[Image via vt.edu]

The Defense Advanced Research Projects Agency (DARPA) is bringing a new approach to election cybersecurity by developing a new voting system designed to resist attacks through more secure hardware. The  Washington Post’s Cybersecurity 202 has more:

The Pentagon research agency that played a key role in inventing GPS and the Internet has a plan to make voting machines far more secure against hackers.

[DARPA] intends to re-engineer voting machines’ hardware – basically the machines’ physical components, such as computer chips and circuits – so that many of the tricks hackers use to undermine the software become impossible.

That’s a fundamentally different approach from most cybersecurity efforts, which focus on updating and patching software each time a new bug is discovered.

If the project is successful it will make it far more difficult for hackers to penetrate and manipulate voting systems as Russian hackers tried to do in 2016. There’s no evidence Russian hackers actually penetrated machines, but the effort — along with a broader hacking and disinformation campaign targeting Democratic nominee Hillary Clinton — sparked big questions about the influence of foreign actors in U.S. elections and ultimately led to a probe of possible ties to the Trump campaign by special prosecutor Robert Mueller.

“This seems to us to have a deep national interest,” Linton Salmon, the DARPA program manager leading the project, told me. “Most of us want to make sure our votes get counted.”

One big hurdle: There’s no requirement for voting machine vendors that supply states with election equipment to adopt the new hardware. Once DARPA achieves its goal, it plans to simply publish all the technical details so any organization that wants to build the new machines can do so.

There will likely be intense pressure on election vendors, however, to adopt the hardware from the election security community, which has criticized vendors for being opaque when it comes to protecting their machines.

The project is part of a broader DARPA effort to use more secure hardware to protect against hacking in numerous sectors — including at energy plants, financial services firms and the defense industrial base. By making voting machines their first focus area, the agency hopes to excite the interest of the tech industry and the broader public and to help boost security in a highly important area, Salmon told me.

“The intention…is not just to protect DOD systems. It’s to protect everybody,” Salmon said.

The project won’t be complete for another two years, however, and it often takes a year or more to replace a digital system’s hardware. So it’s unlikely to achieve its goal before the 2020 elections.

DARPA has awarded about $15.5 million to eight grant recipients to work on the broader hardware project. It awarded another $4.5 million to the Portland, Ore., security firm Galois to apply those hardware fixes to real-world threat situations – in this case, voting machines.

Galois unveiled the voting machines at a George Washington University election security conference Thursday. DARPA and Galois plan to take their prototype voting machines to the Def Con hacking conference in Las Vegas in August where they’ll invite ethical hackers to try to crack into them. Hackers at that conference found numerous vulnerabilities in existing voting machines during last year’s conference.

After Def Con, DARPA and Galois want to take the machines on a tour of universities where they’ll do similar live-fire testing with graduate computer science students. Then they’ll bring an improved system back to Def Con in 2020.

“We feel the best way is to open it up to the best hacker community in the world,” Salmon told me. “This is getting us a lot better evaluation than we’d get from hiring five gurus to attack it.”

The revamped hardware won’t protect against all potential software hacks, Salmon warned — just against seven main categories DARPA believes can be fully thwarted with a hardware rewrite and that accounted for about 60 percent of all known software vulnerabilities as of 2015.

There are other categories of hacks, though, that don’t have a hardware fix, so the new system wouldn’t be a replacement for other election-security efforts, such as mandating paper ballots or auditing election results.

This last point is key: while DARPA’s R&D effort is intriguing, it will not replace – and indeed is intended to supplement – other changes to election policy or individual behavior (e.g. phishing awareness) intended to make the entire system more secure. As the article notes, there is also the challenge of driving adoption of the changes via the private sector. Still, it’s a fascinating approach to enhancing the long-term security of the American voting process. Stay tuned …