Advertisement

DARPA wants help answering ‘trillion-dollar questions’ in the smallest way possible

Bill Chappell and the Defense Advanced Research Projects Agency’s Microsystems Technology Office has some great ideas. But it’s looking for help.

DARPA’s MTO held an expo Friday presenting the new areas in which it’s concentrating its efforts to facilitate collaboration between academics, industry and government that will continue to push the boundaries of microsystem technology.

Chappell has organized the office’s focus into four main areas, or what DARPA calls “thrusts:” electromagnetic spectrum, decentralization, information microsystems and globalization.

“Today is about saying DARPA is the place where you do the big inventions, the capstones of your career that will hopefully change the world in a very literal sense,” Chappell said Friday, adding that the thrusts are a “re-evaluation of why we are doing some of the microsystems we are doing.”

Advertisement

Some concerns in each thrust mirror issues the federal government and global tech community have been wrangling with in the past few years.

The information microsystems thrust is focused on processing the massive amounts of data the world continuously creates, even as engineers are beginning to reach the end of Moore’s Law, or the ability to double processing speed every two years.

Joseph Cross, a MTO program manager, used an interesting analogy to sum up the current state of processing systems: Imagine you are rich and you employed a very smart butler. A year later, your accountant fires the butler and replaces him with two butlers who cost less, but aren’t as smart. The next few years, your accountant keeps adding cheaper-but-dumber butlers to save money.

“Until one day, you open your balcony doors and out on your veranda are 1,000 drooling butlers,” Cross said. “The fact that they are dumb is fine if you want them to paint a fence. It isn’t fine if you want them to make a hollandaise sauce for your breakfast.”

The “hollandaise” Cross is referring to is the vast amount of data the military produces that needs to be utilized. To give perspective to the amount of stress the modern military puts on data processors: One Global Hawk drone uses 500 percent of the bandwidth the entire U.S. military used during the Gulf War.

Advertisement

“The number of transistors has been going up, that’s good, but the ways to use them efficiently have either hit a wall or [are] projected to hit a wall from an engineering perspective,” Chappell said.

Another growing concern for the military is component counterfeiting, which DARPA is looking to tackle through the globalization thrust.

Daniel Green, another project manager, presented slides of transistors and semiconductors that have been forged or tampered with in order to look like components normally sold by Intel.

“In the past, we’ve sort of buried our head in the sand,” Chappell said. “One of DARPA’s roles is to be the leading indicator of what’s going to happen and to avoid strategic surprise.”

While studying how counterfeit components end up in global supply chains is more a business focus, Chappell said it will ultimately be beneficial for the military.

Advertisement

“We have to embrace to global supply chain, but we have to do it in a way so you trust the components that show up on shore,” Chappell said. “If you only upgrade your technology every 20 years, with the fact that we are getting new electronics on a yearly basis, that doesn’t really work for us. We have to show [the military] how to get there through advanced research.”

While the focus of the MTO has been to make things as small as possible, they also want those looking to get involved to figure out how to make things cheaper.

“Cost is now killing us,” DARPA Director Arati Prabhakar said. “It’s a self inflicted wound, but it’s a very serious issue.”

But Chuck Wolf, deputy director of DARPA’s Adaptive Execution Office, said anyone who steps up will have eager “beta testers” ready and willing in the military.

“It doesn’t need to be earth-shattering,” Wolf said. “These improvements can be incremental.”

Advertisement

“We’ve always done advanced computing,” Chappell said. “When that stops, who is going to figure out these trillion-dollar questions?”

Latest Podcasts