By Don Dugdale
A future of ultra-small, ultra-fast, ultra-concentrated electronics, offering more services available more quickly to more people and on a wider scale than ever before was described by a panel of UniForum '95 futurists.
Assigned to discuss the most important developments taking place in computing were Samuel Fuller, vice president of corporate research for Digital Equipment Corp.; Tsvi Gal, vice president of information technology at Bank of America; and Dennis Georg, director of Hewlett-Packard's computer systems laboratory. Panel chair Allen Paller, vice president and director of open systems for Computer Associates, described the three as having "a vantage point on technology that is a little higher and farther forward than most of us."
One of the issues facing users in the coming year, Gal said, is expanding use of the Internet, which will no longer be just a thing for "techies" to play with. "We really liked it when it was our toy," he said. "It was a little nerdish but it was fun." Then he offered a warning. "By and large the Internet is not secure. If a company is getting into the Internet, they should be very careful, because once a person is on your network, he's in and can jeopardize your entire network," Gal said.
The power of the Internet will show up in the home to a greater degree, Gal predicted, especially in home banking. "It will give the power of doing everything that you can do in a branch from your home, except for getting money," he said. "I see a major breakthrough in working with small companies and providing them a full range of services over the Internet. It can eliminate the need for a small business to go to multiple sources to do their business." He said the use of the Internet by businesses-only about 5 percent of total Internet traffic today-will increase dramatically.
Bandwidth availability, which is dependent on public utility regulation, may spell the success or failure of many Internet applications, Fuller indicated. "We now have the issue of what kind of political reform will go forward that will deregulate the industry and will or will not allow this to really blossom," he said. "Largely, the politicians in Europe and North America have in their hands the ability to either accelerate this transformation by freeing up the communication lines or slow it down by keeping a lot of the rate structures in place. If you continue to have to pay what you pay for a T-1 line today, it will be very hard to have effective multimedia communication nationwide. People can't afford to pay over $100,000 a year just to have that multimedia connection."
The so-called information superhighway will evolve from today's Internet, Georg said. He added, "I think it's going to involve perhaps a lot of standards that grow from the Internet."
Fuller also noted that capacity in almost every technological area is doubling every 18 to 24 months. "About the only critical technology that's not doubling every couple of years is battery capacity," he said. "How long will this continue, and will we continue to satisfy the need? The smallest critical dimension is the gate oxide width in CMOS transistors, which was decreasing by 13 percent per year from the 1960s through the early 1990s. For that to continue, in 10 years, namely in 2005, the width of that oxide layer would be thinner than the diameter of a silicon atom. So something is going to happen before 2005.
"The other piece of reality in our industry is that we go through a new generation roughly every 18 months, and we go through a major generation every three years. Quite literally, the industry gets turned on its head every 10 years. As we look at it today, we've got multiple generations and an entire industry revolution to go through before this rate of increase is going to slow down."
Georg said more than two generations at the current improvement pace are likely. "What we call instruction-level parallelism, or the amount of work that can happen in a given machine's cycle, is going to do in general purpose computing what it's done in loop computing in the past. I see a great deal of optimization of codes, faster transistors, and more dense memories. I believe this trend is going to continue."
"A growing share of our business and personal lives is going to depend on computing," Georg said. "I think of systems in the year 2000 as being faster, smarter, smaller, and more specialized. The most significant change is how systems might be smarter. Smarter computing is going to support measurable improvements in management, adaptability, measurement, scalability, and availability. These are some dimensions that we're not meeting expectations on today. I think the industry is going to pay more and more attention to the integration of legacy and next-generation applications and data."
Connecting different levels of computers is not the problem, Georg said. "The real problem is in integration." Scalability is going to have to be stretched several orders of magnitude with new design paradigms, he said.
More computing power will also be needed, Georg added. "If we look at specialization in visual computing, if we move from 2-D to 3-D, from interactive to real-time, from modest resolution to high resolution, we need four orders of magnitude more processing and memory power to do that. It's hard to believe that, in the future information-rich world, we're going to be satisfied with the [visual technology] that we have today."
He also predicted more mobile and wireless computing, and faster storage retrieval with greater storage density. "Faster is also a trend," he said. "Faster is one of those trends we like to talk about that gets a lot of hype, but I think it's also very predictable."
If these experts are right, the near future holds exciting possibilities. It will be equally interesting to see how well the IT industry meets these challenges.