The disk rotation speed bottleneck
I’ve been referring to the disk (rotation) speed bottleneck for years, but I don’t really have a clean link for it. Let me fix that right now.
The first hard disks ever were introduced by IBM in 1956. They rotated 1,200 times per minute. Today’s state-of-the-art disk drives rotate 15,000 times per minute. That’s a 12.5-fold improvement since the first term of the Eisenhower Administration. (I understand that the reason for this slow improvement is aerodynamic — a disk that spins too fast literally flies off the spindle.)
Unfortunately, random seek time is bounded below, on average, by 1/2 of a disk’s rotation time. Hence disk seek times can never get below 2 milliseconds.
15,000 RPM = 250 rotations/second, which implies 4 milliseconds/rotation.
From that, much about modern analytic DBMS design follows.
Comments
7 Responses to “The disk rotation speed bottleneck”
Leave a Reply
I don’t get it. We’ve been stuck at 15K RPM for a decade now. I still remember Seagate touting the first 15K RPM Cheetahs in February of ’00. Yeah, those first drives were hot and noisy (and the second generation drives only slightly less so), but they got the job done. By ’03, Seagate and Maxtor (steward of the former Quantum) were producing comparatively muted, much cooler, and (with the arrival of Ultra320 SCSI) even faster…15K drives. And that’s where we’ve been ever since.
Why is that? The escalation from 10K to 15K happened in, what, less than five years?
I thought I remembered hearing tell about 18K drives, but (and perhaps it’s the lateness of the hour) am drawing a complete blank now. I don’t think there are any widely-available/mass-market 18K drives, however.
[…] has at least as high an annual improvement as Moore’s Law shows for chip capacity, the disk rotation speed bottleneck notwithstanding. Disk will long be much cheaper than silicon for data storage. And cheaper silicon […]
@SuperNintendoChalmers, you may be enlightened by
http://www.storagesearch.com/nibble-futureofhdds-servers.html
Robert Cringely is part of a group trying to make drives using stainless steel foil for platters. He was thinking they could run at 30K RPM – twice the speed, half the seek. The price difference with SSD is enough that he expects them to be commercially viable. They will also hold 3X as much data as standard drives due to additional platters.
Original post:
http://www.pbs.org/cringely/pulpit/2006/pulpit_20061026_001143.html
Most recent post:
http://www.cringely.com/2009/06/the-mouse-that-roared/
The theoretical upper end for RPM for a 3.5″ disk platter is roughly 70k. At this rotational speed the outer edge of the disk will be travelling about 330 m/s, just under the speed of sound.
In other words, there is little room for seek time improvement for conventional platter design.
(Don’t take this disk to Denver!)
That last bit sounds like the title of a geeky country song! 🙂
[…] average wait time before you can read data from a specific place on a disk. At 15,000 RPM or less, it can’t be below 2 milliseconds, even if disk heads moved along the radius at infinite speed. In practice, the best figures are […]