LTO-8 tape media patent lawsuit cripples supply as Sony and Fujifilm face off in court

A courtroom showdown over a patent infringement allegation that Fujifilm brought against Sony is crippling supply of LTO-8 tape media stateside and further afield, and a resolution doesn't appear to be imminent. Sony has confirmed to The Register that it is currently not supplying the US market with LTO-8. The case dates back …

[Anonymous South African Coward]

The spokeswoman added, "To date, Fujifilm has not offered LTO8 for sale."

whut? then why the squabble?

IMHO a patent can be enforced only if the patent holder makes use of it, produces items which will then fall under the patent

no items or production, no patent protection needed, so why bicker about it then?

[Aitor 1]

Wrong

If you hold the patent, you can prevent others from using it, but you do not have to use it yourself.

[Mage]

Re: Wrong

Sadly true. The OP is maybe thinking of Trademark protection, which is complex.

v

The Elephant in the room? Why are ANY patents allowed on any recent kind of tape cartridge? Magnetic recording was first demonstrated in the Victorian Era. Changing the backing media, the magnetic formulation, recording mode, the binder glue, the track layout or the casing are not real innovations. They are incremental developments.

The biggest problem is the USPTO.

[Anonymous Coward]

Aitor is correct

It is a trademark that's "use or lose", not a patent. You can patent something and sit on it if you like, as say an oil company might do if someone found a way to build Back to the Future's "Mr. Fusion" device. But they could only block it for the 20 year patent term, then it would be public domain.

[CountCadaver]

Re: Aitor is correct

Until they successfully lobby congress etc as its "harms american businesses" "stifles investment" et al

[Anonymous Coward]

Re: Aitor is correct

Much more likely that Congress would be lobbied to nationalize the fusion patent. The US has no interest, beyond that of a few oil companies which are a tiny fraction of the GDP, in maintaining the oil economy.

The US government would be fine with causing the oil price to crash to $10/bbl, because the middle east would suddenly get a LOT simpler when we don't need to kiss ass with Saudi Arabia. It would topple most of those governments, but promises of "we'll build you shiny new fusion plants free of charge if you let us dismantle all your nuclear plants and cart away all your radioactive materials" would be cheap when measured against our all to frequent wars.

It would also make South America a lot simpler as well, especially since we don't have an Israel there we feel we need to protect as if it was US territory.

[martinusher]

Re: Aitor is correct

>dismantle all your nuclear plants and cart away all your radioactive materials" would be cheap when measured against our all to frequent wars.

Alas, you're confusing cause and effect. Our interest in the ME in the control of oil resources has created a huge business opportunity providing the materials to protect those interests -- its the 'military industrial complex' that Eisenhower warned us about. If we removed the need to protect our interests we'd have no need to spend vast sums on our military which I reckon would do far more damage to our economy than high oil prices ever did.

We're in a classic junkie dilemma. We will do anything to get our next fix - we can rationalize any behaviour, explain away any misdeed because the fix is paramount. Its killing us but we can't stop because we fear the consequences of stopping so much that we just carry on, ignoring the future. We need intervention.....

[Anonymous Coward]

Re: Aitor is correct

Unfortunately we now have the "global war on terror" which was the twin virtues (from the position of the military industrial complex) everywhere instead of just the middle east and can NEVER be won, which is where the cold war went wrong for them.

So they no longer need the pretense oil provides, and could even avoid having the large military itself, when you can spend an unlimited sum on survellience, satellites, drones, and so forth to fight the GWOT.

[Anonymous Coward]

Petrodollar is very real thing

"The US has no interest, beyond that of a few oil companies which are a tiny fraction of the GDP, in maintaining the oil economy."

This is patentably false: The value of dollar itself is directly tied to oil and oil price.

Remove the oil and value of dollar will be zero in months.

So basically everything measured in dollars, including GDP, is depending on oil.

[Alan Brown]

Re: Aitor is correct

China's the coming provider of those nuclear plants - ironically using technology the USA threw away in 1972.

[Alan Brown]

Re: Aitor is correct

"You can patent something and sit on it if you like, as say an oil company might do if someone found a way to build Back to the Future's "Mr. Fusion" device"

Or if you're an oil company you can buy up and sit on critical Nickel-metal hydride battery patents, forcing automotive battery development down the more dangerous lithium-ion path instead and taking Toyota's entire californian fleet of RAV-4 EVs off the road.

https://en.wikipedia.org/wiki/Patent_encumbrance_of_large_automotive_NiMH_batteries

[InsaneGeek]

Sure this will be great on the long term

It's not like tape sales has been shrinking year after year or anything. Surely this wouldn't cause a number of the shrinking count of companies to just move to disk backups vs tape or anything. That would never happen

[Anonymous Coward]

Re: Sure this will be great on the long term

Yep. Dickheads.

[Nick Kew]

Re: Sure this will be great on the long term

Is there enough of a market (that can't easily switch to AN Other technology) to matter?

If yes, that's another innocent victim of US patent nonsense. If no, it's merely a case of suppliers aiming at foot and pulling trigger.

[J. Cook]

Re: Sure this will be great on the long term

There isn't a good, viable alternative to LTO for tape; everything else except for the venerable 3592 series. Everything else in the tape arena is dead, or too small.

[doublelayer]

Re: Sure this will be great on the long term

"There isn't a good, viable alternative to LTO for tape;"

That's true, but disk backup is getting cheaper, and it will start running up against tape's market even more than it already has. Tape's benefit used to be that it was the cheapest way to store a bunch of data per gigabyte. That's still the case, but its margin over hard drives is falling due to innovation in disk and lower demand for disk because SSDs are more popular for hot storage. If the price of tape media is pushed much higher, the margin will become so minimal as to be ignored.

The other problem is that, in order to use the tape, one also needs to purchase readers for the media, and those can be very expensive. Even the standalone LTO tape readers that read only one cartridge at a time and don't load it for you can cost several thousand. That's enough to deny them all of the very small business market and most of the small to medium business market. The last thing they need to do right now is convince large businesses that there will be problems getting media for their drives and replacing that media.

[HobartTas]

Re: Sure this will be great on the long term

"The report showed a record 108,457 petabytes (PB) of total tape capacity (compressed) shipped in 2017" https://www.lto.org/wp-content/uploads/2018/03/LTO-Shipment-Report-Release_2017-FINAL.pdf so even for 10TB hard drives you would need 10,845,700 of them plus the servers and data centres to hold them, it does mention "compressed" so roughly halving that number it's still about 5 million drives for that year alone and actual tape sales in total capacity are increasing each and every year. LTO9 with 24TB uncompressed cartridges may come out at the end of this year.

[doublelayer]

Re: Sure this will be great on the long term

But that's about the same number of tape cartridges shipped, too. In terms of physical size, we're almost exactly the same. The major issue is cost. If you were just to buy that much disk and that much tape, the tape would definitely be much cheaper. Most people, however, are buying tapes and readers. For example, if you bought ten tapes and ten disks, but also needed reader for each, the combined price of the tapes and reader would now exceed the price of the disks and a standalone enclosure. If you scale up to the area where you need more than a standalone single-cartridge reader, the price of a library that can handle a lot of media is once again much more than a multi-disk system.

That's why tape is limited to things at scale. There used to be a time where a small business would back up their local data to a single set of tapes, but that market has almost entirely died to be replaced by disk cartridges or cloud services. Large businesses will eventually reach the point where they're going to obsolete their old LTO systems and have to decide whether they're going to invest in the latest LTO version or something based on disk. If LTO keeps getting more expensive while the increased demand for hot storage keeps disks innovating and being manufactured, the difference in price will fall and may fall enough to lose LTO a big customer.

[Crypto Monad]

Re: Sure this will be great on the long term

> There used to be a time where a small business would back up their local data to a single set of tapes, but that market has almost entirely died to be replaced by disk cartridges or cloud services.

There used to be a time when small businesses would run their own servers, but that market has almost entirely died to be replaced by cloud services.

[Anonymous Coward]

Re: Sure this will be great on the long term

"Most people, however, are buying tapes and readers. "

No, they are just buying tapes. A lot of tapes: One set for every weekend, one set for every day in between, at least 2 sets for off-line and off-site storage.

Basic backup procedure.

And one tape drive/robot/tape subsystem, which typically lasts until next version(tm) of tape is available, 10 years.

Disks typically last 3 years and when it fails, it's all dead at once. Then you miss a part of your backup set which means backup is useless. When we are talking about backup, 3 years is temporary storage.

Disks are good for that.

"... or example, if you bought ten tapes and ten disks, but also needed reader for each,"

Fictional example. Even tape subsystems (with hundreds or thousands of tapes) don't have more than few readers as those push hundreds of megabytes per second each, not really necessary.

[doublelayer]

Re: Sure this will be great on the long term

"And one tape drive/robot/tape subsystem, which typically lasts until next version(tm) of tape is available, 10 years."

First, tape generations come more quickly than that, so they must be skipping a few. Second, your need for tape readers and writers is directly proportional to your requirement for tapes, after a given point. If you have multiple petabytes to back up, you will need more than a single tape drive. You will need at least a large automatic tape library, and quite possibly multiple smaller automatic tape libraries for different locations. You will also need a backup reader, because they might not make these readers after three more generations have been released.

However, even if you don't have multiple petabytes and are using many fewer tapes, you still need a drive. That's why there is a point at which a certain amount of data storage in disk is cheaper than that amount in tape, because the equipment required to use the storage is so much cheaper for disk than tape.

For example, a post below has stated prices for 30 GB of LTO7 at $400 and the same amount of disk at $1300. If you add in the $3000 reader for the tapes, you get

Price of tape = 3000+(400/30)*t where t=number of terabytes stored

Price of disk = (1300/30)*t

In other words, the price of disk is lower than the price of tape below a hundred terabytes. If we conclude that we need twice as much disk as tape because you consider disk extremely unreliable, it's still cheaper under fifty terabytes. And that's why there is very little use of tape in small business or places that don't generate a lot of data. A local business can keep eight copies of their data on disk and still be even with the price of a tape backup system. The problem is that the increasing price of tapes and readers due to these legal disputes and the decrease in manufacturers while disk prices are falling is pushing that boundary higher. Of course tape is going to be used when the company is large enough and has a very large dataaset. While I don't run our backup systems at [current employer, which has a very large amount of data], I'm entirely certain there are rooms full of tapes around me.

[Alan Brown]

Re: Sure this will be great on the long term

" If you have multiple petabytes to back up, you will need more than a single tape drive. "

Your rate of backup determines the number of drives you need for that task. It's usually a small number.

Your rate of disaster RESTORE determines the number of drives you need for that task. It's usually significantly higher.

Don't spec for #1 and find yourself caught badly short when #2 happens.

and $3000 is a _very_ cheap tape drive. they're closer to $16k apiece for LTO7 when in a quantum library and if you're backing up anything more than 1-200Gb/day you WILL need a robot.

[Alan Brown]

Re: Sure this will be great on the long term

"But that's about the same number of tape cartridges shipped, too"

The difference being that tape cartridges only "consume power" when installed in a drive. The rest of the time they're passive pieces of plastic which are remarkably resistant to abuse.

Not that I trust iron Mountain. My preference is to pay £5k a pop for fireproof datasafes holding 850 tapes.

[Crypto Monad]

Re: Sure this will be great on the long term

> Tape's benefit used to be that it was the cheapest way to store a bunch of data per gigabyte

And that when stored properly, it has much longer data storage lifetime than (spun-up) hard drives.

And it uses less physical space, and doesn't need power.

Plus, for many backup and archive applications, the fact that it is off-line is actually a bonus - your crypto malware can't easily get at it.

[Tomato42]

Re: Sure this will be great on the long term

> it has much longer data storage lifetime than (spun-up) hard drives.

only as long as you have a working drive to read it. LTO backwards compatibility is limited to 2 generations, and that doesn't include the interoperability issues because you are using a niche technology in a corner-case scenario with reader, writer and the tape coming from 3 different companies.

[Jakester]

Re: Sure this will be great on the long term

I quit using tape for backups years ago. Opted for portable hard drives for daily off-site storage and internal hard disks for on-site backups. My backup systems are running Linux and use Back-In-Time for the backups. Because the backup systems "pull" data from the servers, the Windows side of the network never accesses the backup drives.

Sure there is a risk of an infection attacking the Linux systems and corrupting the data on the backups, but not any more than the system running a tape backup getting infected and corrupting the tape backups.

[Anonymous Coward]

Re: Sure this will be great on the long term

"... but not any more than the system running a tape backup getting infected and corrupting the tape backups."

What?

You realise that *no-one* is having only one set of backup tapes? And at least one set is off-line and off-site?

While basically no-one is having multiple copies of disk backups.

So no, the risk of corruption is inversely relative to amount of backups: Disk: 1, Tape: As many you like.

[Peter Gathercole]

Re: Sure this will be great on the long term @Tomato42

Disk is not likely to fare much better. Try connecting a SCSI II LVD drive to a modern computer, let alone an SSA, ESDI or ST506 drive.

With the advent of NVMe m.2 SSD's, I suspect that the days of SAS and SATA are numbered as well.

And it won't be enough to store the necessary adapters alongside of the disks, because we've seen the demise of ISA, EISA, MicroChannel, PCI and PCI-X as adapter busses.

You might be lucky and have USB survive as an interface for a bit longer, but this will rely on the OS's continuing to support legacy disk. And whatever encryption you're putting on the disk!

Long term data storage is in a bit of a mess. I pity future generations of historians when so much of todays information will no longer be available.

[J. Cook]

Re: Sure this will be great on the long term @Tomato42

Heh- USB to SCSI-I external adapter, then to II LVD via a set of converters. it ain't pretty, it ain't fast, but it's functional. I did that for (of all things) connecting a SCSI Jazz Drive to an eMac to transfer the data off the jazz drive to the new system.

One of the 'big' off-site backup companies (a rusty hill, as it were...) at one point offered data translation services, and had some of the old technologies sitting in their storage facilities for just that reason.

[doublelayer]

Re: Sure this will be great on the long term @Tomato42

"I pity future generations of historians when so much of todays information will no longer be available."

I think they'll be fine on that score. We certainly make it hard to get at some data, stored on hardware that might not be readable, but we also create and preserve a lot more data. First, we print out a lot of completely irrelevant information because paper, ink, and the means to make the ink say something without a lot of human effort are so cheap. Second, we have places whose job it is to keep a lot of data in a decentralized manner. Some of these places have a clear purpose that would be useful to historians, like academic paper storage and wikipedia, but as useful as the Internet Archive is, I doubt they'll really need the multiple identical captures of dead sites.

I pity future generations of historians when they'll have to read so much junk. They'll get a lot of useful accurate information, but they'll have to read a lot of internet users insulting one another, people's social media posts, and mangled or false information on the way.

[Alan Brown]

Re: Sure this will be great on the long term @Tomato42

>> "I pity future generations of historians when so much of todays information will no longer be available."

> I think they'll be fine on that score.

I point you to the time and effort required to retrieve a working copy of the BBC Domesday book.

[Anonymous Coward]

Re: Sure this will be great on the long term

"only as long as you have a working drive to read it. LTO backwards compatibility is limited to 2 generation"

Totally irrelevant for most users, I'll say. We have machines with LTO3 installed sitting in the corner just in case we need to restore something from 1990s.

[Alan Brown]

Re: Sure this will be great on the long term

> > it has much longer data storage lifetime than (spun-up) hard drives.

> only as long as you have a working drive to read it.

This is why you have migration policies ensuring that the data moves long before the drives expire.

I've had data on DLT, LTO2, 4 and now 6 and I can retrieve it in an hour or less.

On the other hand I've had people show up with 9-track NASA tapes from the 1970s they're hoping to read (The answer is "sure - go see this company in Southhampton - the charge is about £400 per reel") and the guy with 3500 exabyte tapes he thinks I'm going to hand feed to a single drive (that's been dead for years) when he wants a restore is vastly outta luck (not my circus, not my monkeys)

[doublelayer]

Re: Sure this will be great on the long term

That's correct, as well. However, many places use spun down disk for offline backup. It may not have a perfect lifetime, but those who choose it usually fall into one of these categories:

1. They keep backups for emergencies and don't need decades of reliability.

2. They're worried enough that they will have data corruption that they keep several copies of data. Having chosen this, they can get similar reliability with a greater number of hard drives.

3. They want convenient access to data at multiple sites in case of emergency. They choose disk because they can afford to buy several readers and leave them in the backup location unused, or even keep a server that can accept them into a bay so a person at the remote location can load a drive for them to access remotely. Buying that many tape readers would be prohibitively expensive and most locations that specialize in data storage of the kind they just improvised charge a lot more and are in the cloud services group.

Nothing is perfect and will store data forever without risk, but as disk price falls, the price of reliable enough is falling too.

[ibmalone]

Re: Sure this will be great on the long term

Interesting IEEE Spectrum article last year (don't think it's paywalled), about this. https://spectrum.ieee.org/computing/hardware/why-the-future-of-data-storage-is-still-magnetic-tape

The short version is spinning disc technology got developed over the past couple of decades while tape languished, tape is now catching up in terms of innovation because there's more room to push the technology. You can see this in HD capacities topping out in the past few years, but tape continuing upwards (nobody is seriously talking about 24TB HD, but it's the next step on the LTO roadmap). Of course, if SSD can keep its current pace it may obsolete both in a few years, but only if this tape infighting means they lose enough market share that development stalls.

[schubb]

Re: Sure this will be great on the long term

30TB SSDs are available today, no where near consumer availability at 12k USD per drive, but they exist and are in use.

[ibmalone]

Re: Sure this will be great on the long term

That's cool, but I was distinguishing HD (which doesn't look like it'll go much past 12TB) from SSD. And nobody uses tape for primary storage. They use it for depth.

30TB SSD 12k USD.

36TB HD SAS ~ 1.3k USD

36TB LTO-8 ~0.9k USD (if you could get it)

30TB LTO-7 ~0.4k USD (normal -7, couldn't find M8 pricing quickly)

If you have some need to keep TB of data for months or years you'll probably think twice before buying SSD for each copy. Maybe in 5-10 years 30TB will be pretty affordable, maybe if the LTO lot manage to make up they'll have 64-128TB tapes by then.

[J. Cook]

Re: Sure this will be great on the long term

Yep. an external SAS LTO7 drive is a shade over 3 thousand brand new.

If the SAS or SATA interface standard remains stable (and there's nothing that says otherwise) I can see SATA SSDs eating the majority of tape's lunch for longer term backups, or offline archiving.

(for what it's worth, [RedactedCo] is using tape for long term archiving and compliance-required backup data storage of certain data types. While we don't do cloud (for compliance and political reasons) we do have off-site backup that works... well enough for our purposes.)

[Anonymous Coward]

Re: Sure this will be great on the long term

"but disk backup is getting cheaper, "

Yes but you know what? Disks fail.

Several decades faster than tapes. That's the problem.

Even powered-off disks lose data relatively fast compared to tape so no, disk backup isn't a long term solution.

As a short term solution it is good.

[Alan Brown]

Re: Sure this will be great on the long term

"but its margin over hard drives is falling due to innovation in disk and lower demand for disk "

The corollary to that is as disk sales fall it becomes uneconomic to make them and the price will either climb or they'll go out of production (probably both)

As for tape sales falling: dollar sales yes, but capacity per dollar keeps going up - and if you stay a couple of generations behind it can be surprisingly cheap (or just hold off for 12 months - the initial tape price usually halves)

[rcxb]

Re: Sure this will be great on the long term

LTO (and cheaper hard drives) killed off just about everything else. Oracle/Sun StorageTek T10000 was recently discontinued in favour of LTO-8. Only IBM is still hanging on with the 3592, and that's ironic, since IBM is nearly the last surviving LTO member. Half the LTO-8 drives out there have HP's extremely-vented design, and the other half have Big Blue's prominently big blue eject button.

A generation back, there were a handful of manufacturers of LTO drives and tapes, and the competition kept prices reasonable. Now there's only two drive manufacturers, and two media manufacturers, and this story nicely outlined the media manufacturer's current refusal to manufacturer media...

The organization really shot themselves in the fet on this one.

[John Savard]

Re: Sure this will be great on the long term

This would be the perfect time to resume production of T10000 if it's unencumbered by the Sony and Fujifilm patents on LTO-7 and LTO-8.

[Mage]

Re: resume production of T10000 if it's unencumbered

I think like most HW Oracle acquired from Sun that the T10000 is essentially dead. Oracle uses LTO.

"The T10000D drive was announced in September 2013 with a native, non-compressed 8.5 TB". The T10000E development, supposed to deliver 12TB, was cancelled in 2016. Also it's using helical recording thus drive life might be poorer. More expensive cartridges which are inherently larger as it has both spools. Technically a single spool cartridge is possible as Panasonic had such for 1/2" video in 1970s. It used a stiff leader about 3" long rather than the pin on LTO.

[Mage]

Re: stiff leader about 3" long

I only repaired one Panasonic cartridge video. Most non-broadcast VCRs/VTRs then were u-matic, N1500 or 1/2" Reel to reel EIAJ. I had a look and actually the single spool cartridge wasn't a particularly Panasonic, but EIAJ-2. Exactly the same format of B&W and optional "color-under" as 1/2" EIAJ reel to reel, but a single spool in a cartridge.

I think BBC Vera was the dead-end already obsolete linear video recorder.

[Anonymous Coward]

Sony and Fujifilm tape media patents

How exactly can you patent a plastic tape coated with a magnetic material?

“Dan Keldsen, director of customer innovation at object storage maker Wasabi”

Whenever I see people refer to themselves as possessing innovation, I automatically assume they don't have any, like having a director of beef innovation at Angus Steakhouse.

[Warm Braw]

Re: Sony and Fujifilm tape media patents

At least the consumers at Angus Steakhouse volunteer the words "well done" from time to time, though it's probably rare elsewhere...

[The Original Steve]

Re: Sony and Fujifilm tape media patents

Bravo Sir. Bravo

[John Savard]

Re: Sony and Fujifilm tape media patents

It's true that Ampex already did that, and that one has expired. But LTO-7 and LTO-8 use a lot of fancy technology that was novel to achieve high storage densities, and that technology was very legitimately patentable.

[Mage]

Re: very legitimately patentable.

Then the Patent system is broken.

Even the major innovations since 1930s shouldn't really be patentable:

1) Plastic tape with oxide coating replacing wire or metal bands

2) Putting spools in a cartridge (1958 RCA predates CC or LJA 8 track)

3) The 1/8" cassette with no spools (1962)

4) Lear Jet single spool endless loop 8 track (stupid and not the only version of the idea).

5) Traverse & helical recording with an FM carrier for Video. Both mid 1950s in Ampex. Development started in 1952. Their early audio machines based on WWII German designed tape.

6) Digital data on multitrack linear and simulating video on helical.

Things like CrO2 or "Metal" instead of Ferric are minor. Video drove the development of "better" tape density.

The last major innovation was "helical", a natural variation of Ampex transverse on 2" tape so as to use skinnier tape such as 3/4" and 1/2". Akai even did a 1/4" helical in 1970s. The 44.1KHz CD Audio sample rate was to suit archive on a helical tape system originally developed for video.

[Mage]

Re: very legitimately patentable.

Also note that single spool cartridges (like LTO uses) where the tape end is pulled out are really old. Panasonic had one for video in the 1970s. High speed high capacity linear tape predates helical, so not novel for LTO.

[Anonymous Coward]

Re: very legitimately patentable.

>Then the Patent system is broken. Even the major innovations since 1930s shouldn't really be patentable:

Don't get me started on copyright, you'll make me turn green and shred yet another pair of trousers.

[aks]

Re: Sony and Fujifilm tape media patents

fancy technology in the tape or just in the reading and writing of them?

[Charles 9]

Re: Sony and Fujifilm tape media patents

How about new magnetic substrate materials such as Barium Ferrite (being used now in LTO-7 and up) and perhaps Strontium Ferrite in future?

[phuzz]

Re: Sony and Fujifilm tape media patents

"How exactly can you patent a plastic tape coated with a magnetic material?"

They don't obviously, instead they're patenting small improvements. As a (completely fictitious) example,a manufacturer could patent the idea of adding carbon fibres to the plastic tape, thus making it stronger, so they can make it (even) thinner and thus cram more tape into the same size cartridge (increasing the total capacity).

I imagine there's also some pretty trick technology in the read/write heads as well.

Over the last twenty years, LTO tape has gone from ~5000 bits/mm to over 20,000 B/mm. You don't get that without some kind of clever technological tricks (including in the manufacturing), and that's just the sort of thing that can be patented.

If you're really interested the patents they're arguing over seem to be (all US patents): US7029774B1, US6674596B1, and US6979501B2.

[ItWasn'tMe]

Long Time Overdue?

See title ^

[Anonymous Coward]

WTF

Two bald men squabbling over a comb

[Dwarf]

So how come ...

They got through 6 iterations of LTO and didn't have an issue, then they fall out over a dwindling sales stream.

Whats so fancy and innovative about LTO7 and LTO8 then ?

[rcxb]

Re: So how come ...

> Whats so fancy and innovative about LTO7 and LTO8 then ?

LTO up to 5 used Metal Particulate (MP) tape. LTO-6 was compatible with both MP and BaFe tape. LTO-7 and 8 moved to BaFe only.

[HobartTas]

Re: So how come ...

Actually it was the change from GMR heads to TMR heads which only IBM makes as they have had several years experience with TMR technologies http://www.insic.org/news/2012Roadmap/PDF/24_Roadmap%20-%20Heads%20-%20FormattedV5.0.pdf Up to LTO7 drive's heads were GMR and could read either type of tape media whereas LTO8+ are TMR heads and apparently MP tapes trash the TMR heads in short order if used. LTO drives can't tell what type of tape is inserted and since LTO6 can use either the LTO.ORG decided that LTO8 drives won't accept LTO6 tapes as it can't distinguish between LTO6 MP and LTO6 BaFe tapes and therefore the only safe course of action was to deny backwards compatibility and program the LTO8 drives to reject all LTO6 tape cartridges.

[Anonymous Coward]

Bye-bye tape drives

Just-in-time lawsuit. Bye-bye tape drives. Buggy whip manufacturers would be proud.

[Charles 9]

Re: Bye-bye tape drives

And what would you replace it with, especially for archival and off-site storage? I don't recall any other tech matching tape in those departments.

[J. Cook]

Re: Bye-bye tape drives

That's the same argument I had to make for keeping tape around at [RedactedCo]. It helps that there's a regulatory compliance directive that forces us to keep certain types of data around for a set period of time...

[StargateSg7]

Re: Bye-bye tape drives

You can use lasers focused on a 3D-XYZ glass cube to do permanent data writes (i.e. WORM = Write Once Read Mainly) at data capacities as high as a petabyte for a 10 cm per size glass cube.

You're basically popping a nano-sized bubble in a permanent silica media at high laser power and use another much lower power laser to read the data.

It's much more expensive since the glass cubes have to be high-purity low-creep silica + boron glass BUT it's a PERMANENT archive which should last into the MILLIONS of years! And at one petabyte per litre, what media can even come even close to matching that?

.

[rcxb]

Re: Bye-bye tape drives

> And at one petabyte per litre, what media can even come even close to matching that?

Every single bit of similar storage vaporware that has been hyped by companies seeking stupid investor dollars for the past half century but never makes it out of the lab...

[Charles 9]

Re: Bye-bye tape drives

"You can use lasers focused on a 3D-XYZ glass cube to do permanent data writes (i.e. WORM = Write Once Read Mainly) at data capacities as high as a petabyte for a 10 cm per size glass cube."

Where can I buy this technology on the market right now? Because what you describe has been in development for over 20 years with the only thing to show for it the short-lived disc-based HVD, which was itself quickly eclipsed.

[Charles 9]

Re: Bye-bye tape drives

Allow me to update myself. The technology is still out there, but apparently the practicalities of making it temporally stable and high capacity have rendered it extremely niche in nature. An organization called the Arch Mission Foundation seems to hold most of the cutting edge of holographic data storage, which they're currently using for the purposes of information preservation. If the technology matures and becomes cheaper to implement, a version of this for WORM archival purposes may become more practical for firms with large archival needs.

[rcxb]

Re: Bye-bye tape drives

> And what would you replace it with, especially for archival and off-site storage?

You get backup software that supports encrypting and uploading it to various "cloud" storage providers. Pay a monthly fee for however much you need, add to it as needed, checksum it once in a while, and then transfer one of the sets to a cheaper cloud provider when they come along.

Tape (here) is cheaper than hard drives here, but what about tape (here) versus hard drives in Romania? What about when you factor in the cost of the tape library? The shrinking cost of spinning drives, combined with the advantages dynamic allocation and easily accessible online data, means many, many organizations are going that way for more and more of their backups. Sure, any decent sized company still has a tape drive around for important long-term archival, but they're throwing FAR fewer tapes at it, and seeing less reason to upgrade to the next gen as their tape usage shrinks.

Of course that's just the latest trend. Off-site SAN replication and snapshots over ever-faster internet connections was the thing a few years back, which took a big bite out of tape as well.

[doublelayer]

Re: Bye-bye tape drives

This really depends on how much data you have. If you're a place that has a relatively small amount of data that you just really don't want to lose (say 20 TB or less), spinning disks are fine, as a few duplicated sets can be placed at all available offsite locations. If you have somewhat more than that, you can do a cost calculation for services like Amazon's Glacier, your own disks, or tape. When you start getting into the area of a bunch of data which was previously a bunch of tapes, you can redo those calculations as most of the cloud storage providers have some extra benefits for storing a very large amount of data there. As decreased number of producers and lawsuits push up the price of tape cartridges, increasing capacity of tapes and disks reduce the utility of massive libraries, and prices for storage fall across disks, SSDs, and cloud storage, the math is going to change.

[Anonymous Coward]

Re: Bye-bye tape drives

"If you have somewhat more than that, you can do a cost calculation for services like Amazon's Glacier, your own disks, or tape."

Glacier is basically write only.

Company calculated that if they want to get anything back from there, it costs about the same as tape subsystem and tapes to it.

Didn't really make sense as backups need to be verified from time to time.

[-v(o.o)v-]

Re: Bye-bye tape drives

And after a disaster, how do those hard drives in Romania get the data back to you?

Over the Internet?

Better have a 10G upstream connection, then.

I won't even bother to calculate how long it would take to download any larger than "small" amount of data. Can those cloud storage services even send the data back at 10G? Especially if the location needing DR does not directly peer with the cloud provider (which is usually the case)?

Since we are talking about LTO-8, we are obviously backing up mid-high double digits to few hundred TB minimum here.

[rcxb]

Re: Bye-bye tape drives

> And after a disaster, how do those hard drives in Romania get the data back to you?

The better providers will offer to rush ship a drive to you, for a price.

> Better have a 10G upstream connection, then.

Bollocks. LTO-7 has a maximum speed of just 2.4Gbits/sec. The far less pricey LTO-6 is only a bit over gigabit speeds. And with a smart restore, you only need to pull down a fraction as much data as you would need to pull off of a tape, so even a gigabit connection may be FASTER to restore than a tape. No old fulls and several incrementals to reconstruct the current state. You'll only be pulling a virtual full, and even then, if you're able to start from something like a month-old backup, then it'll only transfer over the network a small set of differences against your current data set. Modern backup solutions are just that advanced.

Honestly, if a little downtime is that critical to you, you've already got a replicated DR site... at least for all your most important systems. For important systems, even having an off-site tape isn't nearly fast enough, and isn't much faster (and may be much slower) than pulling it from the cloud.

When your entire datacenter gets wiped out by an asteroid, and you've got to order all new servers and start over completely from scratch, a little extra time to pull down your backup isn't going to be your biggest problem, as you're scrambling around to build everything up again from scratch. You're just going to be ecstatically thankful someone was smart enough to save a tertiary copy and leave you the means to access it at all.

[seven of five]

Re: Bye-bye tape drives

> LTO-7 has a maximum speed of just 2.4Gbits/sec. The far less pricey LTO-6 is only a bit over gigabit speeds.

per drive. Small libraries have four drives, we have 24 and there are many companies bigger than us. No matter how smart you restore, that 12TB SAP database is going be be around 12TB to restore. Same to that fileserver. I´ve been a TSM administrator for more than 20 years now, tape has been pronounced dead for at least as long as that. And I have read things like rcxb´s comment for the same length. We will see.

[Charles 9]

Re: Bye-bye tape drives

"You get backup software that supports encrypting and uploading it to various "cloud" storage providers. Pay a monthly fee for however much you need, add to it as needed, checksum it once in a while, and then transfer one of the sets to a cheaper cloud provider when they come along."

And how do you do that without running over data transfer allowances, given you're talking bulk data in the multi-TB range. Sorry, but at those sizes, Sneakernet is much cheaper. Not to mention not as reliant on third parties staying in existence.

[rcxb]

Re: Bye-bye tape drives

> how do you do that without running over data transfer allowances, given you're talking bulk data in the multi-TB range

Modern smart backup software only needs to complete one full backup, EVER, and that can be spread out over as long of a period as you want to. After that, it's an "incremental forever" strategy, where only the highly deduplicated, rolling hash, incredibly well compressed differences need to be transferred.

This strategy works great for the vast majority of workloads. "Sneakernet" would only be useful in exceptionally rare circumstances, like a massive DVR with already highly-compressed data that (strangely) also cycles out its full data set on a perhaps weekly basis.

You need to modernize your IT skill-set because you're seriously missing out on immensely useful modern developments. Take a look at Borg and Veeam just for a start.

[seven of five]

Re: Bye-bye tape drives

That "modern software" you refer to is known as "Adstar distributed storage manager" to the longer serving backup operators here. Or "Tivoli storage manager" to the younger ones. iirc it is now called "IBM Spectrum Protect". That piece of software is 30+ years old, so don´t get all shouty about it.

[rcxb]

Re: Bye-bye tape drives

> That piece of software is 30+ years old, so don´t get all shouty about it.

TSM didn't get client + server side deduplication options until v6.2 in 2010.

[tocha]

Re: Bye-bye tape drives

Dedup has no benefits during restores, for example your 12TB dedup data needs to be rehydrated so you actually restoring complete 12TB and even slower than restoring from tapes. Also no matter how you backup your data to a deduplication system, it is rarely stored contiguously on disk. Keep in mind defragmentation or housekeeping from backup dedup storage repository can cause heavy performance hits, it's common issue with SHA-256 hashing.

Tape has been IT data protection strategies for many years. Now cloud has emerged as a key technology for data protection improvement. While some may think that cloud can replace tape, and others saying the opposite, tape is not dead, the two technologies are actually complimentary when used in conjunction. Both offers different strengths to address specific IT challenges. Best practice is to keep two copies for redundancy, why not one copy on tape and the other backup stored in the cloud

[rcxb]

Re: Bye-bye tape drives

> Dedup has no benefits during restores, [...] you actually restoring complete 12TB and even slower than restoring from tapes.

Bollocks. Client-side dedupe does not do that... at all. If your 12TB dataset gets deduped and compressed down to 3TB, then it's only 3TB that gets transferred over the network, whether for a full backup, or a full restore. And it can do better than that if you're not completely starting from scratch.

> Keep in mind defragmentation or housekeeping from backup dedup storage repository can cause heavy performance hits

Again, client-side deupe does not have that issue.

> the two technologies are actually complimentary when used in conjunction

I mostly agree. I'm all for a monthly archive tape, and similar. But tape's benefits keep getting smaller and smaller, so large companies are going to have fewer libraries and cycle through far fewer tapes, as they do more and more online and cloud backups. Mid-sized companies will decide they don't need to upgrade to newer tape tech. Smaller companies will get a big benefit out of being able to forego the expense of tape entirely. Mid-sized companies will get there in the not-too-distant future.

[Orv]

Re: Bye-bye tape drives

I've been using cloud backup at home for about three years. In that time I've already had to switch providers once because one of them exited the home backup market. Switching meant doing the entire full backup from scratch, for each machine, which took over a month and put me over my cable data cap.

For home use it's worth it because no one is going to use tape at home, plus it automatically gives me offsite. But there's a definite hassle factor.

[John Brown (no body)]

Re: Bye-bye tape drives

"Tape (here) is cheaper than hard drives here, but what about tape (here) versus hard drives in Romania? What about when you factor in the cost of the tape library? The shrinking cost of spinning drives, combined with the advantages dynamic allocation and easily accessible online data, means many, many organizations are going that way for more and more of their backups."

Long term archive on tape, properly stored, is well understood and can stay reliable and readable for many years. Has anyone got any info on the lifetime of data stored on an HDD kept in ideal conditions? Are there drives made with this use case in mind?

[rcxb]

Re: Bye-bye tape drives

> Has anyone got any info on the lifetime of data stored on an HDD kept in ideal conditions?

You're thinking about it the wrong way, entirely. It's a sea change, not a direct substitution.

Offsite and encrypted but online data gets scrubbed and checksums verified on a routine basis. Malfunctioning hard drives are replaced as needed. You rely on the integrity of the system as a whole, not put blind faith in every single individual component (like a tape).

[Charles 9]

Re: Bye-bye tape drives

But then there's a Catch-22. What if you need to restore OFFLINE to get back ONLINE?

[rcxb]

Re: Bye-bye tape drives

> What if you need to restore OFFLINE to get back ONLINE?

What did you have in mind? You need internet access, account information and key config info to access your cloud services, encryption keys, and some place to grab the backup software, but that's about it. You may only need a small piece of paper to hold all that info, tucked in your wallet (with several people having copies). Fireproof safes are another good option.

You'll have at least as much trouble boot-strapping from tapes. Still need a copy of the backup software to get started, and special hardware (tape drive/tape library) and systems need to be configured to access it.

[Charles 9]

Re: Bye-bye tape drives

What if you can't trust online backups or have a legal requirement to keep things in house?

[doublelayer]

Re: Bye-bye tape drives

Then don't use cloud systems. Your options:

1. Tape. Relatively cheap (for now), known as relatively reliable. That said, don't just assume a single tape will last as long as they say, as there is always some variability and you care a lot more than the manufacturer does when you ended up with one that varied too low.

2. Disk, spun down: Cheap and reducing in price. Reading is easy. You'll have to duplicate a bit more because disks are not as reliable as tape, but the people who make claims like the data vanishing in a few years are exaggerating.

3. Disk, spun up locally: Uses power, processing, and internal network bandwidth. However it lets you run a system as described above where your integrity is guaranteed well because drives can be replaced. Your constant availability also makes it easier to take incremental backups more frequently, which is a useful benefit. Access time is also reduced.

4. Flash, spun down. It's known as very reliable, and restoration is very fast once you've connected it. It costs an arm, a leg, and your favorite coworker, though.

[Charles 9]

Re: Bye-bye tape drives

"Flash, spun down. It's known as very reliable, and restoration is very fast once you've connected it."

Not over the long term. Some of them have been shown to "leak" if left unpowered for a lengthy period. Tape is still unmatched for on-site or otherwise-self-controlled (legally required in some fields) cold storage and archival.

[Anonymous Coward]

Re: Bye-bye tape drives

"You get backup software that supports encrypting and uploading it to various "cloud" storage providers. "

... and when the provider decides your file types aren't supported they are deleted without a warning. too bad. Or go titsup and there goes your backups, all of them.

Cloud is nothing but someone else's computer you have zero control over, not a place for a backup you have to have.

[Uncle Ron]

Every Two Years...

The IT industry has been predicting the death of tape every two years for the last 50 years. But because tape offers characteristics that are not duplicated by any other storage media, it has survived. Some of these characteristics are: 1) Continues to be relatively cheap, 2) Movable from one (broken) machine to another (working) machine, 3) Easily movable offsite, 4) No RAID rebuild delay, which can be MUCH longer than recovery from tape, 5) Can easily and provably be made non re-writable, 6) Can be made archivally (30+ years) permanent . I like tape.

[-v(o.o)v-]

Re: Every Two Years...

Indeed.

There just isn't any technology available that can meet all the use cases where tape is needed currently.

If there was then tape *would* already be dead!

Fact is, tape is not going anywhere until some new breakthrough is found in storage media. 3D storage was mentioned earlier. Cheap NAND flash could *maybe* be an option if adapted to some kind of cartridges.

Durability, transportability, speed, ...

[Charles 9]

Re: Every Two Years...

I don't think NAND can be trusted for archival. Some forms of the tech have been shown to "leak" if left to sit for more than a year or two. As for rust drives, you'd need something with a longevity standard like RDX, which currently tops out at 4TB per unit.

[rcxb]

Re: Every Two Years...

> There just isn't any technology available that can meet all the use cases where tape is needed currently.

No, but the vast majority are better suited by other technologies.

> If there was then tape *would* already be dead!

Even floppy drives are still hanging onto life. Tape sales have been in steep decline for decades now. It's been dying for a long time, and it won't take very much to finally kill it. This current patent squabble certainly has the potential to kick it off. Or the LTO group might learn their lesson and ensure this nonsense NEVER happens again. Then perhaps tape could retain a strong niche for quite some time. But there is no possible future where tape dominates the landscape once again.

[Anonymous Coward]

Re: Every Two Years...

"Tape sales have been in steep decline for decades now"

That's because most people and companies believe they don't need backups as "everything is in the cloud".

Backups also cost money and need a management that can think further in the future than 1) next quartal, 2) next year or 3) next stock option release date.

I'd bet >90% companies fall in any or all 3 categories. That's why.

[Anonymous Coward]

Re: Laws of physics.

Theoretically, any tech that is added elsewhere in the chain (CPU/GPU/HDD/RAM etc) will increase fidelity of manufacturing that can also be applied to tape.

Thus, tape will scale with the rest, just as RAM/HDD/CPUs have scaled accordingly depending on supply, demand and tech (or laws of physics ;) ).

I guess you could even do exotic things such as taking MEMS or Racetrack memory and applying it to miles of length of tape. :P

[Dedobot]

I have 250+ lto4 and 220+ lto6 tapes on the shelf so far. Good luck with flash storage for massive archives.

[l8gravely]

Bwah hah ha!

I've got something like 18,000+ pieces of LTO-4 media kicking around in Iron Mountain. Talk about a horror show to restore! And I'm not even sure if that includes all my old SDLT tapes... must check out that setup as well...

Tapes aren't going anywhere. The other problem is filling tapes fast enough so they keep streaming, and not shoe-shining back and forth. That's what kills performance and tape heads. So yes, you do want disk in front so you can spool there as fast/slow as you like, then rip it off to tape super fast when ready.

[Anonymous Coward]

Off-site backups

At work, I've been sitting near the windows for decades. I get to watch the fluffy clouds drift past, the robins pulling worms from the lawn, and the 'Iron Mountain' company truck (for off-site backups) coming and going.

The truck is getting smaller. It used to be a large truck. Then a smaller cube truck. Then a normal white van. Now it's a Ford Transit Connect, the little tiny cute van based on the Ford 'Fungus' ecobox.

At this rate, it'll soon be a guy on a skateboard.

(I understand that they don't yet trust backing up sensitive data over a wide area network because of security, in spite of the fact that all this sensitive data is already on a wide area network with multiple geographically-distributed locations.)

[HobartTas]

Re: Off-site backups

If you look at this document here https://www.lto.org/wp-content/uploads/2018/03/LTO_Media-Shipment-Report__CY17.pdf you'll see on page 4 that tape capacity shipped is going up every year but at the same time on page 6 the number of cartridges sold is slowly declining and I presume that this is because of individual tape capacity doubling every generation and you might be right about the "guy on a skateboard" if they are carrying 1 PB of data sometime in 2026 stored in five 192TB LTO12 cartridges. https://en.wikipedia.org/wiki/Linear_Tape-Open

[Orv]

Re: Off-site backups

Never underestimate the bandwidth of a guy on a skateboard with a backpack full of tapes. ;)

[PM from Hell]

Re: Off-site backups

The Iron mountain trucks have got smaller, we went from large open reel tapes to 3490 cartridges then LTO but the majority of the space in the truck was boxes and boxes of paper. The amount of paper has reduced radically year on year and may soon disappear altogether.

Even with mirrored data centre running visualised servers constantly snapshotting still like to know I can get back to a known state and tape can be a huge help here. If you run a complex workload with an ERP system at the core and a number of Line of Business systems feeding it you are actually running a very complicated distributed database and managing restore points across all those applications can be tricky. getting back to a know state after a database corruption then being able to control the roll forward gives greater potential for a recovery without significant data loss. It is the nightmare scenario but in health, social care or regulated industries it does need planning for.

[Charles 9]

Though there could reach a point where rust on sticky tape can hit the same problems spinning rust does and physics starts getting in the way. Any info on how close tape is to those physical limitations?

[HobartTas]

Several years ago researchers got to 220TB for BaFe tape https://www.fujifilmusa.com/products/tape_data_storage/innovations/barium_ferrite/index.html and then IBM got to 330TB with sputtered media https://arstechnica.com/information-technology/2017/08/ibm-and-sony-cram-up-to-330tb-into-tiny-tape-cartridge/ and I've also read of Strontium Ferrite as a future replacement for Barrium Ferrite so who knows where the limit is.

[Charles 9]

The article you describe is a couple years old and was all theoreticals, plus some of the tech such as IBM's TMR head is now already on the market. About the only way you can improve this is to use a different media composition, but even the ars article noted only about a 50% improvement over a few years ago. So while there's still headroom, I don't think there's as much headroom as we'd really like given the insatiable hunger for storage.

[carl0s]

I used to think that the name Linear Tape Open meant that it was an open standard unencumbered by patents.

[Charles 9]

Just because a standard is open doesn't mean it is unencumbered. An open standard simply means anyone can sign up and read the specs (meaning it's not kept behind closed doors as trade secrets). Gaining the actual means to produce them (getting past the encumbrances) is another matter.

[Anonymous Coward]

Missed the word OPEN in the format's acronym?

Oh, I do hate Sony. LTO's origins as a cross-manufacturer compatible system to keep tape competitive bred the name Linear Tape Open. I can see this format war killing tape; if it is not quickly closed. I even picked up an older LTO-5 drive for home backup purposes due to mistrust of cloud service reliability. I like having my irreplaceable holiday photos and videos stored on tapes in multiple locations of choice; not to mention audio recording masters of the (dubious) content I've put together over the years in Reason! Having it all on tape that will probably outlive most of the drives intended to read it. Overkill? Probably. But it's the only backup system where I have more or less full control of the whole loop - with the exception of the supply chain for the drives themselves. Fortunately my next upgrade on tape drive is a few years away yet; and even then would be to LTO7 (due to the two-version-step backward compatibility in the standard).

[martinusher]

So what you're really saying....

...is that an internecine dispute between two Japanese manufacturers of a product that's important to customers is messing up those customers. So much for being 'customer centric'.....

[Anonymous Coward]

Why does a US court action stop supply in the Middle East? Article is not clear at all.