Abstract
A group of disks may be interleaved to speed up data transfers in a manner analogous to the speedup achieved by main memory interleaving. Conventional disks may be used for interleaving by spreading data across disks and by treating multiple disks as if they were a single one. Furthermore, the rotation of the interleaved disks may be synchronized to simplify control and also to optimize performance. In addition, check- sums may be placed on separate check-sum disks in order to improve reliability. In this paper, we study synchronized disk interleaving as a high-performance mass storage system architecture. The advantages and limitations of the proposed disk interleaving scheme are analyzed using the M/G/1 queueing model and compared to the conventional disk access mechanism.

This publication has 7 references indexed in Scilit: