Slashdot Mirror
220TB Tapes Show Tape Storage Still Has a Long Future
alphadogg writes: IBM and Fujifilm have figured out how to fit 220TB of data on a standard-size tape that fits in your hand, flexing the technology's strengths as a long-term storage medium. The prototype Fujifilm tape and accompanying drive technology from IBM labs packs 88 times as much data onto a tape as industry-standard LTO-6 systems using the same size cartridge, IBM says. LTO6 tape can hold 2.5TB, uncompressed, on a cartridge about 4 by 4 inches across and 2 centimeters thick. The new technologies won't come out in products for several years.
They use tapes to store all that data they get from smashing tiny bits together. Totally forget how much one of their tapes hold, but at the time I remember thinking it was a lot.
What is the write speed of this technology ?
If it is too slow you will have to resort to more complex backup strategies, which is never a good thing.
Although it would still be acceptable for archiving I presume.
4 inches by 4 inches by 2 cm weighs 2 troy ounces and gets 40 rods to the hogshead.
Wake me when tape is reliable AND costs 10% of the $/GB of hard drive storage.
Worthwhile for enterprise... maybe. I haven't even looked at a tape backup in decades, but I do not relish paying more for a single tape than an entire 2TB HDD... as a consumer, or even as an enthusiast. It's cheaper and possibly more reliable to do backups to BD-R at this point, or simply use redundant HDDs as backup devices.
Write Once Read Never
That's how AWS Glacier rolls
And much like tape, it seems like the random mixing of imperial and metric measurements won't ever go away, either :)
Site & blog: http://www.mayaposch.com
what decade is this? 2000? wouldn't it be easier to use a 3 TB external USB drive that has random access? Suppose I need to access one file in the middle of the tape. The tape drive would need to roll the tape from the beginning to the middle.
Sure, the tapes themselves may be cheap. But the drives are quite expensive.
comment first, facts later. http://chem.tufts.edu/AnswersInScience/RelativityofWrong.htm
Most enterprises are likely using some form of virtualization and backup facilities using disk snapshots/etc (Veeam is awesome BTW). The need for near real-time restore kills tape backup for anything other than long term archival from my perspective.
Properly stored tape has a lifetime of decades. See if your hard drive will spin up after sitting for one decade.
Only the State obtains its revenue by coercion. - Murray Rothbard
You said "flexing". Hehe. Hehehe.
"...The new technologies won't come out in products for several years."
Er, several years?
For a minute there, I thought they were referring to the restore time for a full cartridge.
Capacity isn't really the problem with tape media. It's sitting around waiting for ages if you ever have to actually execute a full restore of that much data from tape.
Not quite sure why it remains a viable solution for that reason alone, especially in this era of the InstaTwitterVine level of instant gratification. Spinning rust in the cloud might be a bit more volatile, but it will likely always be a hell of a lot faster.
Bad for everything else. It also slows down any work considerable if you suddenly need access to a file in the middle of the tape.
How long does the tape media last until it deteriorates to the point where it becomes unreadable?
I've abandoned my search for truth; now I'm just looking for some useful delusions.
Last time I looked tape was as expensive as hard disks per GB and also a lot slower too and needed a seperate and quite expensive drive.
All in all, I do not care about tapes anymore.
"Life is short and in most cases it ends with death." Sir Sinclair
... as long as you need it to minus one second.
Knowledge is how to play a game, intelligence is how to win, wisdom is knowing what game to play.
Why isn't anyone reporting on IBM's advances in punch card microscopy? By reducing the size of the punch hole and using modern encoding systems, they've greatly increased the amount of data that can fit on a standard Hollerith card. Those who can't wire the plugboard of an IBM 407 are going to be left behind.
> on a cartridge about 4 by 4 inches across and 2 centimeters thick
10 x 10 x 2 cm
4 x 4 x 0.8 inches /sigh
Just read HP is bowing out of cloud storage for the most part. Other then supplying in house cloud servers for large customers. I think its clear consumers are reluctant to embrace cloud storage in any meaningful way. Perception of privacy issues, lost information, access limitations because not everyone in America has great internet service. These all lead to people not trusting cloud storage for mundane file storage. Even companies are re thinking the cloud and how much risk it involves vs having internal solutions. Really, when you think about it, many people do not have that much to store, and basic storage is all they need. I myself use very minimal storage in the cloud and most businesses I know of still use a local storage option. Any comments I have read about cloud storage always involves some sort of compatibility issues, files missing, or files lost, even corrupted files on occasion. Many people still do not even bother with any kind of backup. Local or otherwise. The other problem is that unless storage is free, many don't buy anything extra. This has to weigh on any company supplying storage on how they pay for the infrastructure?
(micro)etching on gold. unless we can find a way to lay down diamond or corundum layers and etch onto them. scientology is on to this.
Someone should send a shitload of these to the IRS.
When Fascism comes to America, it will call itself Anti-Fascism, and tell you to give up your guns.
"I don't understand why anyone would use tape. After all I've never used it/not for years/etc."
If you feel the urge to say words like this, just stop after the first 3 words. You don't understand. That's enough right there.
Some people have volumes of data that you cannot fathom. Some organizations have use cases that you haven't encountered. Maybe even some organizations make decisions that' could be handled another way, and maybe that different way might be better too. But's that's speculative based upon NO information.
The continuing demand for tape pretty well speaks for itself though.
Tapes have never gone out of use for large databases. The tape storage is cheap per bit compared to other formats. We know the life of a bit on tape is finite, and we know the random access time of tape is horrible. However, suppose you are providing a reliable backup service. You will have at least three copies of every record at any time, with probably a fourth archive kept separate for legal reasons. Ideally, the three copies will be in different geographical and economic zones, so you can survive the total loss of one. You will be checking these copies against each other and re-writing the data onto fresh media at regular intervals. You know the archive is good because you have a scheme for checking it and re-writing it at regular intervals. If you compare a tape in an actively maintained archive against a hard disc you keep on a shelf and never read, then the tape archive will probably be the safer of the two.
Tape is not really a consumer product, even if the tapes and tape readers are affordable . I doubt if many consumers have the discipline to maintain their own archives to this standard. I know of several good-sized companies that have kept tape archives that turned out to be no use when they had to be read. I long for the day when crystalline molecular memories will give us moles of stable bits in a few tens of grammes of material. But until then, tape seems to work.
> Tape isn't dead, but it's not worth it for small quantities
The cheapest LTO-6 drive on NewEgg is $1500, and Sony has the tapes for $18/TB. External hard drives are running about $35/TB. So you need ~90 TB for cost crossover on sheer data volume, not considering usability and reliability. So I would agree, with those kind of prices, you might want to *start* thinking about tape when you get to 100 TB, because 1 drive isn't very reliable. It might work for backup storage, since you can get by with a broken tape drive for however long your backup cycle is.
Actually, the article gave all metric primary measurements, and English in parentheses for enough of them for the metric-impaired to understand the scale.
"...about 10 by 10 centimeters (4 by 4 inches) across and 2 centimeters thick"
So apparently, it was the OP who took the queue from NASA.
Funny how a discussion on tape has devolved into one about types of RAID.
The most "enterprisy" array can still be toast due to an electrical event, deliberate corruption, accidental corruption, fire or theft. That's when you want the option to restore to new hardware from tape.
Also if you have many TB of data that you never want to lose but nobody is likely to look at this year.
The current highest capacity is 8.5 TB native in StorageTek (From Oracle) T10000D. The highest capacity from IBM is 2.5 TB. Needless to say, IBM needs some publicity with imaginary product to fight against real product. Here is one year old article on the same topic. The imaginary capacity has gone up from 185 TB to 220 TB.
I think you're out of date by several months. IBM TS1150 holds 10TB uncompressed per tape cartridge, not 8.5TB (or 8.0TB -- to get up to 8.5TB is a little "weird"). Sustained data rate is 360MB/s, not 252MB/s. Yes, LTO state-of-the-art is well behind both IBM TS1150 and Oracle STK T10000D.
I am the technical architect for the Diagnostic Imaging solution for a number of National Health Service Trusts in the south of England. Currently, to provide the imaging for CT, MRI, a bit of cardio and a whole lot of plain xray and ultrasound, we have an archive of about 300TB of data. Replicated, and with local caching that is about 1PB of deployed storage. I am paying heavily for a second site replication, old school kind of solution, and it really doesn't scale well. Now I am looking at adding Digital Pathology to the shared imaging solution, and I really don't think that having two data centres growing to keep an online second copy of all of that data is viable; not least due to the cost of electricity of keeping a second copy spinning.
So, alternatives to this are:
Keep a second copy in the 'Cloud'. No clue on what my RTO would be.
Tape: Defined RTO, RPO, mostly passive means I don't have to worry too much about power consumption. I can keep multiple generations of data on tape if so required. My existing applications support Information Lifecycle Management and Online/Nearline/Offline storage heirarchies
Why would I NOT want a high density tape store as part of my solution?
> Tape isn't dead, but it's not worth it for small quantities
The cheapest LTO-6 drive on NewEgg is $1500, and Sony has the tapes for $18/TB. External hard drives are running about $35/TB. So you need ~90 TB for cost crossover on sheer data volume, not considering usability and reliability.
People who quote that hard drives are cheaper than tape always leave out the cost of electricity and reliability. If I'm going to tape, that is one of the reasons whether it's backup or archiving. I can take that tape out and store it w/o power for years and reliably read it back.
How long can you reliably do that with a hard drive? The mfg don't design drives for that, they design for always powered up drives. If I need that, I probably need to test for it and that can change with models and firmware settings. So you might have those costs for a powered off drive. If it doesn't last as long, you have medium exchange.
HDs are more delicate. I can reliably ship a tape cross country and read it on the other side. USB hard drives, not so much. I can put tapes in a vast array that a robot retrieves from so human hands don't damage them transferring them. I can't do that with HDs.
I put ~ 20 GB on 4mm DATs in the 90s (1.3 GB/ea) and read them back 10 years later. Drive was ~ $1k, tapes $10 so the cost was ~ $1200. Disk was $100/GB (probably more, 4GB drives came out ~ 97) so I would've need $2000 of them. If I needed to keep them spun up how much was 10 years of electricity + the SCSI interfaces ($200 * 2?) to keep them running + the enclosure to put them in (20+ drives? ).
Costs for the drives *today* would be much lower, but the electricity over 10 years is still there.
A good backup strategy involves reading the tapes back on a fairly frequent basis to make sure your tapes are still readable. Regardless of it's reliability, tapes do drop bits at a similar rate to hard drives at rest. You're very lucky that you have read 100% of the DAT tape without any issue (or maybe you didn't notice it, bit errors are hard to notice until you need that particular bit). Also, most companies really don't care about their data from 10 years ago (and if they do it's because they want it GONE in case of discovery).
Back then, tape WAS the cheaper option. Today, hard drive storage is on par as far as investment cost. Most tape strategies involve a large(r) tape robot and multiple heads which is where the expense comes in (also energy costs). With hard drives that is less of an issue and hard drives can also be spun down. Hard drives are also better at random access which is generally what you'll need when restoring day-to-day backups. Massive failures are usually not the problem, backup restoration usually boils down to that user wanting to get a version of that Word document from a year ago but they forgot what it was called back then. Reading through a tape for that kind of stuff is SLOW to the point of being infeasible.
Tape backup is still a good solution but it's being outgrown because hard drives are faster and for most people what it can provide is 'good enough'. Tape is great if you have so much data that you need it because of the density (a rack can hold 100's of PB worth of tape but only ~5PB worth of hard drives).
Custom electronics and digital signage for your business: www.evcircuits.com
Yep, they may also have lightning strike them from a toaster or freeze solid if I walk them past a fridge. How do you survive in the modern world with such a misconception?
You think that above poster would have worked something out about field strength by the way the magnetic stripes in their credit cards don't get wiped by trips in an elevator. A bulk eraser exposes tapes to a couple of orders of magnitude more of a field than if the tape was resting directly upon the casing of the elevator motor, let alone inside the elevator which is only going to be near the motor at the top of the shaft.
There are 10 TB hard drive right now and 10 TB SSD:s coming in 2016. It's not hard to see a race between these manufacturers resulting in large discs. With tape manufacturers dragging their feet there could be 220 TB or larger drives on the market before the tapes hit production.