Failure Management and Fault Tolerance Techniques in a Multi-Module Computing System Based on Creation and Replication of Checkpoints

Introduction: in order to enhance the efficiency of target information processing, it is necessary to adopt new approaches to the rapid detection and recovery from failures and faults to minimize the impact of such issues on the overall computing system. Purpose: to outline a technique for failure management and fault recovery in a multi-module computing system. This system implements periodic saving of calculations (checkpoints) and their exchange between all computing modules. Results: the problem of planning such a computing process has been outlined, including the determination of the optimal number and time points for creating checkpoints. The time points for creating checkpoints are determined based on the law of distribution of time points of computing module failures. Practical significance: the results of the simulation modelling calculations conducted as part of the proposed approach demonstrate the feasibility of implementing the proposed technique.