Abstract

Software systems often share common vulnerabilities that allow a single attack to compromise large numbers of machines (“write once, exploit ev-erywhere”). Borrowing from biology, several researchers have proposed the introduction of artificial diversity in systems as a means for countering this phenomenon. The introduced differences affect the way code is constructed or executed, but retain the functionality of the original system. In this way, systems that exhibit the same functionality have unique characteristics that protect them from common mode attacks. Over the years, several such ap-proaches have been proposed. We examine some of the most significan