Striping data across platters in a single hard disk

  • Thread starter Thread starter tony
  • Start date Start date
Thanks everybody (including those on the other fork!) for clarifying
why it wouldn't work!
 
Nope. Thermal effects are _much_ larger than many people
imagine. 3.5" aluminum platters grow 44 um from 20-40'C.
That's about 440 tracks!

If you want to scare a mechanical engineer sometime, tell
her to multiply Young's Modulus of Elasticity by the linear
thermal expansion coefficient. The resulting stress per
degree _is_ scary: Many materials will yield below 200'C
temperature difference.


A lifetime ago I attended a lecture by one of IBM's electromechanical
engineers who had been part of the 3330-generation of disk drives.
(10 plattters, 14 inches each. Originally 100MB and maybe 35ms seek
time. Maybe the first servo-based disk drive.)

He said the servo was designed to follow a half-inch out-of-round
track at 3600 RPM. It took huge coils and capacitors to do this.
 
Or more basically, 'just because head0 is above Track0 on platter0, you
cannot assume that head1 is above Track0 on plattert1' .. maybe it was
when you wrote, but it might not be when you read some time later.

Hmm, you sure about that? That must depend on the timeframe and mfr who
made the multi-head drive - go back far enough and there were certainly
some which had the heads in "lock-step" and which wrote/read the data
to/off all the platter sides in parallel.
It might be slightly less risky having two 'side by side' heads on the
same platter (remember 'drums', anyone? (Anyone really ancient I mean)),
but even that probably won't work, given thermal effects.

Oh I remember a story about a FastRand drum which supposedly came off its
bearings and chased the operator around the computer room.:-)
 
Or if it made doing the parallel head thing feasible (lower density makes the
positioning thing easier?). Trade some capacity for higher throughput.

I think it'd be a *lot* of capcity from what I'm hearing from others.
It would take a lot of drives though to saturate a SATA II bus.

I was at the Samsung site a few days ago and noticed that they are
working on hybrid drives (flash mem + rotating media), but the mem
part of it is only 1 GB so it seems more like a cache of sorts rather
than general storage. I can't imagine that the hybrid drives will be
cheap. If a single "striped" drive could be had for the price of a hybrid,
that may be the acceptable price point to shoot for. Maybe getting multiple
platter drives to produce higher throughput is wasted effort given the following
technology:

I guess they have to have a market for all those flash chips they're
planning on producing... apart from Apple.:-)
I read something about perpendicular magnetic platters. How about
getting that to work in some kind of "parallel" fashion to increase throughput?

The only thing I've heard about "vertical" wrt disks is "vertical
recording" which has to do with the alignment of the magnetic particles on
the disk: instead of all lying side by side on the platter, they get stood
up on end - IOW it's about increased density. This has been talked about
for 20 years now and it's suddenly just starting to get implemented.
 
Bitstring <[email protected]>, from the
wonderful person George Macdonald <fammacd=!SPAM^[email protected]>
said
Hmm, you sure about that? That must depend on the timeframe and mfr who
made the multi-head drive - go back far enough and there were certainly
some which had the heads in "lock-step" and which wrote/read the data
to/off all the platter sides in parallel.

Yes, if you go back far enough it was true. Back when disc capacities
were measured in MB .. maybe even single digit GB .. and one whole (side
of a) platter was used for track alignment data.
Oh I remember a story about a FastRand drum which supposedly came off its
bearings and chased the operator around the computer room.:-)

Never seen that. I've seen the hole in the wall caused by a 30"
(diameter) disc stack tearing loose. Luckily no operators were in the
way. In those days you could repair a disc (but not that one!) with a
large spanner and a spare platter, and you didn't even need (much of) a
clean room. The old platters made neat coffee tables too although the
magnetic layer tended to scrape off.
 
George Macdonald said:
I think it'd be a *lot* of capcity from what I'm hearing from others.


I guess they have to have a market for all those flash chips they're
planning on producing... apart from Apple.:-)

Here's a link the Samsung/Microsoft Hybrid HDD release from 4/25/05:

http://www.samsung.com/Products/HardDiskDrive/news/HardDiskDrive_20050425_0000117556.htm
The only thing I've heard about "vertical" wrt disks is "vertical
recording" which has to do with the alignment of the magnetic particles on
the disk: instead of all lying side by side on the platter, they get stood
up on end - IOW it's about increased density. This has been talked about
for 20 years now and it's suddenly just starting to get implemented.

I thought though they could read/write data in layers magnetically similar to
multi-layer optical technology (dual layer DVDs). Here's a link (aug/05):

http://www.samsung.com/Products/HardDiskDrive/news/HardDiskDrive_20050801_0000197917.htm

Tony
 
Here's a link the Samsung/Microsoft Hybrid HDD release from 4/25/05:

http://www.samsung.com/Products/HardDiskDrive/news/HardDiskDrive_20050425_0000117556.htm

This seems to be about fast startup and doing trivial things, like watching
a movie on the plane, without a spinning HDD. Obviously the HDD mfrs would
rather be the ones adding value and getting the return but I get the
feeling that we'd be better off if the flash used for such purposes is on
the mbrd.
I thought though they could read/write data in layers magnetically similar to
multi-layer optical technology (dual layer DVDs). Here's a link (aug/05):

http://www.samsung.com/Products/HardDiskDrive/news/HardDiskDrive_20050801_0000197917.htm

It's a bit muddled but I don't think they're talking about multi-layer
recording here.
 
George Macdonald said:
This seems to be about fast startup and doing trivial things, like watching
a movie on the plane, without a spinning HDD. Obviously the HDD mfrs would
rather be the ones adding value and getting the return but I get the
feeling that we'd be better off if the flash used for such purposes is on
the mbrd.

I was thinking the exact same thing today. The question becomes: "does
flash need a different interface than drives/peripherals (it's memory afterall)?".
It's a bit muddled but I don't think they're talking about multi-layer
recording here.

They are though because they indicate a separation layer between 2 magnetic
layers.

Tony
 
craigm said:
No, any change in temperature causes them to change position relative to each other. Twenty years ago they would have been
close enough, but today track densities are much higher. Then drives had about 600 tracks per inch, today you may see
60000+ tracks per inch.


But the problem is that your assumption that they would be in the same relative position doesn't work.


That is also 20 year old technology. Today all hard drives use embedded servo where servo and data reside on the same
track. This was necessary to deal with the temperature shift of the heads. Also with few disks, this is a much more
efficient use of the disk surface.




To accommodate your concept each head would have to be positioned independently. This would vastly add to the cost of a
drive.

You would also need to replicate the entire read/write electronics chain for each head. Again, a significant increase in
cost.

Now, when you add in defect management, you can run into a situation where you have to map out the same areas on each
surface to keep the data synchronized across heads. This would reduce the capacity of the drive.

There are other technical issues involved that I won't bother to address.

The bottom line is: If you could make it work, the drive would be too expensive for anyone in their right mind to buy. For
those needing high performance, look for a high performance RAID system. (I'm not talking two drive raid 1 inside a PC
either.)

How including multiple sets of heads and electronics into one drive and
funnelling the results into one data interface would be more expensive
than multiple drives illudes me. Doing that sounds conceptually "easy" to
me. How to store the data and combine the data under various read/write
scenarios seems like the tricky part of the project. This hardware design
would work on a single platter drive also of course.

Tony
 
tony said:
How including multiple sets of heads and electronics into one drive and
funnelling the results into one data interface would be more expensive
than multiple drives illudes me. Doing that sounds conceptually "easy" to
me. How to store the data and combine the data under various read/write
scenarios seems like the tricky part of the project. This hardware design
would work on a single platter drive also of course.

Tony

Simple. There are millions of single channel drives made every year.
They are cheap.

_If_ you could make what you want work, the demand would be much lower.
Volume cost considerations would not apply. Additionally,the actuator
you are talking about itself could be incredibly expensive. You would
probably not fit in a normal 3.5 inch drive space.



craigm
 
craigm said:
Simple. There are millions of single channel drives made every year. They are cheap.

So adding another set of heads/actuators would make drives that much more
expensive? Is that the most expensive portion of the drive? Of course the
price would be higher, how much is the question (assuming the electronics
and firmware could be changed to accomodate the addition).
_If_ you could make what you want work, the demand would be much lower.

I'd buy one. :)
Volume cost considerations would not apply.

Oh I bet they would if the throughput could be almost doubled.
Additionally,the actuator you are talking about itself could be incredibly expensive.

The same as the one already in there.
You would probably not fit in a normal 3.5 inch drive space.

An additional set of heads/actuator you mean? (Of course _I_ wouldn't fit! ;) ).

Tony
 
tony said:
So adding another set of heads/actuators would make drives that much more
expensive? Is that the most expensive portion of the drive? Of course the
price would be higher, how much is the question (assuming the electronics
and firmware could be changed to accomodate the addition).




I'd buy one. :)




Oh I bet they would if the throughput could be almost doubled.




The same as the one already in there.




An additional set of heads/actuator you mean? (Of course _I_ wouldn't fit! ;) ).

Tony


I can only point out the issues. If you think you can do it, go for it.
 
tony said:
So adding another set of heads/actuators would make drives that much more
expensive? Is that the most expensive portion of the drive? Of course the
price would be higher, how much is the question (assuming the electronics
and firmware could be changed to accomodate the addition).

Essentially, dual-actuator 200 GB drive BOM and assembly cost is almost as
much as two 200 GB drives BOM (minus platters and motor).The price would be
much greater because of small production volume. Then the system builders
would rather have two actual 200 GB drives for the same cost.
The drive also is not likely to fit into the standard form factor.
 
Essentially, dual-actuator 200 GB drive BOM and assembly cost is almost as
much as two 200 GB drives BOM (minus platters and motor).The price would be
much greater because of small production volume. Then the system builders
would rather have two actual 200 GB drives for the same cost.
The drive also is not likely to fit into the standard form factor.


And the MTBF would be higher and the component that has to be replaced
to fix is more expensive.
 
tony said:
So adding another set of heads/actuators would make drives that much more
expensive? Is that the most expensive portion of the drive? Of course the
price would be higher, how much is the question (assuming the electronics
and firmware could be changed to accomodate the addition).


I'd buy one. :)

And look at all the idiots who spend 6 times as much, per GB, for
a 74 GB Raptor compared to a more typical 300 GB 7200 rpm SATA
drive. If some people are willing to spend 6x as much for a very
small performance benefit, imagine what they'd be willing to do
for something that might offer a more significant performance
benefit.

With the new 150 GB Raptors, the Raptor per-GB premium drops to
about 5x.

That 5x or 6x factor gives us an idea of what an /existing/
subset of the hard drive market is willing to pay for a small
performance gain - and that sub-market is obviously large enough
to satisfy Western Digital or they would have stopped making
Raptors long ago.
 
craigm said:
I can only point out the issues. If you think you can do it, go for it.

A number of these combinations have been implemented in IBM drives at one
time or another, from two actuators to two sets of heads for inner and
outer zones on one actuator, to a surface of head/track for fast swap in
addition to normal actuators. But so far as I know, none of them are
used any more. In fact the market for specialized HDA (head disk
assembly) has pretty much gone away.

del cecchi
 
Alexander said:
Essentially, dual-actuator 200 GB drive BOM and assembly cost is almost as
much as two 200 GB drives BOM (minus platters and motor).The price would be
much greater because of small production volume. Then the system builders
would rather have two actual 200 GB drives for the same cost.
The drive also is not likely to fit into the standard form factor.

The hypothetical striped-platters drive doesn't have to be cheap
enough to compete with ordinary SATA or PATA drives. It merely
has to be cheap enough to appeal to a sufficiently large
sub-market, such as the Raptor market that has already
demonstrated that it is willing to pay huge price premiums for
small performance gains. This hard drive sub-market is obviously
big enough to make it worthwhile for WD or WD would have stopped
making Raptors long ago.


Compare a 74 GB Raptor to a typical 200 GB SATA drive: 2.25
times the price to get 37% of the capacity - a 6x per GB premium.
And I would guesstimate that 80% of the Raptor owners I know
bought pairs of Raptors - not just one drive.

Even the new 150 GB, $300 US, Raptors are back-ordered for 6 to 8
weeks by people who can't wait to pay huge price premiums for
small performance gains. And all three people I know of so far
who have ordered these drives ordered pairs. $300 US can buy you
three 200 GB SATA drives, but there are still enough people out
there who would rather trade a lot of capacity for a little bit
of speed.


And who cares if the hypothetical stripped-platters drive
wouldn't fix in a standard 3.5" drive bays - the kinds of people
who buy Raptors are not going to quibble over a minor detail like
that. This issue is trivial to work around and they'll do it
happily in order to have their computer boot in 45 seconds
instead of 47.
 
Alexander Grigoriev said:
Essentially, dual-actuator 200 GB drive BOM and assembly cost is almost as much as two 200 GB drives BOM (minus platters
and motor).The price would be much greater because of small production volume. Then the system builders would rather have
two actual 200 GB drives for the same cost.
The drive also is not likely to fit into the standard form factor.

I'm not convinced at all that a drive "twice as fast but costs more" would be
"low volume".

Tony
 
Rob Stow said:
And look at all the idiots who spend 6 times as much, per GB, for a 74 GB Raptor compared to a more typical 300 GB 7200 rpm
SATA drive. If some people are willing to spend 6x as much for a very small performance benefit, imagine what they'd be
willing to do for something that might offer a more significant performance benefit.

With the new 150 GB Raptors, the Raptor per-GB premium drops to about 5x.

That 5x or 6x factor gives us an idea of what an /existing/ subset of the hard drive market is willing to pay for a small
performance gain - and that sub-market is obviously large enough to satisfy Western Digital or they would have stopped
making Raptors long ago.

And we're talking about a large potential performance gain with the
dual (or 3! ;) ) head/actuator setup too.

Tony
 
Back
Top