application for P2P does not mean that such an application will not appear in the future.

A decade ago, no one understood how people would use the Web, and now the applications built on it are transforming businesses, markets and personal lives and have driven both the economic boom and bust of the last five years. A P2P technology such as Jxta--which proposes protocols for peer discovery, pipe binding, routing, and so on--has potential on many levels on the Internet, not just on the end-user level. It is as much a technology for network devices to use to communicate with each other as for humans to use across the network.

On the other hand, many promising technologies--for instance, bubble memory and videotex--have faded into obscurity because they failed to find a real problem to solve.

Sophisticated vendors such as Sun must treat their businesses like an investment portfolio. Parts of the portfolio support the business in the near term; other parts are like a venture investment in the future. Five out of 10 of these venture-type investments will fail outright, but two or three will be huge. Give credit to Sun for having venture-type ideas such as Jxta in its portfolio.

Jxta sitting and a-waiting
Jxta is not going to appear in enterprise infrastructures anytime soon, and companies should not waste their resources trying to develop corporate applications on what today is a very immature, incomplete platform. However, corporate research departments should download this code, subscribe to the Jxta Web site, and try it out as an interesting technology that they should be watching.

If Jxta ever does take off, it could become a pressing management concern for IT groups. True P2P applications, which require no server, could easily spread from end-user to end-user through the enterprise, until one day the IT organization wakes up to find that half its network bandwidth is being consumed by applications of which it has no knowledge and over which it has no control. Large enterprises need to watch Jxta and other P2P development efforts to protect themselves and be prepared to support this scenario.

A technology such as Jxta could also become the worst nightmare of the recording industry--or any business that depends on intellectual capital that can be distributed online. The recording industry tamed Napster because Napster was not pure P2P. Like most currently available technologies described as P2P, Napster is actually a hybrid that depends on a central server that can be shut down by the courts.

Jxta promises a P2P technology that would be completely distributed across the Internet, running solely on consumers' desktops, with nothing that central authority can control.

Meta Group analysts Mike Gotta, David Folger, William Zachmann, Dale Kutnick, David Cearley, Val Sribar and Jack Gold contributed to this report. Visit Metagroup.com for more analysis of key IT and e-business issues.

Entire contents, Copyright © 2001 Meta Group, Inc. All rights reserved.