Slashdot Mirror
IBM Prepares 100-Terabyte Tape Drives
Roland Piquepaille writes "It's a well-known fact that we're living in an era of data explosion, and that it's not about to stop. So it's not really surprising that IBM researchers are eyeing 100T-byte tape drive. Yes, you read correctly. They want to increase the capacity storage of their largest units by 250 times, from 400 GB to 100 TB. In order to achieve this goal, they're borrowing "nanopatterning" techniques derived from the microprocessor division. Today, the size of a tape track is about 10 microns. They want to reduce it to 0.5 micron -- or 500 nanometers -- in about five years. IBM doesn't really say when a 100-Terabyte tape drive will be available. But more importantly, the company doesn't say a word about future data transfer rates, which today reach a 80 MB/s. Read this overview for more comments about this problem of data transfer rates."
That's _alot_ of pr0n.
...I mean, I'm sure I could back up my entire life to one of these things... ;)
;)
Seriously, imagine backing up every single thing you've ever heard, seen, or read. 40TB maybe?
Error 407 - No creative sig found
What do they have in mind, they want to build the world's largest Turing machine?
Laser holography ?
If anyone can do it, it's big blue. I remember when they first USB key drives were widely available from IBM. 8mb and 16mb form factors. I talked to a guy I knew at IBM, who smiled at me and told me to expect 1 GB form factors within a year. (Not that I would be able to afford them. :-) )
The boys at IBM can create anything they put their minds to. Marketing is another matter.
Help! I'm a slashdot refugee.
you know, he could at least suggest in his submittion to "Read my overview.."
"Never under estimated the bandwith of a station wagon full of tapes..." http://www.bbc.co.uk/dna/h2g2/A678576
Pshaaah... I say.
Never underestimate the bandwidth of a 100GB tape in my pocket.
ps: We do all realize that, if these ever make it to the consumer market, approximately 99.9% of these will be bought for the "backup" of copyrighted material, right... ?
... a beowulf cluster of these !
At least with some striping, they can resolve the speed issue.
(yay I know it's the 10^14th time we mention beowulf)
I don't know where the solution here will come from, but I expect for the meantime this kind of large capacity will be used more for archival storage of old data than for backup.
Is there any research out there into the data transfer rate problem?
How am I supposed to fit a pithy, relevant quote into 120 characters?
Heck, even if you were to connect a drive like that to the firewire-800 port (800mbps) port on a new (today) PowerMac G5, it'd take over 277 HOURS to fill that tape! (assuming complete bus saturation and an 8-bit-byte)
What is the difference between a small revolutionary change and a large evolutionary change?
Roland Piquepaille and Slashdot: Is there a connection?
I think most of you are aware of the controversy surrounding regular Slashdot article submitter Roland Piquepaille. For those of you who don't know, please allow me to bring forth all the facts. Roland Piquepaille has an online journal (I refuse to use the word "blog") located at www.primidi.com. It is titled "Roland Piquepaille's Technology Trends". It consists almost entirely of content, both text and pictures, taken from reputable news websites and online technical journals. He does give credit to the other websites, but it wasn't always so. Only after many complaints were raised by the Slashdot readership did he start giving credit where credit was due. However, this is not what the controversy is about.
Roland Piquepaille's Technology Trends serves online advertisements through a service called Blogads, located at www.blogads.com. Blogads is not your traditional online advertiser; rather than base payments on click-throughs, Blogads pays a flat fee based on the level of traffic your online journal generates. This way Blogads can guarantee that an advertisement on a particular online journal will reach a particular number of users. So advertisements on high traffic online journals are appropriately more expensive to buy, but the advertisement is guaranteed to be seen by a large amount of people. This, in turn, encourages people like Roland Piquepaille to try their best to increase traffic to their journals in order to increase the going rates for advertisements on their web pages. But advertisers do have some flexibility. Blogads serves two classes of advertisements. The premium ad space that is seen at the top of the web page by all viewers is reserved for "Special Advertisers"; it holds only one advertisement. The secondary ad space is located near the bottom half of the page, so that the user must scroll down the window to see it. This space can contain up to four advertisements and is reserved for regular advertisers, or just "Advertisers". Visit Roland Piquepaille's Technology Trends (www.primidi.com) to see it for yourself.
Before we talk about money, let's talk about the service that Roland Piquepaille provides in his journal. He goes out and looks for interesting articles about new and emerging technologies. He provides a very brief overview of the articles, then copies a few choice paragraphs and the occasional picture from each article and puts them up on his web page. Finally, he adds a minimal amount of original content between the copied-and-pasted text in an effort to make the journal entry coherent and appear to add value to the original articles. Nothing more, nothing less.
Now let's talk about money. Visit http://www.blogads.com/order_html?adstrip_category =tech&politics= to check the following facts for yourself. As of today, December XX 2004, the going rate for the premium advertisement space on Roland Piquepaille's Technology Trends is $375 for one month. One of the four standard advertisements costs $150 for one month. So, the maximum advertising space brings in $375 x 1 + $150 x 4 = $975 for one month. Obviously not all $975 will go directly to Roland Piquepaille, as Blogads gets a portion of that as a service fee, but he will receive the majority of it. According to the FAQ, Blogads takes 20%. So Roland Piquepaille gets 80% of $975, a maximum of $780 each month. www.primidi.com is hosted by clara.net (look it up at http://www.networksolutions.com/en_US/whois/index. jhtml [networksolutions.com]). Browsing clara.net's hosting solutions, the most expensive hosting service is their Clarahost Advanced (http://
Come on, sure its getting IBM exposure, but it does qualify as news. 100TB's is a big leap up, and it's something I'm interested in reading about, even if a company stands to gain from it.
WANNAWIKI Wannawiki WannaWiki WANNAWIKI!
yes, this will be perfect for my....... "archiving" needs.
People are starting to think that having RAID is actually the same thing as backing up your data.
The owls are not what they seem
"Anthropo-scientific research material".
You can do this today. Just make it hold a much bigger reel.
Unknown host pong.
I misunderstood... completely, ignore me =)
WANNAWIKI Wannawiki WannaWiki WANNAWIKI!
The IBM enterprise SAN device -- shark -- is only able to crank data out at about 35MB/s per disk pack, assuming sole access. When you've got multiple systems hitting one disk pack it drops dramatically from that.
So 80MB/s is more than their disk systems can do anyway, unless you're pulling data from multiple packs.
"But actually trying to use m4 as a general-purpose langage would be deeply perverse" --ESR
.05 micron data tracks don't seem very robust to me.
The only thing worse than a catastrophic data loss would be to find out that your backups are unreadable.
The VXA drives from Ecrix seemed to go a long way in terms of increasing reliability of "concentional" tape drives, I'll wait for something similar for this new technology.
LK
"Hi. This is my friend, Jack Shit, and you don't know him." - Lord Kano
99.9% of these will be used by corporation and governmant agencies.
The Kruger Dunning explains most post on
256 times larger :\
Now that's an awesome possibility. It's about time storage-related companies start worrying about reliability and longevity of their storage solutions instead of trying to impress everyone with capacity.
100 TBs is a lot of p0rn.
The next version of Microsoft Office fits on two of these.
If we have one micron now, and 0.5 micron allows for 100T byte tapes, why don't we have 50T byte now?
my karma will be here long after I'm gone
IBM has a subsidiary in China and cannot possibly vet all the Chinese working there. As well, IBM has an affiliate in Taiwan. Of the two groups (i.e. mainland Chinese and Taiwanese), the Taiwanese pose the greater security risk.
The technologies that IBM has developed could be used by the Chinese military. Some Taiwanese, at the request of Beijing, might just xerox the blueprints for this technology and hand it to Chinese intelligence services. One use of high density storage is to store all the intelligence collected by Taiwanese spies on behalf of Beijing.
To IBM, I say, "Tread carefully. Guard your technology. The Chinese wall between IBM-USA and IBM-Taiwan is no protection. Tear it down and sell the Taiwan subsidiary to Acer."
"Nope, it's somewhere here on tape..."
Comment removed based on user account deletion
Comment removed based on user account deletion
Comment removed based on user account deletion
Literally. You can wait your whole life to backup your data, and wait another life time to restore them.
The simple answer is that the data is not stored and to end - - - - - in a linear array. The data is stored in a 2 dimensional grid, using both the width of the tape as well as the linear position on the reel to store data. In many cases, the data is stored in a biased array of rows /////
Thus the data density depends on not just the size of the individual particles storing the bits, but also on the possible arrangement of the grid, be it rectangular, biased arrays, hexoganal, etc.
Additionally, most backup systems include redundancy in the written patterns, to protect against degradation due to environmentla exposure. The most common I am familiar with is the storage of a reversible cyclic redundancy check (CRC) in the written blocks. The block size varies from program to program, as does the compression algorithm chosen.
So if we assume a rectangular array of bits, with mild bias, we get a grid #. If each bit on the grid is halved in size, the data density, barring other changes, is quadrupled. Changes to the pattern made possible may increase this further, as well as advances in the heads.
Current heads can only read at a certain speed, so a trade off is made between spool speed and data density, meaning that not 100% of the space on the tape is lines of data, there is alos white space, unused on the tape. If a better head can read a more densly packed datastream, then you could very well make a 250x increase in total capacity.
It's been a long time since I worked with tape drive technology, so this is just an approximate explanation, of course.
You can have it fast, accurate, or pretty. Pick any 2.
Very informative. I think /. should stop linking to that guy's site. And if it does, then he'll probably find some other way to do the same thing. Maybe he'll make a new account and a new website and spend a little time altering the articles he copies.
Heck, maybe I'll do the same if I'm in the mood for some easy money. What's better than some geeks making money for a geek?
Improve at backgammon rapidly through addictive quickfire position quizzes: www.bgtrain.com
Heh, I have a 250MB (compressed) tape drive in my closet, with transfer rates of 6MB/min. Would take all day and 13 tapes to back up my 2GB system. :)
Is that terabytes or tebibytes?
There are 11 types of people in the world: those who understand unary, and those who don't.
hard to imagine a 0.05 micron track on anything flexible being readable.
Apocalypse Cancelled, Sorry, No Ticket Refunds
They aren't preparing it. They're more than two orders of magnitude, and well over five years away. This is another typical /. "someday we might make a product you want" article. ("Animated holographic XXX playing cards running Linux on Crusoe? Save your allowance kiddies!!!")
From TFA: "researchers say they expect to one day build cartridges that can store as much as 100T bytes of data."
One day?
"in order to store more than the 1T byte of data that IBM is planning for its next-generation products"
"he said cartridges that can store a terabyte of data will hit the market within 18 months."
So, the next generation (only 18 months away) will feature 2.5x more storage. Only 100x more to go. We're practically there!!
"His group of ten researchers hopes to shrink that size down to about 0.5 micron, or 500 nanometers, within the next five years."
Hmm, 2.5x in 18 months, and then a little over twice that long after that they'll have the right core technology in place to get to 100TB.
Steve Jobs hoped IBM would hit 3GHz in one year. Look how well IBM performed on that one.
'"This will carry us all the way to the 100T byte regime," he said.'
So, in five years they hope to have the core technology in place that they'll then have to refine in order to get to 100TB per tape. How long will that take?
"Narayan was reluctant to predict when IBM might bring its first 100T byte tape devices to market"
So, even longer than five years.
AC, whoever you are, you've done outstanding research, and I salute you.
I'd make you a Friend, if I knew your ID.
-kgj
-kgj
My hard drive blew up just yesterday.
Clearly, it is a conspiracy, started by CowboyNeal.
100 terabytes of storage and we're still using magnetic tape... A fragile, 60+ year old technology?
Please, all caring slashdotters, I ask the following of you:
Copy and paste this article as an AC reply every time you see a Roland "Fucky-facey" Piquepaille article. It would remove much AC discussion and it puts quite clearly what so many others have voiced.
THIS THING CAN TURN ON A DIME, MACROSSZERO STYLE ALSO FUCK BETA, ~NYORON
decide what to use. For me, 200-300GB tape drive would be OK if I KNEW the data won't go bad in 25 years (at least) if cassettes are properly stored (cool, dry place, no strong magnetic fields).
And then I'd just dump all my data onto a few cassettes, and send one set to my relatives 5000 miles away.
100TB is useless if you know the data will "die" in a few years.
The analysis of the data transfer 'problem' linked by this article is as simple, straighforward and totally ridiculous. We are all well aware of the dangers of using a simple linear aproximation to predicte the future in regards to microchip improvements and we shouldn't be stupid enough to accept it in this case. Admitedly the author of the comment doesn't make the same mistake of assuming that data transfer rates will be linear with respect to time but he makes the equally unsuported assumption that data transfer rates are linear (in fact a constant multiple) in data storage. Without any reason to believe these technologies progress at the same rate this is no more than a guess.
Moreover, there are particular reasons to believe the data-transfer rates are not a particular problem. For one even if his prediction is correct it isn't clear there is really a problem. In a device often used for backup 3days to fill might not be unreasonable. Moreover, storage transfer rates are often limited by the data transfer rate which is usually far less than the system bus speed or other dedicated high speed communications (interprocessor communication). Thus suggesting it is merely the cost-performance tradeoff which is limiting storage transfer rates not a pure technological problem. Indeed, if the main system bus speed does not increase fast enough so as to allow us to implement tape transfer as it's own specialized bus then memory literally will not be able to produce data fast enough to fill the device anyway.
If the worry is instead that the tape drive simply won't be able to read or write data then one should remember that tape/drive access is extremely parrallelizeable. If the IBM technology for writing the tape does not already proced at high speed nothing prevents stacking several of them next to each other so an entier slice of the tape is read at once.
If you liked this thought maybe you would find my blog nice too:
So much collection, so little bandwidth and CPU power to crunch it all.
It's becoming apparent that more data is not nescessarily a "good thing". As the amount of data collected grows, it's becoming harder and harder to reduce that data to something "meaningful".
When I listen to "Pachelbel's Canon" (it IS the holiday season after all!) I don't need to "know" every single frequency delta for every single instrument.
It's the aggregation that's important.
Seems to me that more data, may actually have the reverse effect of that which is intended. The entire purpose of data collection is to track/process "what's important".
Adding MORE data to the mix just adds more levels of complexity/"data filtering" without a clearly defined benefit (in most cases).
Who cares if the guy buying a phillips head screwdriver at "Home Depot" has "blonde", "sandy blonde", or even "dyed blonde" hair???
It's not relevant to the business objective of selling phillips head screwdrivers.
The 400gb figure is "compressed", i.e. marketing fantasy. LTO3 will be 400gb native, but it's not available yet, and the media will cost a fortune for a while after it's introduced.
Tape these days also has error correction, actually generally multiple layers so you can completely wipe out a region of tape and still restore the data that the region held. You need a RAID to do that with disks.
Tell you what, take a tape cartridge and a disk drive holding the same data. Drop both of them on the floor from 5 feet, which one do you think you're more likely to restore from?
Archiving to disk is just asking for trouble.
Turing machines use an _infinite_ tape. Your puny 100TB tape doesn't quite cut it.
In a sense, the length of tape = capacity. I know there are limits to how much you can wrap around a reel but really, if you took today's tape capactiy in a 9 track format you would have (scientifically speaking ) a fuckload of storage.
I could back up the internet on some travan sludge if the tape was long enough.
Are we primates or are we Mice?!!
Gimmie a 9 track size with SDLT or whatever is best at the moment and add a few extra crc's to the mix, add some splice handling and glue a friggin CD-R to the back of the reel for table of contents info.
Oh yeah... One of the BBC's first radio recording devices used razorblade sized steel to record on. they had to build a whole room with thick walls in case the 'tape' broke because it would unravel and slice through the normal walls... iirc a few techs got cut to ribbons..
Firefox &
Actually it's not the backing up taking 3 days that IT people should worry about.
;).
It's how long it takes to restore the data....
I'm willing to bet it seems longer when the Boss and the Customer is standing behind you making "encouraging noises".
So far the data transfer rate problem is a reality with most tape drives. So much so that it seems like buying 200GB SATA HDDs for backups is a pretty attractive option. Especially IF you can safely hotswap the SATA HDDs.
Most modern ATA HDDs can transfer at 40-60MBytes/sec, sustained. Most tape drives are 1/10th the speed (or worse!), and the price (including media) vs performance+capacity isn't very attractive.
Tape drives could be fine for archival stuff. But they don't seem that useful to me for backups. It's probably OK to take 3 days to fulfill a court order to restore data archived 3 years ago. Heck the boss might be happy if it _has_ to take longer
What about MO drives? Do we have to keep continuing to see MO drives lag hard drives? We have 400 GB or possibly larger drives out already, and we are stuck with 9.4 GB DVD drives?
Are we really stuck with other hard drives for backup?
Award winning hair styling and hair coloring, 100% natural hair extensions, pictures and photos of perms, glamorous updos, bridal hair updos, prom updos, wedding updos, cute hair updos, teen updos, fun updos, down dos, down-dos, teen styles, body waves, for men and women. Framesi, Bio Ionic, Consort, Wella, Paul Mitchell, Matrix, Rusk, Clairol, Shades Q., and more.