This article is more than 1 year old

Researchers propose simple fix to thwart e-voting attack

Running hash makes insider fraud easier to detect

Researchers have devised a simple procedure that can be added to many electronic voting machine routines to reduce the success of insider attacks that attempt to alter results.

The approach, laid out in a short research paper (PDF), augments the effectiveness of end-to-end verifiable election systems, such as the Scantegrity and the MarkPledge. They're designed to generate results that can be checked by anyone, by giving each voter a receipt that contains a cryptographic hash of the ballot contents.

The researchers propose chaining the hash of each receipt to the contents of the previous receipt. By linking each hash to the ballot cast previously, the receipt serves not only as a verification that its votes haven't been altered, but also as confirmation that none of the votes previously cast on the same machine have been tampered with.

The procedure is intended to reduce the success of what's known as a trash attack, in which election personnel or other insiders comb through the contents of garbage cans near polling places for discarded receipts. The presence of the discarded receipts is often correlated with votes that can be altered with little chance of detection.

The running hash is designed to make it harder for insiders to change more than a handful of votes without the fraud being easy to detect.

“This mitigation makes the attack far more difficult and makes it nearly impossible to alter more than a small number of votes,” Josh Benaloh of Microsoft Research and Eric Lazarus of DecisionSmith wrote. “This mitigation also offers additional benefits to many verifiable systems at minimal cost.”

Most verifiable election systems already include a cryptographic hash on receipts returned to voters, so the inclusion of a running hash should be relatively easy to incorporate, the researchers said. ®

More about

More about

More about

TIP US OFF

Send us news


Other stories you might like