Parametric Trace Slicing and Monitoring

Trace analysis plays a fundamental role in many program analysis approaches, such as runtime verification, testing, monitoring, and specification mining. Recent research efforts bring empirical evidence that execution traces are frequently comprised of many meaningful trace slices merged together, each slice corresponding to instances of relevant parameters. For example, a Java program creating iterators i1 and i2 over vector v1 may yield a trace “create〈v1 i1〉 next〈i1〉 create〈v1 i2〉 update〈v1〉 next〈i1〉” parametric in vector v and iterator i, whose slices corresponding to parameter instances “v, i → v1, i1” and “v, i → v1, i2” are “create〈v1 i1〉 next〈i1〉 update〈v1〉 next〈i1〉” and, respectively, “create〈v1 i2〉 update〈v1〉 ”. Several current trace analysis techniques and systems allow the specification of parametric properties, and the analysis of execution traces with respect to each instance of the parameters. However, the current solutions have limitations: some in the specification formalism, others in the type of trace they support; moreover, they share common notions, intuitions, even techniques and algorithms, suggesting that a fundamental study and understanding of parametric trace analysis is needed. This foundational paper gives the first solution to parametric trace analysis that is unrestricted by the type of parametric property or trace that can be analyzed. First, a general purpose parametric trace slicing technique is discussed, which takes each event in the parametric trace and distributes it to its corresponding trace slices. This parametric trace slicing technique can be used in combination with any conventional, non-parametric trace analysis technique, by applying the later on each trace slice. As an instance, a parametric property monitoring technique is then presented, which processes each trace slice online. Thanks to the generality of parametric trace slicing, the parametric property monitoring technique reduces to encapsulating and indexing unrestricted and well-understood nonparametric property monitors (e.g., finite or push-down automata). The presented parametric trace slicing and monitoring techniques have been implemented and extensively evaluated. Measurements of runtime overhead confirm that the generality of the discussed techniques does not come at a performance expense when compared with existing parametric trace monitoring systems.