A formal approach to property testing in causally consistent distributed traces

Abstract

A formal framework for the analysis of execution traces collected from distributed systems at run-time is presented. We introduce the notions of event and message traces to capture the consistency of causal dependencies between the elements of a trace. We formulate an approach to property testing where a partially ordered execution trace is modeled by a collection of communicating automata. We prove that the model exactly characterizes the causality relation between the events/messages in the observed trace and discuss the implementation of this approach in SDL, where ObjectGEODE is used to verify properties using model-checking techniques. Finally, we illustrate the approach with industrial case studies.

Techniques and Toolset for Conformance Testing against UML Sequence Diagrams

“…Regarding the representation of execution traces, we follow the approach of [ 13], which uses a single event to model synchronous (instantaneous) communication and a pair of send and receive events to model asynchronous (non-instantaneous) communication. In [ 13], it is also proposed a translation procedure of a partially ordered execution trace, containing both synchronous and asynchronous communications (but not interaction operators), into a system of communicating automata, with one automaton per process (lifeline) and one 'message delay' automaton per asynchronous communication, which product yields the possible ways of interleaving events. Despite the different outputs, our translation procedure follows some of the principles of their approach, adding the support for interaction operators and other UML SDs' features.…”

Modular System Verification by Inference, Testing and Reachability Analysis

“…The work [17] relies on observed traces to construct automata models of communicating components which are then model-checked using user defined properties. An objectflattening technique [11] is used to collect system behavior and then invariants are calculated on the behaviors to check against the new version of the system.…”

“…This work is more related to regression testing. Moreover, the system behavior is observed while the system is running as in [17]. In this paper, we rely on testing communicating system by stimulating it through external inputs and then use the observations to obtain tunable approximated models.…”

Reverse Engineering Models from Traces to Validate Distributed Systems – An Industrial Case Study

“…Such monitoring is often used in telecommunication networks for debugging purposes, relying often on protocol analyzers and network tracers. In the chosen industrial case study, we focus on the applicability of a model checking based validation technology we developed earlier [5] to end-to-end testing of a 3GPP UMTS radio network [11]. Our goal is not only to demonstrate to testers that MDD can be of help to them, but also to identify research direction for improving the applicability of this technology.…”

“…This type of analysis is also known as runtime verification or passive testing of distributed systems [2]. In this context, an approach has been developed [5,6,7] that takes as input an execution trace of the system obtained during the execution of a test scenario. Such a recorded trace, which contains a causally ordered sequence of send/receive events or messages exchanged between system components, is reverse-engineered to produce a model in the form of a system of communicating state machines that reflects the system behavior as it occurred during the execution of that particular test scenario.…”