Post reply

SSD Performance & index fragmentation

  • February 15, 2013 at 1:12 pm

    #289473

    I'm not quite sure where is the best place to post this question but here goes.

    My understanding is that if I were to install Windows onto an SSD drive then I would no longer want to defrag the hard disk - I was wondering if I were to start using flash storage arrays of one sort or another (I don't have a variety in mind, just a question out of interest) would index fragmentation still be an issue and need dealing with - presumably it would become an issue eventually but the rewriting activity would need to be minimised?

    Does anyone know of any studies into this sort of thing with SQL Server & SSDs?

    Thanks

  • February 15, 2013 at 2:49 pm

    #1588517

    Brent Ozar and Paul Randal both have posted research and findings on the topic. Google or Bing are your friend there. I just pulled back a bunch of relevant results.

    There are no special teachers of virtue, because virtue is taught by the whole community.
    --Plato

  • February 15, 2013 at 2:50 pm

    #1588520

    Fragmentation, probably not. Low page density (a side effect of fragmentation), hell yes. Having half of your expensive SSD going to waste because of low page density isn't the greatest of ideas.

    Gail Shaw
    Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
    SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability

    We walk in the dark places no others will enter
    We stand on the bridge and no one may pass

  • February 15, 2013 at 3:26 pm

    #1588540

    Thanks

  • February 15, 2013 at 3:27 pm

    #1588541

    GilaMonster (2/15/2013)

    Fragmentation, probably not. Low page density (a side effect of fragmentation), hell yes. Having half of your expensive SSD going to waste because of low page density isn't the greatest of ideas.

    I agree. I think Jonathan Kehayias did a blog post about that (and also about sequential IO actually being faster on SSDs too, just like on rotating media).

    Best,
    Kevin G. Boles
    SQL Server Consultant
    SQL MVP 2007-2012
    TheSQLGuru on googles mail service

  • February 25, 2013 at 2:23 pm

    #1590942

    Yes, SSD's (and SSD's in RAID configurations) are indeed still faster for sequential access than they are for random access (significantly, in the case of writes, slightly in the case of reads). Spinning disks are significantly faster in sequential access for both writes and reads.

    Note that I've provided some SSD performance results in the

    RAID and Its impact on your SQL performance

    thread.

    Testing techniques and how to load the output into SQL Server comes from the FusionIO thread.

Viewing 6 posts - 1 through 5 (of 5 total)

You must be logged in to reply to this topic. Login to reply