Slashdot Mirror
IBM Says Polymer Memory Could Be Ready By 2005
prostoalex writes "Polymer memory is hardly anything new, and we already had HP and Princeton announcing their prototype. In a Forbes magazine article IBM promises polymer memory that's five times cheaper than current flash memory, and expects the first devices with polymer data storage systems to be delivered possibly by 2005. IBM's Zurich Lab published this article last year with description of Millipede."
And does anyone remember what crazy, non-magnetic-plate memory technologies that IBM was saying in 2001 would be ready by 2003?
Just checking.
Irritable, left-wing and possibly humorous bumper stickers and t-shirts
To capture the market, this stuff has to either be:
1. Cheaper than flash or HDs.
2. More durable than flash or HDs (or even CD/DVDs)
3. Be faster than flash/HDs/optical media.
By the time this stuff comes out, trying to beat one of the three is going to be tough - by that time all of those existing technologies will be VERY mature. I'm already able to buy hard drives for super-cheap, so logically, flash is the intended target. The question is, by the time this stuff comes out, will hard drives become so tiny, cheap, and robust, that it's not flash that is the main competitor, but magnetic hard drives?
Of course, if IBM wants to give me petabytes of super-stable long-term storage that will fit in a shoebox, and only cost me a few hundred dollars, who am I to argue? At the very least, if it can replace tape, that might be enough to ensure a place for it, assuming optical hasn't totally displaced that market by then...
I want to see a RAM tech that allows for non-volitle (i.e. keeps its data even without power), and unlimited re-rewrites. This would be a great tech for laptops or PDAs as they could suspend very very easily and boot up to same state. This would be a fabulous tool as battery tech seems to be going nowhere fast.
"I have great faith in fools: Self confidence my friends call it." ~Edgar Allan Poe
This year I had the chance to go to the VLDB (Very Large Databases) conference in Berlin. The keynote speech was about this Millipede project.
I must say everybody in the audience was really impressed: from one side the technological aspects, bordering on nanotechnology, were very interesting. Seeing almost the same principle of vinyl discs miniaturized is really fascinating.
The other really interesting point is the impact that such a storage system will have for our systems.
Imagine, you have 10 Tb of space: what will change in the way you handle data? Probabily the first impact will be the disappearance of the deletion of files: why not keep all the old versions of a file if you have all this space? We could use it as we use packet writing on a CDRW. Or what if your iPod could store some Terabytes of data and restit to a lot more of shock (acceleration)?
The speaker made clear that the storage capacity is huge, but the performances are more or less the same of an HD from today: still the Millipede is highly parallelizable.
I think we must see these new storage technologies not merely as bigger HD, but as something different, with lot of space, but with a bit less of performance.
If you see it from a business perspective, remember that IBM sold its HD division to Hitachi about one year ago: it seems clear that they are going to concentrate themselves on new storage technologies.
Anyway, the future looks really interesting!
Hard drives are great, mature technology, however, they--in my opinion--suck big time. First of all, they have moving parts making them prone to sudden death (thus why RAID exists). Second, they're slow as hell hence why people buy SCSI. But even SCSI isn't fast enough. I mean, nowadays the bottleneck for most computing tasks is the hard drive. Give me DDR RAM fast, solid state long-term storage and I'll be very happy.
As for CD-Rs and DVD-Rs, I burn a lot of them because they're so damn cheap. But I hate it. I once scraped off the surface of a CD-R coaster and almost cried at how easy the stuff flaked off. Not to mention there's no reasonable consensus as to how to properly label the damn things. I mean, you can't write on them, you can't label them...the only thing you can do is take a tiny sharpie and write on the inner circle, which doesn't do me much good. Even though there would be cost and size increases, I would love it if CDs and DVDs had caddies a la Mini Discs.
Yes, I agree these technologies are cheap and mature but I really wish there was some alternative. So, I for one welcome our new micro-millipede masters (terrible name, btw, I have centipedes in my apartment and they freak the hell out of me even though I know they're good to have around 'cause they eat other bugs).
It's rather astonishing isn't it? here we are looking at technology that can increase our storage to petabytes (probably tons more) of storage, and a year or two ago we were worrying about, what was it?, 10, 20, even 30 gigs not being enough? In this world of 300 gig harddrives, and 120 meg floppy disks, rewriteable cds, and such, you have to wonder, when will our advance stop? and how would such a stop in our advance affect us? would we be able to cope with such? but this kind of advance, it is amazing no doubt, i am just in awe at this, and wonder if such storage technologys as in sci fi novels (like the plasma storage in such, (was it asmiov?)) will soon become a reality.
Don't forget -- if we move to this technology world-wide, we'll have mass-storage media that will probably survive an EMP. OK, the actual reader itself will be toast, but the media will survive.
I keep thinking: I want to record something about myself for future generations that will, in one form or another, survive. Right now my best bet for that is printing onto acid-free paper and having it bound, or doing microfiche. This potentially could solve that problem!
"But always she's the spectre of uncertainty I first endured, then faded, then embraced..."
If you don't use video compression, then for today's data transfer speeds, storing and viewing uncompressed video is grossly inefficient--you'll still have to contend with the relatively slow data transfer rate. Assuming a USB 2 connection gives you a 480 Mb/s data rate with the overhead of start and stop bits[1] this gives us 480 Mb/s / (8 data bits + 2 start/stop bits) = 48 MB/s. Suppose a compressed two-hour movie takes up 800 MB of space (divx ;-) can provide a 10:1 compression[2]). This would take 800 Mb / 48 MB/s = 16.6 seconds to transfer[3]. If this movie were originally an 8 GB DVD, the same transfer would take 166.7 s. But this is still with mpeg2 compression. Fully uncompressed, the movie requires *far* more space. According to this link, 4 minutes at the "CCIR-601 digital video standard"[4] would take 4.7 gigabytes. 2 hours into 4 minutes is 120 min / 4 min = 30 [units]; 30 * 4.7 GB = 141.0 GB. Over a USB2 connection, that takes 141.0 GB / 48 MB/s = 2937 s to transfer, or almost 50 minutes. And we're not even considering the data write rate on the device OR the bandwidth load on the device's bus (the CCIR-601 standard can take up between 165.9 and 270 Mbit/s).
:)
;-) Wiki page
Devices always make a practical tradeoff between the bandwidth requirements of the data stream and the computing power required to decode the stream's frames.
Incidentally, I learned a lot about the different compression schemes out there. Thanks for your prompt
- Roey
Notes:
1. I looked up the USB spec but couldn't verify whether it is synchronous or asynchronous (and even then if it uses start and stop bits), so I assumed it uses start and stop bits. It shouldn't make that big a difference anyway.
2. DivX
3. Universal Serial Bus Wiki page
4. CCIR-601 Wiki page