Title:

Static Analysis of Cache Side Channels

Abstract:

Side-channel attacks recover secret inputs to programs from physical
characteristics of computations, such as execution time or power
consumption. CPU caches are a particularly rich source of side
channels because their behavior heavily impacts execution time and can
be monitored in various ways.

CacheAudit is a platform that enables the automatic, static analysis
of such cache side channels; it takes as input a program binary and a
cache configuration, and it derives formal, quantitative security
guarantees for a comprehensive set of side-channel adversaries, namely
those based on observing cache states, traces of hits and misses, and
execution times.

In this talk will present the foundations and architecture of the
CacheAudit platform, and the results we obtain when analyzing library
implementations of symmetric cryptosystems such as AES or Salsa. I
will also show how CacheAudit can be used for engineering proofs of
security of leakage-resilient cryptosystems on platforms with
concurrency and caches.