In partial fulfillment of the Requirements for the Degree of
Master of Science


Ling Zhang
will defend her thesis

Algorithm and Layout for On-Line Disk Failure Recovery



Abstract

In this thesis we attempt to identify and evaluate the primary techniques that can be used to address the failure recovery problem in redundant disk arrays, discussing both the organization of data and parity in the array and the algorithms used to recover from failure of two successive disks. The two key issues addressed are the data layout, the mapping by which data and parity blocks are assigned to physical disk blocks in an array, and the reconstruction algorithm, which is the technique used to recover data that are lost when two disk failures successively.

The data layout techniques this thesis investigates are variations on the de-clustered parity organization, a derivative of RAID Level 5 that allows a system to trade some of its data capacity for improved failure-recovery performance. But parity de-clustering does not tolerate more than one failure. To address this, we present a layout for two successive disk failures, using our odd-even recovering method and the parity de-clustering organization. This layout has the ability to recover from successive failures of two disks without losing any data, it can reconstruct any successively two-disk failures on-line, and it has minimal impact on system performance as observed by users.

Date: Friday, October 27, 2000
Time: 11:30 AM
Place: 550-PGH



Faculty, students, and the general public are invited.
Thesis Advisor: Dr. Ernst L. Leiss