|
by Jon Byous
June 7, 2001 -- As thousands of developers took their seats at this morning's keynote address, they watched a live onstage scene from Survivor -- the JavaOne conference version. Considering the popularity of wireless devices among the attendees, there's a good chance that the players onstage were the only people in the room without immediate access to a network. There's No Services Like the New Services Today's keynote address focused on the future of networking and, most especially, the big questions we face in making that future better. Evolve Through Decomposition Using a set of slides outlining the history and possible future of network technologies, Greg Papadopoulos, Sun's senior vice president and chief technology officer, showed how networks have become more distributed and personalized to individual needs.
"Over the years, network computing has evolved through the fundamental principles of decomposition and specialization," said Papadopoulos. "At each stage of network development, we have to work out a new model and set of patterns for how to build the next system." He continued, "It's a constant evolution, but decomposition and specialization remain. And as these models and patterns evolve, they have to coexist with the accumulated previous versions." Then he warned, "Don't be forced into thinking there is going to be one solution that is going to solve all of the world's problems." Indeed, old models are slow to disappear. "In fact, one of the issues that we're all having to deal with is that the network becomes more distributed and larger in scale; there are many more different and interesting places on it we want to visit; and we become more and more aware of the vagaries of the network. Errors, disconnected operations, and bandwidth latency become more obvious." In the new model of networking, Papadopoulos wants us to look deeper into the network and identify the weaknesses and problems. Users should be able to see a service's latency and limitations. "As we establish the model and patterns for a larger and much more distributed network, we have to ask, What is the minimum set of assumptions and requirements that would enable distributed devices to join together in a community?" Joy Has Entered the Building Bill Joy, Sun's chief scientist, then joined Papadopoulos onstage to talk about new ways of looking at the next stage of network evolution.
Papadopoulos asked Joy about his current philosophy of network computing. Joy answered, "Network computing needs to be more social, democratic, probabilistic, not deterministic. It's not about the technology dictating how we interact but about the way you and I want to interact with the technology. This is the most important place to be right now." He continued, "The growth of the Web has barely begun, and the challenges are significant. No one technology out there today can actually address or access all of the information, content, and data on the expanded Web, until now." With that, Joy described Sun's new peer-to-peer technology, Project Juxtapose (JXTA, pronounced "juxta"). An industry-wide research project led by Sun, JXTA is designed to enable developers to create flexible, interoperable applications that are available on a wide variety of devices, overcoming the limitations of peer-to-peer (P2P) distributed computing. The goal of Project JXTA is to leverage distributed computing by providing direct access or communication to resources from one node (device) on the network to another, without any centralized server control. It includes a core (peer groups, peer pipes, peer monitoring, and security), services, and applications. JXTA will give users much more focused web access: for example, the ability to discover and join peer groups, chat securely within a peer group, share files with a peer group, and view a list of shared contents available on a remote peer.
It is also designed to be agnostic to programming platforms (such as the C and Java programming languages), systems platforms (such as the Microsoft Windows and UNIX® operating systems), and networking platforms (such as TCP/IP and Bluetooth). Joy commented, "Since the release of the JXTA code to the Java community just six weeks ago, we have had over 50,000 downloads of the source and binaries, and numerous community members have already begun projects." Sun expects many different implementations of JXTA technology to emerge in the months ahead and will let the market determine which ones will prevail. Currently, the first JXTA prototype implementation is available on http://www.jxta.org. Panel Discussion: Evolve by Asking the Right Questions In closing, Papadopoulos and Joy were joined onstage by a panel of technologists, including Adam Bosworth, founder, chairman, and CTO of Crossgain; Tim O'Reilly, president of O'Reilly & Associates; James Gosling, Sun vice president and fellow; Rob Gingell, Sun vice president and fellow; and Mike Clary, Sun vice president, Project JXTA.
Discussing the future of network technologies, the panel agreed on several points. The first is this: We know we are going to move toward more and more complex, distributed, and diverse networks. The challenge is to find the minimum set of assumptions we can make about the needs of such a network. The key issue is, "What kind of infrastructure can be put in place to be used by everyone but owned by no one?" The crowd roared with spontaneous applause. Another point of general agreement was one made earlier by Papadopoulos, now echoed by Gingell: "I think we need to look at how we think about problems in the first place. Network technologies of the past were designed to hide the problems of the network and shield the user from its performance issues. We've got to stop doing that." Everyone agreed: That's got to change. And, as Gingell said, "It's time to pull the rug out from underneath the idea that you have a perfect machine and a perfect network. It's just not true, and we have to deal with it." That comment sparked the mention of Peter Deutsch's article, "The Seven Fallacies of Distributed Computing," and, like a Greek chorus, the panel members automatically began reciting its list of incorrect assumptions. "The list, with the addition of an eighth fallacy, can be found at http://java.sun.com/people/jag/Fallacies.html," said Gosling. The Community Is the Network And again, for the third day, the river of devoted Java technology developers flowed out of the massive hall to begin building an incredible future for all of us, line by line, node by node.
See Also
Day 1 Technical Keynote: Discovering The Possibilities
Day 4 Conference Keynote: Stay Open, Stay Busy, and Stay Cool | |||||||||
|
| ||||||||||||