A Hypergraph Theory-based Approach to Distributed System Reliability Evaluation

The paper deals with the problem of fault tolerance in distributed systems using the theory of hypergraphs. This theory appears to be extremely suitable for modelling a distributed system and the redundancy mechanisms adopted to improve its fault tolerance. The paperprovides theoretical conditions for calculation of the probability of system functioning according to the number of resources, the number of applications and the probability of a fault occurring in each resource.On the basis of the results obtained, an algorithmic solution to the problem is proposed. The computational complexity of the solution isreduced by an algorithm proposed by the authors, also based on the theoretical considerations put forward in the paper.