A month of MrT stats

System Config:
2 Node 3PAR T400 (Max:4)
24GB Cache (Max: 48GB)
32 x 4Gbps Fiber channel host+disk ports (Max: 96 x 4Gbps Fiber Channel)
4 x 1Gbps iSCSI host ports (Max: 16 x 1Gbps iSCSI)
200 x 750GB SATA-II disks [150TB raw] (Max: 640 disks/300TB today, expecting 1280 disks/1200TB in the future)
Software: Thin provisioning, Dynamic Optimization, Virtual Copy(R/W Snapshots)
Wide striping data across every spindle, combination of RAID 5+0(3+1), RAID 5+0(5+1), and RAID 1+0
Consumes: 2x40U cabinet & 2 pairs of 208V 30A active/passive power circuits ( collapsing to single cabinet soon)
Maximum controller performance rated at 160,000 IOPS(to disk), 112,0000 SPC-1 IOPS, 3.2 Gigabytes/second of throughput to the hosts.
(Of course you must have sufficient number/type of disks to reach that performance)

Assuming we stick to SATA disks this system provides linear scalability all the way up to 1,777 spindles, I'm expecting the
system to max out at 1,280 spindles after software and some minor hardware adjustments. The ability to re-stripe data across
all spindles on the fly is critical to scaling incrementally.

Disk Stats

All Fiber ports stats (both Host+Disk combined(and averaged where needed))