Abstract
In the recent scenario, capacity of the storage system increases significantly due to increasing demands. Data availability is a major concern in the context of storage systems. In order to achieve high data availability, a Maximum Distance Separable (MDS) code such as Row-Diagonal Parity (RDP) is used in RAID 6 storage systems. In order to ensure continuous availability of data, even a failure in a single disk cannot be tolerated and hence it must be recovered at the earliest. Existing RDOR-RDP, a modified scheme of RDP that achieves minimal disk I/O access and reduced recovery time when compared with conventional scheme but yet fails to achieve load balancing during the recovery process. In this work, Parity declustering based load balancing is proposed, which provides optimal recovery time when compared to RDOR-RDP.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
