RSS_RAID a Novel Replicated Storage Schema for RAID System

Abstract

Nowadays, due to the emergence of big data and its critical rule in most applications, data availability is big concern. For some applications, even the lack of a piece of information is not acceptable and have great drawbacks on results. Therefore, the storage reliability and guarantee of rapid data recovery is one of the main concerns. In case of disk failure, due to high storage volume, a lot of time is required for data recovery and this greatly decreases data availability. A new storage schema named RSS-RAID is presented in this paper. In this schema, disks are divided into groups with the same number of disks and data are stored as strips between disks with a particular algorithm based on a reversible hashing function. One advantage of proposed schema in comparison with similar models is that the location of the blocks is pre-known and when disk failure happens, number of missing blocks is clearly known and recovery algorithm do not need to search copies of missed blocks on the replica disks to recover them. This increases the recovery speed and causes more availability of data. Proposed schema is completely fault tolerant in case of one disk failure and fault tolerant against concurrent failure of up to three disks in the case that failed disks are located in the same group.