Enterprise Storage Bench - Microsoft SQL WeeklyMaintenance

Our final enterprise storage bench test once again comes from our own internal databases. We're looking at the stats DB again however this time we're running a trace of our Weekly Maintenance procedure. This procedure runs a consistency check on the 30GB database followed by a rebuild index on all tables to eliminate fragmentation. As its name implies, we run this procedure weekly against our stats DB.

The read:write ratio here remains around 3:1 but we're dealing with far more operations: approximately 1.8M reads and 1M writes. Average queue depth is up to 5.43.

Microsoft SQL WeeklyMaintenance - Average Data Rate

We don't see perfect scaling going from 4 to 8 controllers but the performance gains are tangible: +42% over the RevoDrive 3 X2 and nearly 3x the performance of a single Vertex 3.

Microsoft SQL WeeklyMaintenance - Disk Busy Time

Microsoft SQL WeeklyMaintenance - Average Service Time

Average service time continues to be where the Z-Drive R4 really dominates. The use of 8 controllers in parallel appears to be able to significantly reduce average service times when queue depths skyrocket. The R4 CM88 is now over two orders of magnitude (136x) faster than a single Vertex 3—and 227x faster than the Intel X25-E. Again we see that the RevoDrive 3 X2 is much slower than it should be here, possibly pointing at a firmware bug or some other enhancement on the Z-Drive R4.

Enterprise Storage Bench - Microsoft SQL UpdateDailyStats Final Words
Comments Locked

57 Comments

View All Comments

  • geddarkstorm - Wednesday, September 28, 2011 - link

    ^ This
  • cervantesmx - Wednesday, September 28, 2011 - link

    I agree 100%
  • GTRagnarok - Tuesday, September 27, 2011 - link

    "We have a preproduction board that has a number of stability & compatibility OCZ tells us will be addressed..."

    I think a word is missing here.
  • icrf - Tuesday, September 27, 2011 - link

    Also, you missed the protocol on the last link on the first page (the one to ssd bench) and it 404's now
  • Anand Lal Shimpi - Tuesday, September 27, 2011 - link

    Fixed both! Thank you!

    Take care,
    Anand
  • FATCamaro - Tuesday, September 27, 2011 - link

    OOH OOH Let me guess!!!
    Never??

    As in :
    "We have a preproduction board that has a number of stability & compatibility OCZ tells us will NEVER be addressed..."
  • vodkapls - Tuesday, September 27, 2011 - link

    Isn't it the fact that the revodrive 3 x2 use asynchronous memory that makes it so much slower than the r4 ?
  • Anand Lal Shimpi - Tuesday, September 27, 2011 - link

    Great catch! I hadn't even thought of that but it's definitely a possibility :)

    Take care,
    Anand
  • jebo - Tuesday, September 27, 2011 - link

    I just can't take OCZ seriously from a reliability standpoint. I would love to know what the failure rate is like on OCZ's desktop offerings. I personally am in the process of my 3rd RMA of an OCZ SSD during the past 2 years.

    I think Intel, Crucial (or, judging by the last review, Samsung) will make my next SSD. I can only rebuild windows and piece together backups so many times before I say enough is enough.
  • dilidolo - Tuesday, September 27, 2011 - link

    What's the point to develop enterprise product if no enterprise is going to buy?
    I don't think any enterprise will trust OCZ.

Log in

Don't have an account? Sign up now