Is Jini out of the bottle?

Depending on whom you ask, the Jini networking technology, introduced three years ago January, has either stalled or is just now gaining momentum. Sun Microsystems Inc. believes it is the latter.

“Today we are still in the early adopter phase,” said Franc Romano, Group Marketing Manager of the Jini technology network group at Sun. “We’re starting to see tangible results now,” Romano said, referring to the 75 commercial implementations that are shipping in products now.

This month, Palo Alto, Calif.-based Sun moved Jini forward with the release of an initial set of guidelines for supporting and accommodating existing security standards within the Jini architecture. But Jini’s long-term viability is still in question, according to analysts.

Jini was dreamed up by Distinguished Sun Scientist Bill Joy as a network architecture that allows objects developed in Java to be moved from one device to another. For example, with a Web-based travel portal, Jini would be used to move Java-created interface objects that allow the sharing of data across various disparate reservation systems built on different platforms. Jini passes the objects to each disparate system and thus enables the different applications to ostensibly communicate with one another by sharing data. Without Jini, a lot of programming would be required to identify and tag all of the data needed to make the systems interoperate.

But Jini’s role as an intelligent transport layer to deliver services over disparate and inherently fallible networks has come into question. This has occurred especially as Sun’s own Jxta initiative – another Bill Joy project – and other architectures based on delivering XML code continue to rise in popularity. But Sun officials are quick to reassure Jini developers that the technology is not going away anytime soon and Jxta is not a threat.

“Comparing [Jini to Jxta] isn’t even like comparing apples to oranges,” said Jim Waldo, distinguished engineer in the Jini network group. “It is more like comparing apples to Oldsmobiles.” Waldo pointed out that Jxta is a simple way for devices to discover one another across a network. Jini discovers other devices, services, or applications by what they do via its lookup functionality.

But some analysts aren’t convinced that demand for Jini will be widespread. Web services, to the extent that they exist today, are being used mainly for linking applications inside an organization or among established business partners, which makes Jini’s ability to “look up” and connect with distant Jini-enabled computers of questionable value, said Mike Gilpin, a research fellow at Cambridge, Mass.-based consulting company Giga Information Group.

Despite the questions of its validity outside of Java-enabled network – although Sun officials say most would be hard-pressed to find a network that wasn’t running Java somewhere – Jini is moving forward.

Earlier this month, Sun presented an initial set of guidelines for supporting and accommodating security standards with Jini. Part of an initiative dubbed the Davis Project, the security guidelines will allow developers to support security functions such as mutual authorization between client and server, to access control between two services across a network, and to support a variety of security implementations with differing protocols, algorithms, mechanisms, and policies.

“Our idea is pretty straightforward,” said Bill Scheifler, distinguished engineer and chief architect for the Davis Project. “We’re not going to invent new algorithms. We want developers to choose any mechanism they want.”

The initial network security programming model has been presented in Sun’s Overture release. The hope is that some of the 80,000 registered Jini users will bang away at the release to improve its functionality. A year from now, Sun believes, the architecture will be complete. Sun officials will then demonstrate that the Jini technology is viable, thus attracting more attention from the developer community. Until then, nobody can say Jini is dead.