Distributed and Parallel Databases

, Volume 2, Issue 3, pp 243–260

Performance and design evaluation of the RAID-II storage server

Authors

  • Peter M. Chen
    • Department of Electrical Engineering and Computer ScienceUniversity of Michigan
  • Edward K. Lee
    • Digital Equipment Corporation Systems Research Center
  • Ann L. Drapeau
    • Computer Science Division, Department of Electrical Engineering and Computer ScienceUniversity of California at Berkeley
  • Ken Lutz
    • Computer Science Division, Department of Electrical Engineering and Computer ScienceUniversity of California at Berkeley
  • Ethan L. Miller
    • Computer Science Division, Department of Electrical Engineering and Computer ScienceUniversity of California at Berkeley
  • Srinivasan Seshan
    • Computer Science Division, Department of Electrical Engineering and Computer ScienceUniversity of California at Berkeley
  • Ken Shirriff
    • Computer Science Division, Department of Electrical Engineering and Computer ScienceUniversity of California at Berkeley
  • David A. Patterson
    • Computer Science Division, Department of Electrical Engineering and Computer ScienceUniversity of California at Berkeley
  • Randy H. Katz
    • Computer Science Division, Department of Electrical Engineering and Computer ScienceUniversity of California at Berkeley
Article

DOI: 10.1007/BF01266330

Cite this article as:
Chen, P.M., Lee, E.K., Drapeau, A.L. et al. Distrib Parallel Databases (1994) 2: 243. doi:10.1007/BF01266330

Abstract

RAID-II is a high-bandwidth, network-attached storage server designed and implemented at the University of California at Berkeley. In this paper, we measure the performance of RAID-II and evaluate various architectural decisions made during the design process. We first measure the end-to-end performance of the system to be approximately 20 MB/s for both disk array reads and writes. We then perform a bottleneck analysis by examining the performance of each individual subsystem and conclude that the disk subsystem limits performance. By adding a custom interconnect board with a high-speed memory and bus system and parity engine, we are able to achieve a performance speedup of 8 to 15 over a comparative system using only off-the-shelf hardware.

Keywords

RAIDHIPPIdisk arraycrossbarmass storage systemI/Oinput/output

Copyright information

© Kluwer Academic Publishers 1994