Proceedings of the Hawaii International Conference on System Sciences, January 1993, pp. 40­49.

Disk Subsystem Load Balancing: Disk Striping vs. Conventional Data Placement

Gregory R. Ganger, Bruce L. Worthington, Robert Y. Hou, Yale N. Patt

Department of Electrical Engineering and Computer Science
University of Michigan,
Ann Arbor, MI 48109­2122

Abstract

The I/O subsystem is becoming a major bottleneck in an increasing number of computer systems. To provide imroved I/O performance, as well as to accommodate growing storage requirements, disk subsystems are increasing in size. A major hurdle to obtaining the performance availble from these large disk subsystems is load imbalance, or disk skew. Dynamic data placement, the conventional load balancing technique, is usually inadequate to deal with load imbalance because it is forced to accept atomic data sets with rapidly changing access patterns. We name this rapid fluctuation floating load imbalance and distinuish it from the conventional view of load imbalance, reerred to as fixed load imbalance. Dynamic data placement also becomes increasingly difficult as the number of disks in the subsystem grows. Disk striping at a high granularty is suggested as a solution to floating load imbalance, the atomic data set problem and the complexity of balaning large disk subsystems. Disk striping uniformly spreads data sets across the disks in the subsystem and essentially randomizes the disk accessed by each request. This ranomization effectively handles both fixed and floating load imbalance. Unlike dynamic data placement, disk striping does not become more complex as the number of disks inreases. While a more optimal load balance may be possible for some very well understood and well­controlled environments, disk striping should provide significantly improved load balance with reduced complexity for many applications. This improvement will result in shorter response times and higher throughput.

Available Online