Yes, I admit it, I still dream of Jeanie. I grew up in the ’70s, and what boy could resist Barbara Eden? But now that I’m a little older, I no longer believe in genies or magic. In fact, I’ve become quite skeptical. Now, Sun is asking me to believe in Jini. Jini is Arabic for magic, but it is also the name of Sun’s strategy for enabling spontaneous networks. Sun states that Jini will “connect anything, anytime, anywhere.” Yeah, right. Time to change the channel from I Dream of Jeanie to Cops. I mean, give me something real here. Yes, Jini is a very attractive technology, but can Sun really perform the magic required to fulfill this promise?

Hocus Pocus

Scott McNealy, president and CEO of Sun Microsystems, says Jini is real “plug-and-play not plug-and-pray.” Remember the great fanfare about the plug-and-play feature of Windows 95? Microsoft’s plan was to make the installation of PC peripherals easy. The problem with the Windows operating system is that for a device to work on your 95 machine it must have a device driver compiled specifically for Microsoft Windows installed.

Jini’s plug-and-play is implemented entirely differently from Windows’ plug-and- play. If you plug a Jini-enabled device into the network, a discovery packet drops onto the network. Meanwhile, a Jini lookup service is listening for these discovery packets. After the device gets a response from the network, it uploads its services as a proxy to the lookup service. Any other network device is then able to invoke the services of the newly attached device by downloading the pertinent Java code from the lookup service. At that point, the client can invoke any of the services that the remote device supports. OK, perhaps that’s too much detail. Let’s just say the client folds its arms and nods its head while blinking, and the service of the new device is magically invoked. Sun likens Jini’s discovery process to the phone tone in a standard telephone network (to the point that they are considering using the JavaTone brand name to describe the Jini environment). But as William Blundon of JavaWorld said in Blundon’s Corner (www.javaworld.com/javaworld/jw-10-1998/jw- 10-blundon.html): “One does not connect to a dial tone, one connects to a telephone network. The dial tone indicates you have done so successfully. To get a JavaTone, you will need a JiniNetwork.”

Who Needs It?

Whether or not you believe in Jini’s magic, you may be wondering if you really need its features. Sun’s Richard P. Gabriel, editor of www.jini.org, is sure that you will. In a Web article called “The New Community for Jini Connection Technology,” Gabriel says:
“People are carrying cellphones, pagers, PDAs, laptops, and digital cameras; each may work in a variety of places—at home, in several offices, in conferences rooms, at client sites....The places they work are less like [traditional] offices and the tools they use are less
like centralized computers than ever. What such people need is a way to make sense of it all, a way to have all these disparate tools made by different companies for different purposes work together.” Gabriel further preaches the promise of Jini technology: “The Jini architecture is built on the idea that small devices and small services can form into larger collections that, perhaps, one day will work together as a congregation.” (Find the full text at www.jini.org/opinion.html.)

Competition

Jini isn’t the only source of networking magic. There are several competitive technologies in the industry: Caltech’s Infospheres, AT&T’s GeoPlex, IBM’s T Spaces, and Microsoft’s Universal Plug and Play. The hype on these technologies sounds strangely similar to Sun’s hype on Jini. A Lucent white paper, for example, explains: “The term infosphere originated in the military and refers to an individual’s collection of interfaces to (possibly remote) software tools, data, appliances, servers, and collaborators.” (See www.infospheres.caltech.edu/papers/ domc96.html.) AT&T’s hype on GeoPlex also sounds familiar: “GeoPlex, a platform developed by AT&T Labs, makes the deployment of services across networks easier and more cost effective. The features in this platform include: security, usage recording, access control, protocol mediation, registration, and authentication.” (See www.ipservices.att. com/internet2/i2_wyg.html.) And IBM’s white paper on T Spaces (www.almaden.ibm.com/ cs/TSpaces/) makes some of the same claims: “T Spaces is a network communication buffer with database capabilities. It enables communication between applications and devices in a network of heterogeneous computers and operating systems.” Both GeoPlex and T Spaces, by the way, are implemented with cross-platform Java. Microsoft, of course, has its own alternative to Jini: Universal Plug and Play. James Niccolai, of IDG News Service (www. javaworld.com/javaworld/jw-01- 1999/jw-01-idgns-msplugplay_p. html) says that Universal Plug and Play is “an initiative designed to allow a broad range of devices such as PCs, printers, and even security cameras to connect as peers over a home network and share resources.”

Three Wishes

After researching Jini, I now know that it is not magic but a well-thought-out technology with great potential. However, Jini has a lot of competition, and if our devices are to work as a “congregation,” they must adhere to the same protocol. That’s a problem; we must either quickly select from the variety of alternatives or wait for one of them to become standard.

When the genie of the lamp asks me my three wishes for the spontaneous networking technology that becomes standard, I’ll say the tightest code, the best security, and the greatest reliability.