State Space Exploration of RT Systems in the Cloud
This work addresses a computational bottleneck for researchers and engineers analyzing real-time systems, but it is incremental as it adapts existing sequential methods to new frameworks.
The paper tackled the state-space explosion problem in real-time systems by comparing distributed and cloud computing approaches for symbolic state-space exploration, showing that cloud approaches are more convenient based on benchmarking tests.
The growing availability of distributed and cloud computing frameworks make it possible to face complex computational problems in a more effective and convenient way. A notable example is state-space exploration of discrete-event systems specified in a formal way. The exponential complexity of this task is a major limitation to the usage of consolidated analysis techniques and tools. We present and compare two different approaches to state-space explosion, relying on distributed and cloud frameworks, respectively. These approaches were designed and implemented following the same computational schema, a sort of map & fold. They are applied on symbolic state-space exploration of real-time systems specified by (a timed extension of) Petri Nets, by readapting a sequential algorithm implemented as a command-line Java tool. The outcome of several tests performed on a benchmarking specification are presented, thus showing the convenience of cloud approaches.