Super User is a question and answer site for computer enthusiasts and power users. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Suppose we have a harddisk with 2 platters with characteristics below:

Rotational rate: 10, 000 RPM
Avg sectors/track: 1000
Surfaces: 4
Sector size: 512 bytes

I was reading "Computer Systems: A Programmer's Perspective 2ed" when I found that it calculates transfer time as if it only uses ONE head to read a sector.
If that's the case, why not use 4 heads to write(read) on 4 surfaces? So when I write a 2K bytes file, each head should only need to wait for the platters to rotate just one sector length instead of 4, thus reducing the transfer time by a factor of 4.
Or even redesign sector to make each sector on one cylinder but on 4 tracks residing same position respectively on 4 surfaces. Each one of (512/4) bytes. So when the hd needs to read a sector of 512 bytes, we only need the disk to rotate roughly 1/4 compare to original time.
The idea looks like RAID 0.

share|improve this question
I don't know what is actually done, but I can say that something that is usually wanted is file contiguity. That can't happen if parts of the files are written on different platters. – soandos Jun 15 '12 at 3:04
In fact, that doesn't have anything to do with file contiguity. It's only a way to organize the internal mechanism when implementing a sector(the last part of the question). In this scenario, all data in a "logical sector" is on one cylinder, and the actual data splits in 4 parts landing on the same physical position on different surface. This is within a sector, when fragmentation is about sectors placement of a file. – WiSaGaN Jun 15 '12 at 3:10
up vote 9 down vote accepted

To read or write all heads at once would require duplicating all the circuitry involved - head amplifiers, bitstream decoders, even the microprocessors and cache memory. I don't know how disks are actually designed, but it seems unlikely to me, especially given the price pressures in the industry. Besides, even the stream from a single platter is likely to be near the interface limit - it's the seeking and rotational delays that kill you.

share|improve this answer

I think that's already how it works :


A cylinder comprises the same track number on each platter, spanning all such tracks across each platter surface that is able to store data (without regard to whether or not the track is "bad"). Thus, it is a three-dimensional structure. Any track comprising part of a specific cylinder can be written to and read from while the actuator assembly remains stationary, and one way in which hard drive manufacturers have increased drive access speed has been by increasing the number of platters which can be read at the same time.


this quote from Wikipedia don't say it all, the guys from clearly say :

Only one head can read from or write to the hard disk at a given time. Special circuitry is used to control which head is active at any given time.

so this is for the low/middle-end HDD, BUT I came across few OLD discussions here and here where they are referring to the technique you are describing as "internal RAID" and I actually don't know enough about today's HDD technology.

share|improve this answer
Interesting quote, where's it from? – Mark Ransom Jun 15 '12 at 23:01
It's from wikipedia, with no reference though. – WiSaGaN Jun 16 '12 at 4:32
From your link, John Rose said "Maybe it's because a certain number of hard drive sectors are always marked as "bad", since no drive will ever have all their billions of sectors functioning perfectly. If you did the "internal RAID" thing, if one sector failed, the drive would need to mark it as bad on all 2 or 3 platters.which I guess would really start cutting down on usable drive capacity...". I think it makes sense too. – WiSaGaN Jun 18 '12 at 2:10

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .