Towards Distributed Runtime Monitoring with C plus plus Contracts

Design by Contract supports the specification of the intended behavior of a software component at its interfaces. Out of these specifications, runtime monitors can be generated. While those runtime monitors check the correct behavior of the software component, they are usually executed on the same processor as the monitored software and therefore prone against the same possible hardware failures. This work proposes an approach for distributed runtime monitors in order to increase the overall robustness of the system. Runtime monitors will be derived from contract specifications and then check the fulfillment of assumptions and guarantees during runtime. While deployed onto a multi-processing system the contract checks can detect permanent hardware failures. Also, there are some scenarios where the parallel execution of the contract checks might be sensible. Therefore, this approach arguably increases the safety benefit of contracts and at the same time, leaves potential to further decrease the added execution time overhead. A first prototype of our approach was implement and tested. Experiments regarding the performance showed that the execution time doubled when distributed runtime monitors were used.