Foundations and applications for secure triggers

Abstract

Imagine there is certain content we want to maintain private until some particular event occurs, when we want to have it automatically disclosed. Suppose, furthermore, that we want this done in a (possibly) malicious host. Say the confidential content is a piece of code belonging to a computer program that should remain ciphered and then “be triggered” (i.e., deciphered and executed) when the underlying system satisfies a preselected condition, which must remain secret after code inspection. In this work we present different solutions for problems of this sort, using different “declassification” criteria, based on a primitive we call secure triggers . We establish the notion of secure triggers in the universally composable security framework of Canetti [2001] and introduce several examples. Our examples demonstrate that a new sort of obfuscation is possible. Finally, we motivate its use with applications in realistic scenarios.