DARPA calls for 'revolutionary' software development
- By Kevin McCaney
- Aug 11, 2015
Military researchers are cooking up a number of ideas to improve security by fundamentally changing how software is created or networks operate, with plans for such things as transparent computing and self-adjusting software that could last 100 years. But they don’t want to stop with just their own ideas; they’re also looking for ideas they haven’t thought of.
The Defense Advanced Research Projects Agency’s Information Innovation Office (I2O) has opened a new suggestion box with a Broad Agency Announcement seeking revolutionary approaches with the potential to guarantee trustworthy computing and greatly improve a program’s interaction with human users.
“I2O seeks unconventional approaches that are outside the mainstream, challenge accepted assumptions, and have the potential to radically change established practice,” the announcement states.
What researchers are not looking for is anything being addressed by any of I2O’s other projects or, for that matter, any product a company or other research team has completed or received funding for, or anything that could be considered “evolutionary.” “Proposed research should investigate innovative approaches that enable revolutionary advances in science, devices, or systems,” according to the BAA.
I2O has two primary goes for the project.
1. Empower the human within the information ecosystem, including systems that can understand human speech and other forms of communication; derive information from diverse media; learn, reason and apply knowledge gained through experience; and respond intelligently to new and unforeseen events.
2. Guarantee trustworthy computing and information. DARPA is looking to change the paradigm in which computer security is based on finding vulnerabilities and then issuing patches—noting that sometimes the patches themselves can introduce new vulnerabilities. Researchers are looking to create “software that is inherently resilient to attack and computing architectures that can be rapidly restored following an attack,” the BAA states, before listing the areas of research that could support that kind of software.
DARPA said it expects to fund a limited number of proposals, with funding amounts depending on the quality of the proposals. Given the shoot-for-the-stars nature of the project, the agency said it is interested in hearing from non-traditional performers—that is, not just from contractors or universities that regularly work with the military—and it is giving potential proposers plenty of time to think it over; abstracts in response to the announcement are due June 10, 2016, with proposals due Aug. 2, 2016.
Kevin McCaney is a former editor of Defense Systems and GCN.