Reachability analysis on distributed executions

Abstract

The paper presents basic algorithms for trace checking of distributed programs. In distributed systems, detecting global properties requires a careful analysis of the causal structure of the execution. Based on the on-the-fly observation of the partial order of message causality, we show how to build the lattice of all the reachable states of the distributed system under test. The regular structure of this graph makes it possible to build it with a quasi-linear complexity, which improves substantially the state-of-the-art.