IEEE Santa Clara Valley Section  

IEEE Magnetics Society

IEEE Magnetics Society
Santa Clara Valley Chapter
Meeting Presentation Summary




Tuesday, February 19th, 2013

Western Digital, 1710 Automation Parkway, San Jose, CA
Directions and Map
Cookies, Conversation & Pizza too at 7:00 P.M.
Presentation at 7:30 P.M.

HDD Ghosts, Goblins and Failures
in RAID Storage Systems

Dr. John Elerath
Elerath Reliability Consulting Services

Abstract

Data storage architectures employing Redundant Arrays of Independent Disks (RAIDs) were created to improve reliability over the typical non-redundant hard disk drive (HDD) used in most personal computers. As the number of HDDs in a storage system increases, the probability of failure and subsequent data loss increase as well. As background, this presentation briefly introduces the popular RAID architectures and then explores the causes of HDD failures. The effects of the failures are then grouped into two major areas, each of which has a different impact on RAID reliability. As HDD capacities continue to increase beyond the multi-terabyte capacity, the number and frequency of latent media defects will increase disproportionately when compared to the number and frequency of HDD operational failures. RAID system architects must understand these effects when designing mitigating processes, such as RAID scrubbing, and the HDD magnetics designers must try to improve designs to keep the number of, and rate of added media defects low.


Copy Of Presentation


Biography

Photo of Dr. Jon Elerath Dr. Jon Elerath is a reliability consultant, specializing in reliability of electro-mechanical systems. Previously he was a senior staff reliability engineer at SolFocus, a leading supplier of High Concentrator Photovoltaic (HCPV) systems. He has focused on hard-disk drive reliability for more than half his 35-plus-year career, which included positions at NetApp, General Electric, Tegal, Tandem Computers, Compaq, and the IBM disk drive group where he applied reliability techniques to HCPV trackers, hard disk drives, RAID data storage systems, nuclear safety systems of fast breeder reactors, electronics and robotics of plasma-etching equipment, and fault-tolerant computers. A major area of interest is reliability of RAID storage systems. He has authored over 30 technical publications including “Hard-Disk Drives: The Good, the Bad, and the Ugly” (ACM, 2009) and “A Highly Accurate Method for Assessing Reliability of Redundant Arrays of Inexpensive Disks” (IEEE, 2008). He is active in writing IEEE reliability standards and was Chairman of the Reliability Committee for the International Disk drive Equipment and Materials Association (IDEMA) for over 10 years.
Jon received his BSME and MS Reliability degrees from the University of Arizona, and received his PhD in Mechanical Engineering from the University of Maryland.

 

Contact:



Return to SCV Magnetics Society Homepage

SCV Magnetics Society Webmaster: Tom Gardner (SCVMagSociety@gmail.com)
Last updated on