Securing Causal Relationships in Distributed Systems

Abstract

In a distributed system, it is often important to detect the causal relationships between events, where event e(1) is causally before event e(2) if e(1) happened before e(2) and could possibly have affected the occurrence of e(2). In this paper we argue that it can be essential to security that a process determine, in the face of malicious attack, how two events are causally related. We formulate attacks on causality detection in terms of causal denial and forgery, formalize possible security goals with respect to causality, and present simple algorithms to attain these goals in some situations.