FEATURES & COMMENTARY
San Diego, CALIF. — Barnaby J. Feder reports that for more than 40 years, the Defense Advanced Research Projects Agency has had the daunting assignment of making sure that no enemy or, for that matter, any current friend has access to tomorrow’s technology faster than does the Defense Department.
Darpa, as the agency is known, has parceled out billions of dollars for research projects that can often seem risky, even far- fetched. But the agency assumes that the really disruptive innovations are rarely obvious in their early stages.
“Anytime you propose something that’s clearly achievable, they say it’s not a Darpa problem,” said Jon Sherbeck, vice president for engineering at Mdot Aerospace, a Phoenix-based start-up company that won a small-business award from the agency recently for designing a gas turbine the size of a D-cell battery.
Much of what turns out to be feasible never makes it beyond demonstration projects. Military and Congressional politics create some roadblocks and changing military priorities throw up others.
And yet, no other organization has so frequently if unintentionally served as midwife to the technologies of today’s information- driven economy. Darpa gave birth to the Internet in the late 1960’s. Other projects helped create many of the nation’s most impressive computers, the chips used in cellular phones and vital networking technologies like the ability to send simultaneous signals of many wavelengths down a single fiber optic cable.
What’s next? Well, as the agency told more than 1,000 engineers and high-technology executives who recently attended a three-day symposium in Dallas that summarized its current agenda, the next decade or two is likely to see extremes in miniaturization like molecule-size computers and six-inch spy planes. Machines, men and, perhaps, other living organisms like insects fitted with sensors will be joined in information networks of bewildering complexity for all sorts of military uses. The torrent of information could make today’s network flow seem like a slow trickle.
Breakthroughs in biology and new materials could also create surprising new possibilities. The new materials include amorphous metals that resist corrosion and, unlike normal crystalline metals, become harder under stress.
For a relatively simple example of the agency’s perspective, consider one of today’s military mainstays, the Abrams tank. These battlefield wonders are more maneuverable and lethal than any other tank in operation. Their ability to accurately aim at enemy objects and fire on the move is unprecedented. With twice the range of Iraq’s Russian-designed tanks, they played a major role in NATO’s total domination of Iraq’s larger army in the Persian Gulf war.
But to the advanced projects agency and some Army officials thinking about combat in the future, such large tanks are probably at an evolutionary dead end. Seen through the lens of the information age, the functions of a tank would be scattered across a combat network. A variety of robotic cannons and missile launchers could deliver the firepower, supplementing their built-in intelligence with help from a distant, manned command center. Smaller, quicker armored vehicles with single, large guns and just one or two crew members might also have a role.
A variety of heat, light and motion sensors on the ground and in the air could be networked to provide a much better picture of fighting conditions. Data would pour in not just from the ground vehicles but from satellites, low-flying miniature aircraft and, perhaps, thousands of pollen-size sensors floating like dust or attached to insects.
With military budgets shrinking and private-sector information technology development growing at an explosive pace, the advanced projects agency is increasingly looking for ways to repackage civilian technology for the military as a cost-effective alternative to sponsoring novel projects. Unfortunately, the networks and information systems the private sector is promoting look disturbingly brittle, threadbare and unsuitable for military purposes.
“B2B and enterprise software are low- hanging fruit technically compared to defense needs,” said Shankar Sastry, director of the agency’s Information Technology Office, referring to products that allow businesses to interact with each other over the Internet and software that allows many parts of an enterprise from finance to production to marketing to share data.
Darpa studies are looking at how to create and manage networks of up to 100,000 components that must be quickly able to shift how they communicate. Some components may periodically change functions – a device, for instance, might be monitoring radio signals at one moment and jamming them the next.
Similarly, the agency is financing research on networks of information users. It wants the Defense Department to be able to set up coalitions of up to 100 allies capable of sharing secure information and coordinating logistics within minutes. Just as important, it wants to be able to cut a party off a network instantly and notify all other members.
In both cases, the need for artificial intelligence is obvious. Even if the ability of humans to interact easily with computers were vastly improved by, say, flawless voice recognition software, the amount of information flooding in would be overwhelming. The agency’s response includes research on software that can screen widely diverse data sources for relevant trends, automatically reshape networks and write new programs to respond rapidly to emerging needs.
One project partly addressing such concerns is the effort to develop so-called Darpa agent markup language. D.A.M.L., as the language is known in technical circles, would would create a universal format for telling computers what kind of information is in a data source. It would allow Internet search machines to extract data not just from the World Wide Web but from computer programs, sensors and other machines.
The computer language embodies one of the major changes the agency has confronted in adapting to the information age. While much of the agency’s work still focuses on technology that it would just as soon see stay in American military hands, there are a growing number of developments that will go nowhere unless they also permeate civilian life.