News

Location: home>News Events>News
The paper entitled “A Hybrid Approach of Failed Disk Recovery Using RAID-6 Codes: Algorithms and Performance Evaluation” is accepted by the ACM Transactions on Storage
( 2011-04-08 )

 

On May 27, 2011, our paper entitled “A Hybrid Approach of Failed Disk Recovery Using RAID-6 Codes: Algorithms and Performance Evaluation”(Liping XIANG, Yinlong XU, John C.S. LUI, Qian CHANG, Yubiao PAN, and Runhui LI) was accepted for publication in the ACM Transactions on Storage(TOS). This paper was accomplished by Professor Yinlong Xu’s research group and John C.S. Lui of The Chinese University of Hong Kong. ACM TOS is a premier journal for publishing advancements in storage research and practice. Each year, ACM TOS only accepts no more than 20 research articles. The scope of TOS includes data availability, network protocols, resource management, data backup, replication, recovery, devices, security, theory of data coding, densities, energy-efficiency, and so on.

 

Today, an important requirement of building large storage systems is to make sure information being reliable and available even under the presence of component failures. Therefore, to guarantee that a storage system operates at a high reliability level, a key approach is to repair and recover a failed component as quickly as possible. The paper “A Hybrid Approach of Failed Disk Recovery Using RAID-6 Codes: Algorithms and Performance Evaluation” addresses the problem of how to design efficient recovery schemes for single disk failure in storage systems that use RAID-6 codes. This paper first considers the problem of optimizing the recovery process of single disk failure for RAID-6 codes and proposes a hybrid recovery approach for single failure recovery. Based on the hybrid approach, two recovery algorithms were proposed for RDP code and EVENODD code. The proposed recovery algorithms have the following properties: (1) “read optimal” in the sense that they issue the minimum number of disk reads to recover the failed disk; (2) “load balancing” in that all surviving disks will be subjected to the same amount of additional workload in rebuilding the failed disk.



Copyright 1995-2011, National High Performance Computing Center at Hefei
E-mail: ylxu@ustc.edu.cn Fax:86-0551-3601013