[Info-vax] The Future of Server Hardware?

[glen herrmannsfeldt]

Johnny Billquist <bqt at softjar.se> wrote:

(snip, I wrote)
>> For many years, 3600 RPM disks have been common, so 8.33ms average
>> rotational delay. For servers, now, 7200RPM might be more usual,
>> ro 4.16ms. A million instructions doesn't sound far off.

> 3600 rpm? I didn't even know you could find that anymore. I thought the 
> standard for any cheap home PC was 7200 RPM. 

For current model Seagate, from the seagate.com web site, and
for desktop (presumably 3.5in) drives, the choices seem to be
5900RPM and 7200RPM. For laptop drives, 5400 and 7200RPM.
(Maybe the 5900 is a typo.)  Western Digital has a green
(for eco-computing) line which doesn't indicate the RPM
(at says intellipower) but which some sites claim as 5400.

Looks to me like 5400 is usual for cheaper or more power
efficient home PCs and 7200 for more expensive ones.

> For people who really care 
> about performance, and have money, 15.000 RPM is common, and I think you 
> can also find 18.000 RPM. However, I guess that market is now going for 
> SSD instead.

I am usually a little behind the curve. I believe the fastest
that I have seen (not just read about) is 10,000RPM.

Still, 18,000RPM is five times faster than the big 14 inch
drives of 40 years ago, and 5400 only 1.5 times faster.

> But since CPUs are screaming fast, I'm pretty sure 1 million 
> instructions are probably still true.
> Also, don't forget that rotational latency is just part of the equation. 
> You also have head movement times, which traditionally have been the 
> major part of time spent in a seek.

For the Seagate drives, I see claims of less than 8.5ms average seek.

> 3600 RPM, by the way... That is the rotational speed of the RM03 (or was 
> it just 3000 RPM?). Also the RP04/05/06 if I remember right. Really old 
> disks... Sounds unlikely that todays small disks wouldn't spin any 
> faster than that.

Not so long ago, though, 3600RPM was usual. As you note, you also
have to include seek time in most cases, so making latency a lot
less than seek time doesn't speed up access. On the other hand,
at a fixed bit rate a faster rotation means less data capacity.

Some years ago, IBM made a mainframe drive similar to a previous
model, but with three times the data capacity, at one third the
rotation rate. The IBM 3390-9, near the end of the large
(I believe 14 inch) disk line, has 10017 cylinders of 15 tracks
each and 56664 bytes/track. (IBM often doesn't make them a nice
binary number.) I believe that is a 4200 RPM drive. Then there is
the 3390-27, which I believe has three logical tracks per
physical track, and runs at 1400RPM.

(snip)

-- glen