Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
This short conference talk presents a novel technique for detecting side-channel information leaks before they occur through staged, instrumented interpreters. Learn how researchers from Harvard SEAS and MIT developed a method where timing behavior becomes a first-class value that can be analyzed using classical methods. The presentation demonstrates the effectiveness of this approach against naive, cache-based, and speculation-based timing attacks using an off-the-shelf bounded model checker. The research team, including Cameron Wong, Muhammad Abdullah, Yuheng Yang, Mengjia Yan, Adam Chlipala, and Nada Amin, delivered this talk at the PEPM 2025 conference on January 21, 2025, sponsored by ACM SIGPLAN, ACM SIGACT, and ACM SIGLOG.
