And the read speed?
Does that get a boost as well or are we looking at drives that write 100 times faster than they read?
A team of scientists have published a new way of using heat to store data magnetically, which could increase the speed of hard drives over a hundredfold. Conventional drives use electromagnetism to selectively change the polarity of points on a drive, representing a one or a zero. But according to research published in Nature …
yet another pie-in-the-sky 1000-times-more-this-n-that storage solution that will never materialise, while we're STILL stuck using the same old magnetic-spinny-disk tech we were using 25 years ago. The only alternative now is SSDs, which may be heaps faster but die after 6 months of use.
This? I'll believe it when I see it at my local IT store, like all^H^H^Hnone of the other oh-so-amazing storage inventions that never made it there.
Oh noes! We're still using a technology that gets cheaper and faster and more achieves a higher storage density every year! Whatever is the world coming to? In other news, I drive a diesel. Did you know that the first diesel was demonstrated nearly 120 years ago? And yet the same principles are still perfectly sound, and it remains quite economical and efficient.
New tech will roll in when the old is no longer up to the task. No point bringing it in before then, is there?
Just exmaple : OCZ VERTEX 3 SATA III 2.5" SSD 240GB life span is 2 Milion hours.
2 000 000h / 24h = 83333 days / 365days = 228 YEARS
With a normal filled SSD and very heavy writing/usage of say 10GB data each day 365 days a year you'd be looking at roughly 22 full SSD write cycles per year, out of the 3000 (worst case scenario) available.
Worst scenario then : 3000 / 22per year = 136 Years !
So in the worst scennario if TheRegister keep you message at SSD it will be still to read for quite long time. SSD drive will fail because spark-sure or tunder bolt much easier then because of short life span.
So take care about you power supply ;)
"OCZ VERTEX 3 SATA III 2.5" SSD 240GB life span is 2 Milion hours."
Mean-Time Between Failure is not "Life Span." Here's an example from a reference:
Example: "MTBF of 100,000 hours. [This figure] indicate[s] that in a population of 1,000,000 batteries, there will be approximately ten battery failures every hour during a single battery's four-hour life span. This is because during that one hour, the 1,000,000 batteries were operational for a sum of 1,000,000 hours; dividing total uptime (1,000,000) with MTBF (100,000) we get the average number of failures."
So, one SSD in a group of 2mil dying every hour, while good odds, doesn't ensure your drive won't pop its clogs in the next 2 hours. Next in line is that mainstream mechanical drives barely crest the 1mil MTBF mark...
As for write-endurance, I've heard of numerous drives having hardware/firmware deaths, but not many (if any) write-endurance problems.
I would imagine one of the main technological challenges will be seek time. Moving the mass of a laser in 10ms is not as easy as you might think. Look at the size of the sledges in an optical drive.
Still even with this limitation it could find good uses in systems where huge writes are performed and theres not such a need for random access.
Oh, and do repeated write (==heat) cycles cause a degradation in the material?
You wouldn't have to shift the whole laser unit across the platter, only the focus lens and the micro-mirrors for transferring the bean along the arm. Most laser cutters us this arrangement, as it means the expensive part doesn't suffer from movement induced problems.
Even though the write speed is one hell of a lot faster than the read speeds currently available, I'm thinking that any advancement to the old tech of HDDs is going to be a good thing for some time to come. SSDs might be the flashy new medium for high-speed access currently, but we'll still need cheaper and larger storage for the rest of our medium-low use data. Anything that can equal the gap between the two, however little, is welcome.
It's great writing at those speeds and I see lots of applications for that, in my realm - transaction log of a database for instance; web logs etc..
However, what about a semi or full random workload, will there be latency like we have already on spinning media, has anybody read the paper yet http://www.nature.com/ncomms/journal/v3/n2/full/ncomms1666.html?
It will be great if they can make the technology so we don't have spinning platters and head movements, basically make like flash :)
So it doesn't really say much about how this works (and I don't have $32 just for my curiosity). Does it write on hundreds of parallel tracks simultaneously to get the 100s times increase in write speed, or does the disk spin at 1500000 r.p.m. ? (If it is the latter, how many people does the disk fly through after a containment failure?)
How does it read? Is it still done using GMR heads, or detecting polarisation change of the laser?
"how many people does the disk fly through after a containment failure?"
Cue obligatory random sci-fi quote:
"I come in peace"
"Yeah, and you'll go in pieces, asshole*"
*It was an American film. After all, Arnie never said "Fuck you, arsehole" - although it would have been quite funny if he did.
It's all well and good, nice bit of research etc. But do we really want anything spinning in a laptop, or even a desktop or anything else these days due to the inherent mechanical failures? I don't. There is a chance for high end server storage but the speeds the issue with that. Nice theory though.
Biting the hand that feeds IT © 1998–2021