So are you guys suggesting you short stroke the drives so that only the inermost section of the drive is used? I could see that helping access times slightly but slowing down overall transfer speeds. I mostly only hear of short stroking a drive so it only writes on the outermost parts of the platters, thus your transfer speeds will be faster due to the linear velocity being faster the further you get from the center of the platter. Latency wise, you have to wait for the platter to rotate around to the given read/write position. This gives you the following formula: 1 / (RPM / 60 seconds). Therefore a 7200 RPM drive will give you pretty much on avg. 8.333.... ms access times. A 10000 RPM drive then reduces that down to 6 ms. Of course if you factor in native command queuing then these numbers can be reduced more, and those numbers were assuming worst case scenario where the drive has to perform a full rotation before the heads are positioned at the correct address. But still, RPM's rule latencies.