Guest post by “The magnetic hard disk’s tenure as a critical part of the storage technology mosaic is entering its sixth decade, and it shows no sign of ending any time soon. However, certain limitations imposed by rotating media have been coming to the fore lately, and SSDs, which can in theory resolve all these problems, have long been hailed as the eventual successor technology for mass storage. If UK-based startup Dataslide has its way, though, magnetic recording media will get at least one last hurrah, in the form of a new technology called Hard Rectangular Drive.“
Investing in a technology that will be essentially marginalised with a few years. Seems like a sound business strategy.
Top 5 Alternate Meaning for “Hard Rectangular Drive”
5. Part of a standard Driver’s License test
4. If present more than 5 hours, please seek medical attention immediately!
3. Move used frequently by Kobe in the finals
2. Standard feature on every Phantom console
1. A failed technology from early in the 21st century.
Not sure why you think this would be a failure. We know that there are still issues with SSD. According to the specs this drive is faster then SSD, and doesn’t have the problems of SSDs. Sounds pretty innovative to me. If the cost is down, I’d buy a few.
Sorry, I was just having fun. It might be a good technology, but my gut (as with the first comment) says no. But my gut is often wrong!
So we through a platter around, back an forth, to and fro a very small amount. Nice.
We include “millions” of read/write heads. Not so nice. Albeit, using semi-conductor tech will certainly redue the cost, I doubt we will see a price match for current drives any time soon. Then again, that depends on many factors.
500MBps… I’d like that. I’d really really like that.
I could play like 100 movies at once while compiling Haiku instead of only four. Very nice.
–The loon
I just picked up a 1500GB magnetic drive, on a whim, for $139. A quick NewEgg search reveals that the largest SSDs commonly available are 256GB for $600+. Except for 1 500GB SSD for $1600. So I could RAID 6 drives together, for $3600, to equal the capacity of my $139 magnetic drive. Or RAID 3 of the larger models for $4800. So SSD is 25x to 35x more expensive.
And a glance at the reviews reveals that SSD’s have still have problems. As in “start outs fast but then slows down unless you run some clean up utility”. And reliability seems better for magnetic storage, except in the context of mobile devices which can suffer physical shock.
Sure, SSD makes sense for portable devices, including netbooks, and some notebooks. But why should I even consider one for my desktop? Expensive? Relatively small? More trouble?
SSD’s have a geek-sheik aura which distorts our perception of the reality. And the reality is that boring old magnetic storage is still getting bigger and cheaper at an amazing rate. Not so much *faster*, though. But most SSDs are hardly speed demons compared to regular drives.
Edited 2009-06-25 23:00 UTC
SSD are evolving fast, HDD are not evolving anymore. They will reach 3tb some day, but fail to go beyond. If they go smaller they will hit the physical limit of the storing material and drive will just erase themself over time.
SSD software can fix the slowdown problem and current generation SSD are 3 time faster than HDD, so the future -is- bright for SSD and inevitably dark for HDD (over a long period of time)
I wish the future was brighter than a pen light with flat batteries. Right now here is the price for:
Intel Solid State Drive X25-E Extreme 32GB SATA 2.5″ – NZ$947.25
G.SKILL TITAN Series FM-25S2S-256GBT1 2.5″ 256GB SATA II Internal MLC Solid state disk – NZ$1392
At the current pace, I’ll be close to dead by the time it gets within a price range mere mortals can afford. That doesn’t include the problems when it comes to fragmentation, slow downs, crappy controllers, the fact that one has to use a controller adds a host of power consumption to the equation that would exist if there was a way to avoid having to trick the system into thinking it is a disk.
Magnetic storage will continue to hang around for the next 5 years because not only is it GET (Good Enough Technology) but because the flash producers collude to ensure that the prices don’t drop; as soon as the price of flash storage start to drop – all the flash vendors cut back their product to boost up the price. You can’t honestly expect me to believe that a synchronised cut in production is due to coincidence rather than cartel like behaviour.
Edited 2009-06-26 12:14 UTC
Uhhh… right. The “Great Barrier” that they’ll never get past… just like all the previous Great Barriers that they’d never get past.
Magnetic storage is still a fast moving target. When evolution in that arena actually stagnates, as opposed to pundits claiming it will stagnate, let me know and we can revisit the matter.
But even using your 3TB “barrier” (at an assumed $139) for the sake of argument… SSD would have to come down from $10,000 to $139, about a factor of 72, and do it all in just 1 unit, as opposed to RAIDing 6 or 12 units, just to be competitive with magnetic storage.
That is the worst case scenario for magnetic storage that you have laid out with your “barrier”. And even the worst case future looks bright enough that magnetic storage designers have probably gotta wear shades.
Is it just my n00bness exposing itself or isn’t sata3 already faster than 500Mb/s, 6Gbit = 750MB/s…
“sata” is an interface. The fact that sata3 is *faster* than this drives theoretical peak is a good thing. That means the drive can be run to its full potential and not limited by the bus. As it is, the sata3 interface will be saturated by just two of these drives. So you’d definitely want multiple channels… ;D
It is your “n00bness”. 500_MB/s is very much different than 500 Mb/s.
The figures are so out-of-this-world it’s got to be a joke.
Edited 2009-06-26 09:12 UTC
If you can get millions of heads through lithography techniques, you could probably read the whole drive in a very short period of time. This technology has only 64 heads, but that is still plenty to greatly increase speed. Why would it fail? It is supposedly cheaper, faster, and lower power than SSDs.
Edit: I guess they make no claims about price, but I am guessing it would be on par with conventional hard drives and thus would be lower than the price of SSDs. But then, maybe by the time it is viable, racetrack memory will have taken over and SSDs will be cheap.
Edited 2009-06-26 00:56 UTC
This isn’t a new idea, I saw this many years ago applied to atomic force tunneling by IBM rather than magnetic bits and I am sure it has been done with other memory bit technologies too.
In the IBM invention which so far has not gotten anywhere near to market AFAIK, hundreds of tungsten needle sensors vibrate up an down to sense written atoms beneath. The storage bit array moves underneath driven by piezoelectric motors the same as the above. Perhaps it made sense for Dataslide to just substitute the atomic force strategy for known magnetic technology.
Personally I think it could play a role while hard disks ever so slowly get replaced by SSDs in some markets that need high transaction rates.
As for the slide friction, I am not that concerned, you could stop and start it on a dime, something that is much harder to do with disk spin up and down. I figure “spin up” time might be a few shakes. Gee the internal combustion engine rubs metal on metal forever.
I can see some advantages over spinning disks and heads too, like no head crashes that would effectively kill the whole disk. Perhaps some bit tracks might get corrupted on occasion but ECC can handle that. Even if a head dies, the data on surface could likely be recovered by offsetting the card by a track space.