Ezine

This article can also be found in the Premium Editorial Download "Storage magazine: Boosting data storage array performance."

Download it now to read this article plus other related content.

To get the best performance from your array, you need to tune it to match its workload.

AN I/O REQUEST STARTS at a host running an application or a service, travels through layers of its operating system to a host bus adapter (HBA) and then hops through the SAN fabric until it reaches the storage subsystem. When the storage subsystem responds, this course is reversed. This circuitous route offers plenty of performance-killing bottlenecks, as well as opportunities to optimize the performance of your storage subsystem to provide better service for all I/O commands.

I/O performance optimization activities for SAN-attached arrays can occur at the following three distinct times:

  • Before the subsystem is purchased when choices about drives, interfaces, cache and data replication features are made.


  • At configuration time when choices are made related to RAID level, RAID segment and stripe size; which drives to use to support a RAID group; which LUNs to assign to a RAID group; as well as cache options for the subsystem, RAID group and/or LUN.


  • At runtime when the subsystem services its assigned applications, choices can be made or modified for cache options, LUN-to-controller affinity, application-to-LUN assignment and LUN-to-RAID group assignment.

    Requires Free Membership to View

    Login

This was first published in January 2006

Join the conversationComment

Share
Comments

    Results

    Contribute to the conversation

    All fields are required. Comments will appear at the bottom of the article.
    Back to top