Xiao Qin  [Research Interests | Current Projects | Completed Projects | Research Sponsors | Professional Activities | Back to Home]

MINT: Mathematical Reliability Models for Energy-Efficient Parallel Disk Systems 

[ News | People | Publications | Project Progress | Talks | Implementation | NSF Award Abstract | Project Home ]


MINT Related News

Best Practices: Enterprise SATA Deployment With High Performance And Reliability
12/28/2007 

Application, availability and compliance requirements are driving the demand to store more data online. Significant growth in the quantity of disk-based information that organizations now manage is a result of growing file sizes, number of applications installed, multi-copy data protection practices and the requirement to bring/retain increasing amounts of archived data online. Attempting to address this capacity growth solely with enterprise-class, Fibre Channel (FC) disk would have been cost prohibitive. Conversely, utilizing FC disk would have provided the high performance and reliability that enterprise users have come expect from their arrays.

Over the past five years, server-class SATA drive¡¯s compelling cost/GB has fueled wide enterprise adoption to address challenging data growth and capacity requirements. During this period, the enterprise has deployed SATA as primary storage to drive applications and as secondary storage to create online archives and to backup and restore. Regardless of how SATA is being used, storage administrators have been conditioned to accept that lower performance and reliability are necessary evils in order to enjoy SATA¡¯s cost/GB benefits. Downtime, application crashes, and latency associated with SATA drive failures; silent data corruption and degraded array performance during long drive rebuild events are common complaints.


Contact Us

Updated on 12/28/2007