Practical implications and requirements of diversifying interpreted languages

Abstract

Instruction set randomization (ISR) provides a strong defense against all types of injection attacks, especially in interpreted environments. However, fully enabling a system to benefit from language interpreters that support programs diversified with ISR requires several alterations and considerations. In this paper we identify core challenges related to enabling system-wide interpreter diversification. We also propose possible solutions to each challenge and expand upon the existing diversification schemes for interpreted languages.