There are several key factors to consider
when comparing the specs of HDDs:
(1) rotation speed in rpm, which directly affects
time to do one-half rotation aka average rotational
latency e.g. 5400, 7200, 10000 and 15000 rpm;
(2) time to move the READ/WRITE heads from
track-to-track and from outermost to innermost tracks;
for larger diameter platters the latter "access times"
can be very large;
(3) recording density: either horizontal or
perpendicular magnetic recording ("PMR");
(4) size of the internal cache e.g. 2MB, 8MB, 16Mb and 32MB;
(5) speed of the interface e.g. 133MB/sec, 150MB/sec, 300MB/sec,
400Mb/sec (fibre channel) and the upcoming SATA/6G = 600MB/sec;
(6) the esoteric features like Native Command Queuing (NCQ),
SAS (Serial Attached SCSI), staggered spin-up, time-limited
error recovery, vibration tolerance for RAID setups, etc.
WD's Raptors do not use PMR, and for that reason
WD's models that do use PMR are now about equal
in raw transfer speeds to the Raptor's raw transfer speeds:
"raw transfer speed" here means the rate at which
data bits pass under the READ/WRITE heads.
There is a premium to pay for 10,000 rpm drives;
you may find that the newest 7,200 rpm HDDs with PMR --
like WD's RE2 and RE3 models -- perform quite well
when hosting the OS:
http://www.wdc.com/en/products/productcatalog.asp?language=en
But, for consistent performance overall,
WD's latest VelociRaptors are still very competitive
for hosting the OS, their 10,000 rpm rotation
speed necessarily reduces rotational latency,
and their small platter diameter means that
outer-to-inner head movements are minimal:
http://www.newegg.com/Product/Product.aspx?Item=N82E16822136296
Just use your calculator to compute the time
to perform one-half rotation at 7,200 and 10,000 rpm:
that is "average rotational latency".
MRFS