This paper proposes and evaluates the architecture of RAID level 4 disk arrays. In this architecture, one disk array contains two parity disks, by using a spare disk. Two methods are considered for utilization of these parity disks. One is an extended parity method in which a native parity disk is extended to two parity disks, and the other is a dual parity method in which two disks have the same parity. We construct array models and carry out the performance simulation for the variable write ratio and command arrival rate in normal mode. Array performance of the two methods is evaluated in comparison with other methods for level 5 disk arrays. We also discuss the recovery overhead by estimating the number of blocks to be rebuilt and/or copied back for each model. It is concluded that the extended and dual parity methods for level 4 can stand comparison with sparing methods for level 5 in normal mode of small arrays, and also have the advantage of the recovery overhead.
[1]
Garth A. Gibson,et al.
RAID: high-performance, reliable secondary storage
,
1994,
CSUR.
[2]
Garth A. Gibson,et al.
Parity declustering for continuous operation in redundant disk arrays
,
1992,
ASPLOS V.
[3]
Randy H. Katz,et al.
A case for redundant arrays of inexpensive disks (RAID)
,
1988,
SIGMOD '88.
[4]
Daniel P. Siewiorek,et al.
Fast, on-line failure recovery in redundant disk arrays
,
1993,
FTCS-23 The Twenty-Third International Symposium on Fault-Tolerant Computing.
[5]
Jai Menon,et al.
Comparison of sparing alternatives for disk arrays
,
1992,
ISCA '92.
[6]
John A. Chandy,et al.
Failure evaluation of disk array organizations
,
1993,
[1993] Proceedings. The 13th International Conference on Distributed Computing Systems.