Joint Selection of Providers and the New Comer for Optimal Repair in Distributed Storage Network with Network Coding

Abstract

In distributed storage networks (DSN), server failures occur frequently. Although network coding can provide high reliability of storage in DSN, to maintain reliability, a new server, named the new comer, should be firstly implemented to replace the failed server. Then, new comer should download enough data from multiple storage servers, called provider selected from surviving servers to regenerate the lost data. Therefore, the providers, new comer and data routing have a great impact on regeneration time during repair process. In this paper, we jointly consider the providers and new comer selection (JPNS) problem with two objectives of minimizing regeneration time and minimizing transmission cost under the minimum regeneration time guarantee. Specifically, for the first objective, we firstly model the JPNS problem as a mathematical programming and formulate it as an integer linear programming (ILP). We also propose a Heuristic Providers and New Comer Selection (HPNS) algorithm. For the second objective, based on the providers and a new comer selected by the HPNS algorithm, we give a linear programming (LP) to optimally obtain the routing paths and flows from these providers to the new comer with minimum transmission cost. Finally, we conduct extensive simulation experiments and the results show the effectiveness of the proposed algorithms.