Analysis of fault tolerance improvement in distributed over network data storage

The research object is distributed software-defined storage system and fault tolerance improvement in distributed over network data storage. The article is devoted to solving the optimization problem regarding system performance and ensuring a high level of data storage reliability. The required MTBF (mean time between failures) level of the data storage is determined. The presented mathematical model, obtained from extensive testing and simulation, calculates the system availability for various values of the replication factors. Conclusions are drawn, including the prospects for further research.