Seagate's new line of hard drives

TalkBack 4 of 7:

Next »

« Previous

• Thread View
• Flat View

Multiple arm strategies

I've seen multiple arm drives. I think that the old PDP drive packs had up to 4 arms. Of course these were big disks. (12"? 16") almost an inch apart, and I think spinning at only 3600 rpm.

The speeds involved are not great. The edge of a 15,000 rpm 3" platter is moving at roughly 150,000 inches per minute. 2,500 inches per second. 200 feet per second. 130 mph. About the speed of a Cessna. While there may be turbulence down stream of an arm, this is way too slow for shock waves. Now there may be problem if vibration from one actuating mechanism is transmitted to the other arm. That's a matter of proper shock absorbers in the actuating mechanism.

It shouldn't be unreasonable to design heads that allow laminar flow around the mechanism. Remember when this was first tried, heads were a lot larger.

I'm assuming that most of the heat is due to the air drag between the disks and the container.
One answer to the heat problem is to evacuate all or most of the air from the drive container. This would require good sealing, and a much sturdier container. However glass picture tubes are large bottles of very good vacuum that maintain their seal for years.

If all the air is evacuated, then a different method of supporting the heads at the right distance from the drive is needed. Perhaps electrets?

If a drive maker can come up with a magnetic bearing, then bearing friction vanishes, especially in an evacuated can.

Some experimental setups for flywheel energy storage used flywheels in evacuated containers using magnetic bearings. I recall truly fabulous rpms -- 200,000?. Dental drills, at much smaller diameters, using air bearings and air turbines routinely hit 100,000 rpm.

The following strategies occur to me:
If you could get 2 arms per drive, even if one set was read-only, then log-structured files systems would be easy to implement. Properly designed, frequently used data ends up being physically close on the drive, so seek times are reduced.

If you can come up with an internal shaft support, it should be possible to put more platters in a can. If you had two stacks of 6 platters separated by a plate and a separate actuator for each group, then you could be doing two operations at once. This in effect is two drives in the same box. Some cost reduction is present because the electronics doesn't have to be duplicated. Heat disipation is a greater problem due to the greater thickness.

With a filtered supply of air, it might be possible to both spin and cool the drive with air. (Mental image of a tiny air compressor chugging away inside your computer case...)

In the longer run, I think tiered access will be the key.
Cache sizes of 5-10% of the drive. Operating systems that track history of disk accesses, and attempt to predict what will be needed next and request the drive to pre-fetch into cache. Very short term backup power supplies that allow the drive to write the cache to disk in a second or two in event of a power failure.

Sherwood Botsford
Resident Sorcerer and System Thaumaturgist

Posted by: sbotsford@...   Posted on: 09/09/07 You are currently: a Guest | Members login | Terms of Use

Alert moderator to an offensive message

Subscribe to this discussion via Email or RSS

SponsoredWhite Papers, Webcasts, and Downloads

• VMware Infrastructure: A Guide to Bottom-Line Benefits VMware Frustrated by the costs of maintain ever larger data centers?or building ... Download Now
• Five Steps to Determine When to Virtualize YourServers VMware Server virtualization isn't just for big companies. Entry-level ... Download Now
• The Impact of Virtualization Software on Operating Environments VMware Today's use of virtualization technology allows IT professionals to ... Download Now