XIFER: A Software Diversity Tool Against Code-Reuse Attacks

AutorDavi, Lucas; Dmitrienko, Alexandra; Nürnberger, Stefan; Sadeghi, Ahmad-Reza
ArtConference Proceedings
AbstraktThe enormous growth of mobile devices and their app markets has raised many security and privacy concerns. Runtime attacks seem to be a major threat, in particular, code-reuse attacks that do not require any external code injection (e.g., return-to-libc or return-oriented programming). We present, for the first time, a code transformation tool that completely mitigates code-reuse attacks by applying software diversity to the binary at runtime. Our tool XIFER (1) randomly diversifies the code of an application over the entire memory for each invocation, (2) requires no source code or any static analysis, (3) can be applied to both Intel x86 and ARM Linux executables, and (4) induces a negligible runtime overhead of only 1% in average.
In4th ACM International Workshop on Wireless of the Students, by the Students, for the Students (S3 2012)