Efficient Trace Encodings of Bounded Synthesis for Asynchronous Distributed Systems
share
The manual implementation of distributed systems is an error-prone task because of the asynchronous interplay of components and the environment. Bounded synthesis automatically generates an implementation for the specification of the distributed system if one exists. So far, bounded synthesis for distributed systems does not utilize their asynchronous nature. Instead, concurrent behavior of components is encoded by all interleavings and only then checked against the specification. We close this gap by identifying true concurrency in synthesis of asynchronous distributed systems represented as Petri games. This defines when several interleavings can be subsumed by one true concurrent trace. Thereby, fewer and shorter verification problems have to be solved in each iteration of the bounded synthesis algorithm. For Petri games, experimental results show that our implementation using true concurrency outperforms the implementation based on checking all interleavings.
READ FULL TEXT
