Abstract

Numerous exploits of client-server protocols and applications involve modifying clients to behave in ways that untampered clients would not, such as crafting malicious packets. In this talk, we summarize our research on a method for verifying the consistency of a client’s messaging behavior with the program it is believed to be running, without knowing all of the client-side inputs driving its behavior.  We then turn our attention specifically to cryptographic protocols, which present unique challenges to our method of behavioral verification.  The toolchain we have constructed for verifying a client’s messages explores multiple candidate execution paths in the client concurrently, an innovation that we show is both specifically useful for cryptographic protocol clients and more generally useful for client applications of other types, as well.  In addition, our toolchain includes a novel approach to symbolically executing the client software in multiple passes that defers expensive functions until their inputs can be inferred and concretized.  We demonstrate client verification on OpenSSL to show that, e.g., Heartbleed exploits can be detected without Heartbleed-specific filtering and within seconds of the first malicious packet, and that verification of legitimate clients can keep pace with, e.g., Gmail workloads.