Here is SQLIO from OCZ Revo-X2 (

Dual Six-Core Xeon Nehalem -- This is a PCI-E 480GB Running on W2K8 R2 with the Win 7 drivers (no issues).

Sequential 8K IOPS are 79,000 give or take

Random 8K IOPS are just Shy of 60,000.

We have used both PCI-E Solid State and SATA/SAS drives. We are using MLC and E-MLC in our production appliances as we find that we will never get near the wear out for the writes in the course of 5 years expected life. Also -- note that the size is 480GB. The drive actually has 512 GB of MLC on it. This allows for spare NAND in the event part of the drive gets weak. What would make this drive even better is the addition of a battery / supercap to allow write flushing in the event of a power failure. However all of our appliances wind up in data centers and power outages are pretty rare. So the drive sells for about 1,500 and so to have it mirrored you are looking at $3,000 -- now to get the same IOPS in a spinning disk SAN array you would need to have -- 250 spindles - if they are mirrored then that is 500 Hard disks -- which is a very, very expensive proposition.

Bottom line is that we use the SAN for backup etc. but not for production data. The speed is absolutely addictive and is fantastic for BI / Reporting.

The other factor that isn't discussed much is the simplicity of the operation. No HBA'S, no storage switch, no dealing with Mr. SAN Manager who in my experience can be a bigger bottleneck than the hardware. At the end of the day you have a democratization of storage that adds to the agility of any project.

Here are the results from SQLIO --

sqlio v1.5.SG

4 threads writing for 30 secs to file S:\testfile.dat

using 8KB sequential IOs

enabling multiple I/Os per thread with 8 outstanding

using specified size: 500 MB for file: S:\testfile.dat

CUMULATIVE DATA:

throughput metrics:

IOs/sec: 79244.51

MBs/sec: 619.09

latency metrics:

Min_Latency(ms): 0

Avg_Latency(ms): 0

Max_Latency(ms): 14

histogram:

ms: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24+

%: 99 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

threads reading for 30 secs from file S:\testfile.dat

using 8KB random IOs

enabling multiple I/Os per thread with 8 outstanding

using specified size: 500 MB for file: S:\testfile.dat

initialization done

CUMULATIVE DATA:

throughput metrics:

IOs/sec: 58097.21

MBs/sec: 453.88

latency metrics:

Min_Latency(ms): 0

Avg_Latency(ms): 0

Max_Latency(ms): 12

histogram:

ms: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24+

%: 97 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0