An Optimal Checkpointing Interval for Real-Time Systems

The application of checkpointing as a fault-tolerance measure for real-time services (i.e., services that are restricted by deadlines) raises differ-ent problems than dealing with non-real-time services. Therefore, different criteria are necessary for the assessment of checkpointing in such an envi-ronment. The probability of correct execution before the deadline even in presence of faults (responsiveness) is such a criterion for real-time services. Many checkpointing methods use equidistant checkpoints. For such methods, the number of checkpoints taken during execution time is the con-trolled parameter for maximizing responsiveness. In this paper, we give an exact analysis of the responsiveness of real-time services with checkpoint-ing, using a Poisson fault model and fault detection by acceptance tests at checkpointing time. A simple algorithm is given to obtain a checkpoint in-terval that maximizes the responsiveness of a service. The impact of differ