DataSlide reinvents the hard drive

[fellix]

Hard Rectangular Drive (HRD™) is the name of the revealed technology at DataSlide. A weird mixture of magnetic medium and almost static (solid) operation unfold a promise for massive parallel read/write performance of 160 kOP/s and 500 MB/s per device!

More details here.

 

[Simon F]

Hard Rectangular Drive (HRD™) is the name of the revealed technology at DataSlide. A weird mixture of magnetic medium and almost static (solid) operation unfold a promise for massive parallel read/write performance of 160 kOP/s and 500 MB/s per device!

More details here.

So, in a way, it's sort of an unrolled magnetic drum? Nice.

 

[hoom]

Definitely very interesting.
I'm not entirely convinced though, I don't see quite how you get HDD scale capacity with only a piezioelectric actuator?

I guess if they can do tens of GB, that'd make for a good OS/swap drive though.

 

[MfA]

Simon, apart from the repetition of the read/write heads.

This could be a flash killer, depending on how costly the heads are.

 

[MfA]

Definitely very interesting.
I'm not entirely convinced though, I don't see quite how you get HDD scale capacity with only a piezioelectric actuator

By having huge amounts of read/write heads.

Something like this :

.   .   .   .
 .   .   .   .
  .   .   .   .
   .   .   .   .
.   .   .   .
 .   .   .   .

etc.
Because they are staggered/repeated you only ever have to move the medium in 1 dimension to hit every position on the medium. The amount of distance between repeating patterns of the heads determines the longest distance necessary (and the latency).

PS. the staggering of the heads is a bit conjecture on my part, if they can keep the distance between read/write heads horizontally "small enough" they could simply repeat the lines of heads without staggering only keeping distance in between the lines for wiring/switching. That way the medium would only ever have to be moved the distance between two lines. I think the staggering is quite likely though.

 

[fellix]

This thing is like the GPU to the CPU in the computing world.
If the performance numbers are correct (symmetric R/W op's), you will need SATA-III interfacing to gain the potential of a simple RAID array with those.
Looks like this tech has took all the positives of both HDDs and SSDs and virtually none of the negatives.

 

[MfA]

These read/write head arrays are much more complex than say TFT backplanes, they might not be easy/cheap to produce (although on the other hand, dead read/write heads matter much less than dead pixels).

For instance it seems they need vapour deposition of a sub-micron diamond thin film.

 

[Bouncing Zabaglione Bros.]

I'm really interested in some of these new drive technologies. Speeding up the slowest bottleneck in your system is A Good Thing.

However, I'm always disappointed by:

- Price. The price per GB is always way, way above today's hard drives.

- Size. At a time when 1-2 TB drives are becoming the norm, selling us a 128 GB or 256 GB drive isn't cutting it.

So, yes these are many times faster, but at a fraction of the capacity and at many times the price, they just won't go mainstream and bring the unit price down. Until they get them mainstream and the economies of scale working, I don't see these being really getting anywhere except for the few people that can afford to buy a lot of them at massive expense for a RAID array, or the even fewer people that have a specific application for a small, fast drive and are willing to pay significantly more for it.

They really need to make these new drive technologies cheaper and larger capacity - so then I can afford to buy some.

 

[hoom]

By having huge amounts of read/write heads.

The linked article says 64 heads though, which has to be severely restricting capacity.
I guess that could just be for their proof of concept & they plan to scale the number of heads/capacity for production...

 

[MfA]

Just go to the source ...

any 64 of these heads may be active in reads and writes at the same time

The journalist in question wasn't thinking straight.

PS. some of the people in the comment section seem quite well informed though.

 

[Simon F]

The linked article says 64 heads though, which has to be severely restricting capacity.

No, AFAIU, it was using only 64 heads at the same time

 

[hoom]

OK, yeah if they aren't arbitrarily limiting/ed to such a relatively small number of heads thats fine as long as its not costing too much per head.