State-Driven Testing of Distributed Systems

In distributed systems, failures are often caused by software faults that manifest themselves only when the system enters a particular, rarely occurring system state. It thus becomes important to identify these failure-prone states during testing. We propose a state-driven testing approach for distributed systems, able to execute tests in hard-to-reach states in a repeatable and accurate way. Moreover, we present the implementation and experimental evaluation of the approach in the context of a fault-tolerant flight data processing system. Experimental results confirm the feasibility of the approach, and the accuracy and reproducibility of tests.