Abstract :
Cloud storage systems have been rapidly expanding, which provide users with easy access, low maintenance, flexibility, and scalability. In cloud storage systems, the failure tolerance of storage nodes is essential. As the rapid growth of data, fault tolerance techniques in cloud storage systems have been transitioning replication strategy to erasure codes to achieve higher storage efficiency. Classical erasure codes like Maximum-Distance-Separable (MDS) codes have been the most widely adopted, due to their optimal storage efficiency. However, MDS codes are highly suboptimal because of their high repair cost in failures of storage nodes. For high repair efficiency, one new family of erasure codes called Repair Locality codes are proposed recently. Prior work on Repair Locality codes focused on fundamental bounds, but there were few explicit constructions. In this paper, we present a new class of Repair Locality codes with repair optimality and simple construction for cloud storage systems. Final, we prove the repair optimality of our codes.
