Our database test pattern runs 67% read and 33% write operations, 100% random and purely 8KB block size. In this test, we can see performance of 2.4x if the test volume is limited to only 5GB. Here, the MaxIQ feature can easily buffer all the RAID data onto the X25-E drive.
The performance benefits are still 59%, 23%, and 15% for 25, 50, and 100GB volume sizes. Keep in mind that the test pattern creates a significant amount of write access as well, which MaxIQ cannot accelerate.
The file server test is also 100% random, but it uses 80% reads and only 20% writes. Block sizes are distributed between 512 bytes and 64KB. Please look at this article for details on the IOMeter performance patterns.
In this case, the benefits of MaxIQ over the regular RAID 0 array are 5x, 3.85x, 1.69x, and 1.31x performance boosts for 5, 25, 50, and 100GB on the array test volumes. Clearly, the benefits are noticeable. Let’s look at the Web server results, which are fully based on random read operations.