Database Pattern

In the Database pattern the drive is processing a stream of requests to read and write 8KB random-address data blocks. The ratio of read to write requests is changing from 0% to 100% throughout the test while the request queue size varies from 1 to 256.

We’ll discuss the results of each drive individually.

Hitachi has made some progress in the new drive’s firmware, correcting the defects of the 7K100. Take note that the 7K200 copes betters with the hardest types of load when there is about the same share of writes and reads. The overall performance at long request queues has improved greatly.

First of all, the radically different behavior of the MHV2 BT must be noted: this hard disk has too low results at loads with a low percentage of writes even for its low spindle rotation speed. The other two models are similar, the MHW2 BJ being faster at every load proportionally to the difference in their spindle rotation speeds.

Among Seagate’s drives, the 7200.10 model copes best with the middle part of the diagram, but loses to the older 7200.7 at long queue depths when there are only reads or writes to be done. Interestingly, it is the 2.5” 7200.1 that is the best HDD at 100% writes. The same HDD is considerably slower than its Seagate mates at medium queue depths (4 and 16) when there are more reads than writes.

Comparing all the participating HDDs, Hitachi’s are better than the others when reading at a short queue depth, and the Seagate team are better when reading at long queue depths. Hitachi’s HDDs are also superior at random writing but are closely followed by the 7200rpm Fujitsu MHW2 BJ. When the amount of writes and reads is about the same, the Hitachi 7K200 wins at short queues, and the Seagate 7200.10 is the best at long queues.