Specification mining in concurrent and distributed systems

Abstract

Dynamic specification mining involves discovering software behavior from traces for the purpose of program comprehension and bug detection. However, in concurrent/distributed programs, the inherent partial order relationships among events occurring across processes pose a big challenge to specification mining. A framework for mining partial orders that takes in a set of concurrent program traces, and produces a message sequence graph (MSG) is proposed. Mining an MSG allows one to understand concurrent behaviors since the nodes of the MSG depict important "phases" or "interaction snippets" involving several concurrently executing processes. Experiments on mining behaviors of fairly complex distributed systems show that the proposed miner can produce the corresponding MSGs with both high precision and high recall.