Blind spot - the trouble with optical drives Optical drives should be popular. They hold lots of gigabytes, can be intrinsically WORM and hold data for a guaranteed 30 years or more without the periodic refreshes needed by tape. But they are simply not used as a general data archiving product by business, having become niche products in areas such as medical image …
I don't see the point to this. Optical isn't wonderful, and will soon (hopefully, at least) die out completely - hard drive and flash-based consumer technologies would seem to be a solid indication of this. And, as you say, there are no real, concrete benefits to the move from tape. So yeah, what's the point?
Also, you say "prefessional" movie archive needs.
So uhm, lets's bring back caddies for CDs and DVDs and we are all set. I mean the physical size of the drive is not going to change. And everything that is a mass market now is probably readable in the near future.
There used to be an old caddy standard for CDs you'd just need to produce drives for it. And it would only be a minor change in the drives.
optical media blows for many reasons, so do tapes so do hard drives, so does flash drives
tapes are used because they always have been...
10 years ago tapes were used because they were orders of magnitude cheaper for capacity than alternative technologies e.g. 4x cd burner > £250 and took as long as a tape to write
now hard drives are cheap, but the last 10 years of spam, interoffice porn and possibly business related files are on tape, and the box of 50000 tapes needs to be used
besides tapes are the only thing left that looks remotly mainfraimey (my flashing lights are slowly dissapearing, no big reels of tape, all i got left is a poxy cassete) and the plebs dont have a tape drive on there desktops, so no "wheres the last 5 years of sales data gone?, ummm left it in the laptop i left on the train" scenarios
Proprietary = Doomed
Especially with backup where you know you need long term device support for at least 5-10 years. Affordability is also an issue, lets say my $5-10K device fails, even after its no longer my current backup solution, but now I need to recover something so I have to keep a device working so I have to either buy redundantly and keep an expensive spare or pray that I can still get one several years later.
For this reason I abandoned tape after outgrowing a 1G then 5G then 10G then 20G tape years ago, maybe 1 of which wa backward compatible with the previous ( I think the 10G could read the 5G tapes).
Then I moved to HD-HD backups without removable drive trays to take the disks off site, it was actually faster (no suprise) and cheaper than tape because you could buy an 80G drive for the same money as 2-3 20G tapes. And the HD's don't rip you off by claiming 2:1 compression inherent in their published sizes.
Now I'm waiting for Blueray which will be the equivalent of a 50G tape (using the 2:1 lie!) and will be readable in many devices for no compat problems.
Do unattended HD-HD backup then burn a series of BD disks for offsite as needed.
The tape manufacturers have lost all touch with reality.
So, what is wrong with tape?
Sure, you have to clean the heads on occasion, and the tapes are not cheap, but it is a good standard solution. Tapes are pretty durable, they come in many sizes. The big datacenters have great solutions, tape silos and such, and smaller businesses can buy the cheaper single-tape drives.
Also, since many new tape formats' drives are backward compatable, businesses can keep their old backups without needing to keep a variety of different drives. If a business switched to Blu-ray compatible disks, they would have to discard all of their old stuff or find somewhere to store it. (Yes, I know some new formats are not backward compatible with anything, but they are no less compatible than trying to put an LTO in an optical drive...)
Sure, new companies who are just starting to do backups now could switch to optical, but how is that better than the new guys buying into the same technology as what has been around for a while?
Restores can take for ever if the data is scattered across multiple tapes as in any libtrary, because its sequantial access, all the rest of it. We badly need a decent medium term archiving solution that has reasonable access to the files, good life, easy storage, all the rest of it... The troubl is data storage requirements grow so much faster than the damn MO drives do...
Sure companies have their last 20 years of business data on tapes and need to keep infrastructure around the place to read them if they need it....
But I see more and more companies installing disk libraries... Why? Speed and utility. Their backup windows shrink, sometimes by orders of magnitude. They can hive that data off to tape libraries after the data has passed a certain amount of time without being used...
My company uses disk libraries to keep all our old outlook data on .... there is a delay of about 10 seconds getting data off the disk library when I want to read a really old email.
This article's not about domestic or small-scale backups, it's about enterprise archiving. The main driver for enterprises (and we're talking about backing up 10+TB a night to qualify) is the knowledge that it will work - day in, day out for many, many years. Part of that is the confidence that when one of your STK robots fail, someone will be out in a few hours to fix it. It also means that you are safe in the knowledge that when you need to order another 1000 tapes (at £50 a pop), there will be a ready market to buy them from.
Given these business (note: not technical) requirements, there's really only room for one methodology. It doesn't matter whether it's tape, optical or holographic - so long as it works and the supplier is good enough to provide the ancillary services, which are what REALLY matter. The industry doesn't care about the tekky bits - they just want results. As history would have it, the migration path has decided that starting with 9-track we just happened to end up here. In another generation or two (i.e. 7 or 8 years) we'll have moved on. We can't say what that new something will be - only that there will be one dominant standard and everyone'll use it, not because it's the best but because everyone uses it.
I recently upgraded my old removable IDE caddy with a 500GB hard drive for about £60 and upgraded an old USB HD brick to 500GB in the same way. (In case you're wondering, I've put 350GB of data on them both already, they're my dual independent data store). If I ever get round to fitting an e-SATA adapter to my PC backplate there's a 1TB SATA drive for £97 retail at Saverstore.
If I was buying in bulk from the HD manufacturers, they'd be even cheaper than that and they'll be cheaper still in the future. I can't imagine using optical or tape even for archival backup since HD is so easy to connect into a PC, so easy to buy, only two standards - both well supported (IDE and SATA) and the standard is transparent to OS and application software. It's also fast/random access if you need that.
Clearly you are out of touch with the optical market and performance of storage device per say. A streaming tape drive will outperform a disk drive on pure performance alone. As for Plasmon your observations are correct being a single source technology, but it isn't the only technology. Blu-ray jukeboxes have been taking business from Plasmon for the single source argument. Blu-ray has 50GB on a single side with a roadmap to 200GB. It is being used by schools, MOD, Government, TV companies, book publishers, fire service, photographic agencies etc..
Why is this technology succeeding, because over 180 companies are supporting Blu-ray. You can even archive for 200+ years on Gold disks!
It's not as fast as spinning disk, but that's not the point, it's an archive medium for storing data for 50+ years. It's also a very Green storage technology and incredibly robust.
It take approximately 10 seconds to access a file, which is considerably quicker than having to go through a filing cabinet of archived data.
If you need an archive device with a long life span and proven track record, then a Blu-ray jukebox should be on your list.
The article makes swepping generalizations in error.
We don't actually know optical discs will last for 30 years, nor that any guarantee would mean much if a company just went under. Sure, they claim in some idealized way that the discs "should" last that long, and yes as someone who has lost gigs of data to disc rot when the discs were not ill-treated in any way, I can assure you there's no way we would trust our data stores to such a fragile technology.
Pressed discs, when properly manufactured and handled, are the only ones that are particularly robust over the course of time.
The prior AC is simply wrong about others being out of touch, it is a fool's decision to risk data stores with this unverified tech. Yes it "could" last UP TO many years. Except when any number of things out of your control go wrong. And you won't know until you observe the disc degrading then let's hope it's not bad enough that you can't get the data off to redeposit on another medium.
The key to long lived data stores is not trying to pack bits in the highest density today's tech can barely support. When you have a larger granularity you essentially have redundancy far moreso than just that built into the disc spec.
Remember kids, nobody has had these discs for decades yet. Even the humble 1970's american automobile can last for over 50 years in ideal conditions, but most are a pile of rust by now.
Paris, because once you've lost your 80TB porn stash to a batch of bad bluray discs, you'd need something else to entertain you.
...should be commited to the mighty /dev/null
@Out of touch
If you need an archive device with a long life span and proven track record, then a Blu-ray jukebox should be on your list.
You really seem to be... There's absolutely NOTHING proven about blu ray or about it having a long
life span. The ONLY optical media I'd trust for any long term storage would be CDs.(gasp shock etc...)
Why? because I have 10+ years old CDs that have been badly treated that still work fine and can
still retrieve them fully in this day and age. Also a scratch doesn't nuke a couple of hundred MB(like on DVD) or a couple of GB(on bluray and such).
@Hard drives get bigger and cheaper
It's not a bad idea per-se... But consider this... Unless you have redundancy those two drives will fail
eventually(which kinda doesn't make sense for backups now does it?) And the bigger the drive the
more you will lose in a single failure. I'm getting 2 500gb sata II drives and a sata II hw raid controler...
They'll be mirrored so if one goes I can still retrieve the data. But I also have backups of the most important things and they get offsited to a remote server(we're talking ~5gb) via scp. They are also encrypted so I could even put them onto P2P as omg_big_boobies_naked_chix.zip and I'd have it
distributetly backed up... :)
Anyway whatever one can get from the net is not worth backing up. Everything else... well depends on how much you value your data.
n) Tape technology is primarily driven by the IT industry where reliability, compatibility and long term availability are king. Optical disc technology is primarily driven by the consumer market where product differentiation, speed to market and price are king.
You wouldn't trust your long-term data storage to an essentially consumer orientated medium any more than you'd run your entire enterprise data centre on new-unheard-of-brand-this-week machines bought at 200 quid a pop from PC World.
Optical performance:
Media Load Time 5 sec
Media Unload Time 3 sec
Average Seek Time 35 msec
Buffer Memory 32MB
Max Sustained Transfer Rate - Read 12 MB/s
Max Sustained Transfer Rate - Write 6 MB/s (with verification)
Optical vs. tape performance:
GB <---------------------- Hours ---------------------------->
Tape Read Tape Write Optical Read Optical Write
Time Time Time Time
100 0.2 0.2 4.6 2.3
200 0.5 0.5 9.3 4.6
300 0.7 0.7 13.9 6.9
400 0.9 0.9 18.5 9.3
500 1.2 1.2 23.1 11.6
600 1.4 1.4 27.8 13.9
700 1.6 1.6 32.4 16.2
800 1.9 1.9 37.0 18.5
900 2.1 2.1 41.7 20.8
1,000 2.3 2.3 46.3 23.1
Road map for LTO:
LTO4 (Dec 2006) - 1.6TB (2:1 compression) and data transfer rates of up to 240 MB/second, assuming2:1 compression
LTO5 (Planned) - 3.2 TB (2:1 compression) and data transfer rates of up to 360 MB/second, assuming a 2:1 compression)
LTO6 (Planned) - 6.4 TB (2:1 compression) and data transfer rates of up to 540 MB/second, assuming a 2:1 compression
The total costs of the libraries, including a full set of media, cards (in Euro unfortunately) and the cost per TB is:
Plasmon Optical Libraries Capacity Cost Per TB
24 slot, 1.44TB optical library with one 60GB UDO2 drive, single SCSI Bus 1,440 €4,900.00
24 slot, 1.44TB optical library with two 60GB UDO2 drives, single SCSI Bus 1,440 €6,800.00
32 slot, 1.92TB optical library with one 60GB UDO2 drive, single SCSI Bus 1,920 €4,600.00
32 slot, 1.92TB optical library with two 60GB UDO2 drives, single SCSI Bus 1,920 €6,000.00
80 slot, 4.80TB optical library with two 60GB UDO2 drives, single SCSI Bus 4,800 €3,800.00
174 slot, 10.5TB optical library with two 60GB UDO2 drives, single SCSI Bus 10,440 €3,200.00
238 slot, 14.3TB optical library with four 60GB UDO2 drives, single SCSI Bus 14,280 €4,200.00
238 slot, 14.3TB optical library with six 60GB UDO2 drives, single SCSI Bus 14,280 €4,600.00
238 slot, 14.3TB optical library with eight 60GB UDO2 drives, dual SCSI Bus 14,280 €5,000.00
238 slot, 14.3TB optical library with twelve 60GB UDO2 drives , dual SCSI Bus 14,280 €5,700.00
238 slot, 14.3TB optical library with twelve 60GB UDO2 drives, triple SCSI Bus 14,280 €5,800.00
438 slot, 26.3TB optical library with four 60GB UDO2 drives, single SCSI Bus 26,280 €3,200.00
438 slot, 26.3TB optical library with six 60GB UDO2 drives, single SCSI Bus 26,280 €3,400.00
438 slot, 26.3TB optical library with eight 60GB UDO2 drives, dual SCSI Bus 26,280 €3,600.00
438 slot, 26.3TB optical library with twelve 60GB UDO2 drives , dual SCSI Bus 26,280 €4,100.00
438 slot, 26.3TB optical library with twelve 60GB UDO2 drives, triple SCSI Bus 26,280 €4,100.00
638 slot, 38.3TB optical library with six 60GB UDO2 drives, single SCSI Bus 38,280 €2,900.00
638 slot, 38.3TB optical library with eight 60GB UDO2 drives, dual SCSI Bus 38,280 €3,100.00
638 slot, 38.3TB optical library with twelve 60GB UDO2 drives , dual SCSI Bus 38,280 €3,400.00
638 slot, 38.3TB optical library with twelve 60GB UDO2 drives, triple SCSI Bus 38,280 €3,400.00
The cost per TB of large scale tape libraries is less than €200, again with a full set of media, interface cards, etc.
The cost per TB of disk, including offsite realtime replication, is of the order of optical. So twice the capacity for the same price at greater speed and usability.
Optical is an answer to a question no one is asking.
Sorry for the crappy table formatting but I got tired entering TABs.
Manufacturers were smart enough to develop the 3.5" 'hard' floppy, as opposed to the partly open 5.25" 'soft' floppy. It became a universal hit partly due to being completely encased, no more fingerprints, dust or grit on the media surface.
For some abysmal reason, optical technologies never progressed this far, to a standardized encasement. And has held it back ever since.
And does this insanity date back to the very fact that LP records were not encased, and thus CD's followed this rule? So in effect the current optical technologies are still living the era of open reel tape, never having progressed to the standardized encasement of the C-cassette.
If you can find the capacity of LTO4 tapes, why can't you find the performance spec? I often see LTO4 drives drop out of streaming mode because the disk can't feed the data fast enough. And retrieving a single object from tape is pretty fast as well - it's when you have multiple requests to the same piece of media that things get ugly. As stated, it's archive data, instant access is assumed to be unnecessary, and heavy access should also be assumed to be a non-issue. Otherwise, why are you archiving it???
I just wish the LTO specs were to read all previous generations, rather than just 2 back. No reason I can think of not to have this, and it would make the generation jump a lot easier.
There's no excuse for a storage medium that can be soiled and ruined so easily by fingernails and fingerprints. For this, we can thank the consumer's demand that each new generation be able to play the last's. It's time for an optical medium that is completely encased, like the various Iomega magnetic media we so loved in the early 90s. That would make a worthy successor to the floppy disk.
Didn't panasonic/matsushita already do this or tried some kind of caddy system (still got my 1x panasonic cdrom in the loft somewhere), I'm pretty sure you can buy DVD-RAM discs in caddies too. I reckon its all down to cost cutting, I mean 1 DVD-R disc costs about 12p how much more would it cost if it was incased in a caddy, I'm guessing by atleast 8x, but the difference wouldn't be so much with a 50Gb blueray recordable media probably not even a 20th of the cost of the disc.
Having read the article and associated comments a number of issues have been missed relating to optical. Clearly the author is a Plasmon basher and not aware of optical roadmap.
As an archive medium optical is the only real long term 25+ year choice, not other medium comes close for storing our binary digital data.
Its not fast, but it isn't meant to be if you want something to be read in 25 years do you care if it takes 1 hour or 5ms?
Blu-ray vs the rest. When a company needs to archive data it needs to be written, this process requires a laser to burn a Blu-ray disc which is more permanent than CD or DVD, why? Well a CD or DVD contains a dye which has a laser etch the pits into the surface and we all know what happens to dye over the years (it fades). Blu-ray on the other hand does not use a dye, it uses a crystalline substrate for permanence and crystals last much longer than dye. It is not re-written and is only read when required.
Hard disks are a temporary storage area for holding frequently accessed data. They are not very reliable and after spinning 24x7x365 start to get errors on the surface. Sure we can RAID 5, 6 or dual controller etc.. But I have seen all of this fail due to human error or mechanical failure. They are not rugged enough to be taken offsite and have fragile recording mechanisms.
Tape is a short term archive media 3-5 years. It is designed to take important data offsite for storage and recover a file or server should a disaster happen. If tape is so great, why does every tape manufacturer of a Virtual Tape Library using disk drives?
Solid state disks are new expensive and prone to failure. Just look at the problems the manufacturers are having.
An archiving strategy needs to consider a number of issues:
- Power and cooling for 25 years
- Frequency of access to information
- Reliability of medium
- Industry standards
- Future proof
You should also implement an archiving solution that at anytime makes two copies of any archive data. Be tiered disk - tape - optical depending on time frame and frequency of access.
Any company looking at long term archiving might well choose a Blu-ray jukebox, but should every 3-5 years have a technological refresh to ensure important data can be read and accessed using the current available technology at the time and consider alternatives.
In 5-10 years hard disks will be a thing of the past, we will be using Solid State Discs and optical storage for archiving using holographic technology. Tape will be used to seal boxes.
I've seen over and over these grand "tape vs. optical" discussions. After having lost something around 5Gb of data due to disc rot, I am skeptical of any kind of optical storage. The closest I ever came near something like that would have been the Magneto-Optical band, which was an actually reliable technology. However, my dad was who had the final word back then, and he chose the hideous iomega Jaz drive. Guess who lost a crapload of data; and I wasn't immune to that either. I had a 300Mb backup that my dad decided to move over to one of those cartridges, and all the data went bye-bye. Though most of the stuff I lost was because of those other cartridges, the Iomega zip ones. Remember them???
Meanwhile, the tape camp has gone from DDS-1, DDS-2, DDS-3, DDS-4 up to DAT 72; and all those tape drives are backwards-compatible. While I don't own one myself, I did have access to one, and some of my backups are actually in DDS-4 cartridges. Some of these backups have already outlived my first ill-fated batch of CD's.
SSD's are nice, but they sometimes go crazy, so I'm not quite sure about using them for backup storage. But they are very nice for moving data :)
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
