Why do HDDs suffer from performance loss due to capacity decrease?

15

4

Today I was looking over my HDD datasheet(a Western Digital 500GB WD5000BPKT) and I noticed that for its entire family, the "Performance" row "Host to/from drive (sustained)" is nominally decreasing in read and writing performance due to a capacity decrease:

enter image description here

This makes me wonder why. I didn't notice any other parameter that differs on this table except for the device capacity (750GB to 160GB). I also looked for other families and I found the same behavior (Samsung and Seagate don't show this performance parameters on their datasheets).

So, considering that all disks belong to the same family and are in essence the same, does someone know the exact physical reason for this behavior?

Diogo

Posted 2012-06-25T14:13:59.810

Reputation: 28 202

Answers

17

The storage density of the bigger hard drives is higher than the densities of the smaller ones. With the same rotational speed (7200 RPM), this means that data can be read/written faster.

From Memory storage density # Effects on performance - Wikipedia, the free encyclopedia:

Increasing storage density of a medium almost always improves the transfer speed at which that medium can operate. This is most obvious when considering various disk-based media, where the storage elements are spread over the surface of the disk and must be physically rotated under the "head" in order to be read or written. Higher density means more data moves under the head for any given mechanical movement.

Considering the floppy disk as a basic example, we can calculate the effective transfer speed by determining how fast the bits move under the head. A standard 3½" floppy disk spins at 300 rpm, and the innermost track about 66 mm long (10.5 mm radius). At 300 rpm the linear speed of the media under the head is thus about 66 mm x 300 rpm = 19800 mm/minute, or 330 mm/s. Along that track the bits are stored at a density of 686 bit/mm, which means that the head sees 686 bit/mm x 330 mm/s = 226,380 bit/s (or 28.3 KiB/s).

Now consider an improvement to the design that doubles the density of the bits by reducing sample length and keeping the same track spacing. This would immediately result in a doubling of transfer speed because the bits would be passing under the head twice as fast. Early floppy disk interfaces were originally designed with 250 kbit/s transfer speeds in mind, and were already being outperformed with the introduction of the "high density" 1.44 MB (1,440 KiB) floppies in the 1980s. The vast majority of PCs included interfaces designed for high density drives that ran at 500 kbit/s instead. These too were completely overwhelmed by newer devices like the LS-120, which were forced to use higher-speed interfaces such as IDE.

(emphasis mine)

Dennis

Posted 2012-06-25T14:13:59.810

Reputation: 42 934

Be careful with the comparison to a floppy. The floppy disk (and ST-506) interface has raw data to/from the controller on the host PC. IDE and SATA drives have on-board disk controllers that fully buffers the data after the read head. The PATA/SATA transfer rates to/from the host PC are independent and totally separate from the R/W head operations. Those "sustained" transfer rates in the chart are averages that include idle or dead time on the interface (while the sectors are accessed/read) as well as the full 3 Gb/sec data transfers. – sawdust – 2012-06-25T20:21:38.780

@sawdust: The floppy analogy is merely to illustrate how higher density may result in higher speeds. Saying that the host-to-drive transfer rate independent and totally separate from the media rate is a bit far fetched. The sustained maximum transfer rate is averaged over a long period of time and usually specified for sequential reads. This keeps head/cylinder switches to a minimum, so the media rate is the decisive factor. – Dennis – 2012-06-26T01:39:25.347

OK, that wasn't worded clearly. Since the sector-transfer times of the read-head-to-sector-buffer and sector-buffer-to-host are summed, both are mathematically correlated to the average transfer rate. But these two operations occur sequentially, not concurrently. The sector data is read into the buffer (at a rate based on the platter's bit density). It is then validated. Error correction is applied if necessary, or maybe the sector has to be re-read. Only after verification is the sector data transferred to the host (at the rate of the interface, in this case SATA II 3Gb/sec). – sawdust – 2012-06-26T05:23:05.273

4

Well, all drives are supposedly the same size (as in height, width, depth). So, to fit more data into that size, the data has to be packed at higher density.

The moving parts (like the head) in the HDDs most likely all move at the same speed.

So, if you increase the data density, but the speed at which you move over said data stays constant, you increase overall throughput.

Der Hochstapler

Posted 2012-06-25T14:13:59.810

Reputation: 77 228

1

Yes, but - a typical drive may have one, two, or three platters. The total capacity isn't an absolute indicator of data density. Here's a post (http://rml527.blogspot.com/2010/10/hdd-platter-database-western-digital-25.html) that lists Western Digital 2.5 inch drives with densities of 160 GB, 250 GB, 320 GB, and 500 GB per platter.

– Dave Becker – 2012-06-25T14:32:40.210

3

I believe Advanced Format refers to the use of 4k sectors instead of 512bytes. Among other things, this change meant that fewer bits on the platter needed to be used for ECC codes. As a result slightly fewer bits need to be read to get a given amount of data off the drive; with everything else the same this would result in a slightly higher maximum transfer rate. This is likely the explanation for the difference between the two 500MB drives.

Dan is Fiddling by Firelight

Posted 2012-06-25T14:13:59.810

Reputation: 2 677

1

Well, just a wild guess, but:

A harddrive is divided in several tracks, each one divided into several equally big blocks.

When reading data a harddrive first moves its head to the right track and then waits until the disk has rotated to the right block. For a consecutive read of larger files spanning over several blocks and tracks this movement has to happen quite often. (even more with higher fragmentation)

The bigger drives have either a higher rate of data stored on each track or contain an additional disk. This way the head isn't forced to move this often which effectively boosts the transfer rate.

(read about access times on wikipedia)

Baarn

Posted 2012-06-25T14:13:59.810

Reputation: 6 096

1

In addition to higher bit densities, another possible answer is that the larger HDs have MORE platters / disks. With more platters you see more bits at the same time without having to move the read heads. Also, some drives go to 2 sided platters with same effect

Ron

Posted 2012-06-25T14:13:59.810

Reputation: 11