Robust Declassification by Incremental Typing

Abstract

Security of software systems has to be preserved while they grow and change incrementally. The problem is to make the analysis of their security properties adhere to such a development. In particular we concentrate here on static type systems. Given a non-incremental type system, the algorithm we propose permits using it incrementally, so avoiding to develop new incremental versions of it. As a proof-of-concept we show how our technique permits an incremental checking of non-interference with robust declassification, starting from the classical type system by Myers, Sabelfeld and Zdancewic.