The software, called Jxta and pronounced "juxta," is Sun's contribution to the much-hyped "peer-to-peer" technology made famous by file-swapping programs such as Napster.
Bill Joy, Sun's chief scientist, announced the Jxta software project Thursday at the O'Reilly Peer to Peer conference here, exhorting outside programmers to help standardize Jxta as the basic plumbing for building new peer-to-peer applications.
As previously reported by CNET News.com, the project is a direct threat to .Net, Microsoft's effort to make its Windows operating system the future foundation of the Internet. Jxta is the latest in a long line of Sun projects designed to reduce operating systems such as Windows to mere cogs while people write software that works at a higher level.
Sun hopes to enlist one of Microsoft's worst foes, the open-source community, in the Jxta effort. Jxta will be open-source software, meaning that anyone can modify and redistribute the software without restriction. The open-source community has challenged not just Microsoft but also Sun with successes such as the Linux operating system and the Apache Web server.
What's in it for Sun is a peer-to-peer infrastructure on which higher-level commercial applications can operate, Joy said.
"We have some distributed applications we'd like to run," he said.
Jxta will include standards for how devices in a peer-to-peer network identify themselves and are grouped together, Joy said. It also will include a security mechanism to ensure that distributed programs don't harm the device they're running on, Joy said, in contrast to the e-mail viruses that afflict networked Windows systems.
Mike Clary, head of the Jxta project, said the software will include ways that computing tasks can be linked together in "pipelines" that span a peer-to-peer network. In addition, Jxta will offer a mechanism by which tasks can be monitored and controlled.
Taking on Microsoft
Jxta fits into Sun's vision of the future of the Internet, called Sun One. "This will probably be part of the Sun One platform, but will probably be one of the simplest pieces," Joy said. "We're not turning this into something that's infinitely complicated like (Microsoft's) .Net."
The new project walks a fine line between attempting to control the burgeoning new peer-to-peer technology world and leaving it open to the whims of smaller developers.
Joy disclaimed any desire to control the new market. Jxta is meant simply as a lingua franca that will let separate peer-to-peer applications work in concert with one other, he said.
"We don't want to have a standards body," Joy said. "We are not here trying to get everyone to license this like we're doing with Java. We don't necessarily even want to be the center of this."
Nevertheless, the first release of Jxta--due out in April on the CollabNet site, Joy said--will be code-written under the direction of Sun and Joy's team. And it's not easy to foresee any other company or set of developers taking the same kind of role, even in an ostensibly open-source project, that a company with the market power of Sun has.
Jxta already has some competition as a unified effort. Intel has sponsored a peer-to-peer working group that is already aiming at creating standards for peer-to-peer applications. Although the effort initially proved controversial, the group has already begun work.
Learning from past mistakes
Sun no doubt hopes Jxta will be more successful than its predecessors--two other Sun brainchildren, Java and Jini--in luring the help of the open-source movement.
Well after Java and Jini were created, Sun tried to attract the attention of the open-source community by retrofitting the tools with half-shared, half-proprietary software licenses.
But it's not easy to harness the energy of open-source programmers. "We're not unaware that trying to build these communities and getting people to work together is harder than writing code or even starting a business," Joy said.
Java, unveiled in 1995, was Sun's first attempt to bypass Microsoft. Java's promise, only partly fulfilled six years later, is to let programs run on any type of computer--Windows, Linux, or anything else with the proper Java foundation.
Jini, announced in 1999, was designed as a way to get gadgets such as digital cameras and printers to communicate without requiring computers to act as an intermediary. Despite Sun's promises, though, Jini has largely been a commercial flop.
Joy said Jxta will run well on Java-enabled devices, but won't require Java as a foundation. Here again, Sun appears to be learning from its mistakes: The company is working on a revised version of Jini that also doesn't require Java.
Humpty Dumpty in reverse
Sun's project may be the most ambitious effort yet to bring together a young peer-to-peer world that is quickly fragmenting into dozens of different networks.
Although many of these discrete services operate perfectly well on their own--Napster being the most successful, of course--peer-to-peer developers and investors are increasingly calling for some kind of bridge.
"The risk we face is a maze of balkanized networks," Clay Shirky, a partner at venture capital firm Accelerated Ventures, said in a keynote speech at the O'Reilly conference Wednesday. "We need to talk about interoperability."
This goal, in many of the most ambitious peer-to-peer developers' minds, would be to allow the Net to move a stage beyond today's Internet, which consists largely of tapping into simple text, video or audio services on a Web site or basic services such as e-commerce.
These developers are dreaming of a much more complex network, in which individual computers, wireless phones, powerful servers and databases all work together to offer new kinds of Web services, whether they are software services offered remotely or interactive sharing programs like Napster and Gnutella.