Sun's Project Jxta peer-to-peer application development technology will form part of the firm's One Web services platform, and could also be used to develop applications for server farms and other peer-to-peer software.
Analyst company Ovum said that although peer-to-peer applications are still in their infancy, IT managers might want to experiment with Jxta now to gain some insight into the technology and its potential.
Project Jxta comprises a set of protocols for writing peer-to-peer applications in programming languages for a variety of operating systems, although it has currently only been implemented on Java Developer's Kit (JDK) 1.1.4, with support for TCP/IP and HTTP. However, Sun expects third parties to write reference implementations in other languages, and says that a C version is in the pipeline.
Sun believes peer-to-peer technology will sit alongside existing client-server and Web-based applications. Jxta Source code and software is available free under the Apache License from the Web address below.
Eric Woods, research director of knowledge management at Ovum, said, 'Jxta is not something that is going to worry many IT managers at the moment because it's mainly aimed at developers. IT managers will be wary of the technology because of fears about security and control.'
However, Woods said IT managers would be wrong to ignore Jxta and peer-to-peer systems. 'It would be sensible to look at it in a controlled way. There's always the danger that if people just reject it, it will creep in through the backdoor anyway, so it makes sense to look at it in experiments,' he said.
Woods said Jxta could be attractive for IT managers wanting to build collaborative or file-sharing applications so their companies can work more effectively with customers, suppliers and home workers without having to resort to expensive integration work or the use of electronic data interchange systems.
On the other hand, Phil Dawson, programme director at analyst firm the Meta Group, said Jxta is more likely to be taken up by technical departments using Linux clusters, for example, in research and development work. He believes Jxta would be useful for developing applications that can make use of idle processing power on an organisation's PCs.
'IT managers won't do much with Jxta for the next two years or so, but a lot of our technically competent clients are looking into peer-to-peer, and this is where the main Jxta work will appear,' Dawson said.
However, there is no guarantee that Jxta will be a success, said Woods. 'It's hard to say whether or not Jxta will play a part in the peer-to-peer movement until developers take it up. It's so early in the movement that there will, no doubt, need to be several major revisions of the technology as peer-to-peer evolves,' he said.
Andy Bush, new technologies manager at Sun, was bullish about Jxta's potential, and described it as enabling technology for Sun's One Web services initiative.
'Jxta takes things down to the simple core level like TCP/IP did in the networking world a few years ago. It provides low-level services and allows everything to communicate with everything else  what platform or device it runs on won't matter any more as long as the device can load an application,' he said. Bush believes Jxta has potential in many situations where people need to share information.
'There is a lot of information currently locked inside devices, because it's very difficult to publish it to a peer. But you can use Jxta to build services for staff to unlock that information. This is currently what people are trying to do with knowledge management systems, but it's relatively difficult to add applications to systems that are silos of information because they often do not interoperate very well,' said Bush.
