Eight billion CPUs will ship this year, but a scant 2% will be networked. With this as a starting point, speakers at the three-day Intel Computer Continuum Conference (San Francisco, Calif., March 15-17) urged the audience to explore options for the next 20 years of computing. Furthermore, speakers urged participants to establish a road map for achieving an unbroken computer continuum.
David Tennenhouse, vice president and director of research for Intel Corp., was a keynote speaker. He began by explaining how Intel views its own place in the computer continuum. "Intel is a building-block provider to the Internet economy," he stated. The company's recent introduction of software for the Digital Visual Interface (DVI) will bring content protection to digital video on the Internet. It's an example of how Intel, unwilling to settle solely for core competency in silicon, has moved into other arenas. Tennenhouse, though, quickly acknowledged that the vast majority of good ideas will come from outside of the company.
Tennenhouse believes the industry's challenge, with regard to the computer continuum, is the reinvention of architecture, networking, and software. He also has his eye on those 8 billion microcomputers entering the world this year. More than 95% of them will be embedded into equipment like buildings, houses, vehicles, and clothing.
Then there's the fact that only 2% of CPUs are networked. Tennenhouse asks that we "network the rest of them." From a human perspective he asks, "What will become of the world when the balance becomes networked? Can we envision the coordination and control of so many highly dynamic unattended systems, comprised of tens of thousands of physically unattended devices, enormous numbers of networked devices monitoring and interacting with the physical world in hospitals, factories, offices, homes, vehicles, and the human body? Well, we can't."
Moving from a current computer-centric state with perhaps no more than 2% of all computers networked, to a human-centric state where humans become closely connected to the physical world through computers, is to move us toward a world in which computers truly become ubiquitous. "The Internet continues to penetrate cracks and crannies," says Tennenhouse. But, he's weary and the Internet is tiring him out. "I want my computer to do millions of things for me," he states.
Problems like the ones discussed here result from successes that pile up over time. So also do the major obstacles confronting us on the Internet such as privacy and security. Tennenhouse wrapped up his talk with three concise recommendations; "Let's get 'physical,' let's get 'real,' and let's get 'out'."
By getting physical, he looks to bridge the physical and virtual worlds. Becoming real means driving applications toward real-time operation. He says, "Let's squeeze latency out of every system—from the devices to the applications—and let's enable fine-grained interaction across subsystem boundaries. As for getting out, let's relinquish control over how functionality is implemented. I want my computer to do millions of things for me."
With everything said, there's one fact that must remain inviolate. "People," says Tennenhouse, "are precious." He looks ahead to the time when ubiquitous computing will be harmoniously blended into all aspects of our lives. And above all else, it will be aimed squarely at serving mankind's most human needs.
What do you think the options are for the future 20 years of computing? Send me your thoughts.