Optical discs are not really meant for longterm storage on their own. Disc rot is a real thing.
Edit: I should have gone into a bit more detail. Yes, optical storage discs CAN be okay for longer term storage. But it depends on quite a few factors. The material itself has to be long lasting, the manufacturer has to have good quality control, and the end client should store it in a controlled environment. It’s it better or worse than alternatives. No idea without the actual data. It certainly has better density.
Another associated benefit would be the minimisation of data migration. The discs are said to be highly stable, with an expected lifespan of 50 to 100 years. That’s a huge leap over current data centre HDD based storage systems, which generally move data over to new devices every five to 10 years to avoid data-loss from ageing drives.
No, not really. Generally as a product ages the quality control goes down since demand isn’t there. You can make archival grade CDs that will last a life time, it just costs too much money for anyone to want to pay for it. Plus business have tape which is plenty good for long term storage.
That makes sense. Why improve the process if it costs more to do so and most people don’t need it to last that long? But at least archival-quality CDs are out there.
SSDs aren’t great for long term archival since the nand needs to periodically be refreshed. You can build a better SSD, but that compromises storage capacity. HDDs are better, but they have other issues from sitting around not being used. Disks like these are a pretty good backup method if produced correctly. If is the big key, 100 layers sounds like a lot of layers to manufacture correctly, and you won’t know your dat is gone until it’s unreadable.
But will it be able to replace tape for long term backup? LTO 9 is supposedly available, and up to 18TB not compressed. LTO-10 is 36, and supposedly LTO-14 is going to be 576 TB but that seems overly ambitious.
At these sizes, you could have one or two error correction layers within the disc to let you read the data through errors. I’d be surprised if that isn’t the case. Sacrificing 1-2% of the storage space for better reliability is an obvious trade-off.
True. It also depends on manufacturer quality controls and end client storage environment. I just recall that being a big selling point for DVD and CDs only for some people to find unreadable garbage after a decade. Or at least not easily readable garbage.
*Some optical discs. Others not so much, it’s not an inevitability… M-DISC, introduced in 2009, has a rating with proper storage of one THOUSAND years. They are even readable and writable by most regular DVD/Blueray drives!
That was my question about this. It can store a lot of data, great! But will the media last 10 years or more? For real long term storage it needs to last decades.
Another associated benefit would be the minimisation of data migration. The discs are said to be highly stable, with an expected lifespan of 50 to 100 years. That’s a huge leap over current data centre HDD based storage systems, which generally move data over to new devices every five to 10 years to avoid data-loss from ageing drives.
Optical discs are not really meant for longterm storage on their own. Disc rot is a real thing.
Edit: I should have gone into a bit more detail. Yes, optical storage discs CAN be okay for longer term storage. But it depends on quite a few factors. The material itself has to be long lasting, the manufacturer has to have good quality control, and the end client should store it in a controlled environment. It’s it better or worse than alternatives. No idea without the actual data. It certainly has better density.
This is also meant to compete with LTO tapes. To my knowledge, the current best is LTO9 with a max uncompressed storage of 18TB per tape.
100-200 TB on a disc would be huge as they could replace 5-11 tapes with one disc and have better random seek times.
Hopefully this does not end up like HVD which was promising but ended up dying due to the initial cost:
https://en.wikipedia.org/wiki/Holographic_Versatile_Disc
There are technologies being developed for data archiving that have densities and longevity orders of magnitude higher than this.
I don’t see this ever leaving the lab.
Wasn’t that something to do with the manufacturing process which has been more recently fixed?
No, not really. Generally as a product ages the quality control goes down since demand isn’t there. You can make archival grade CDs that will last a life time, it just costs too much money for anyone to want to pay for it. Plus business have tape which is plenty good for long term storage.
That makes sense. Why improve the process if it costs more to do so and most people don’t need it to last that long? But at least archival-quality CDs are out there.
I wonder if in the context of storing 200 TB whether the added cost now makes sense given what a comparable SSD or HDD equivalent would run.
SSDs aren’t great for long term archival since the nand needs to periodically be refreshed. You can build a better SSD, but that compromises storage capacity. HDDs are better, but they have other issues from sitting around not being used. Disks like these are a pretty good backup method if produced correctly. If is the big key, 100 layers sounds like a lot of layers to manufacture correctly, and you won’t know your dat is gone until it’s unreadable.
But will it be able to replace tape for long term backup? LTO 9 is supposedly available, and up to 18TB not compressed. LTO-10 is 36, and supposedly LTO-14 is going to be 576 TB but that seems overly ambitious.
At these sizes, you could have one or two error correction layers within the disc to let you read the data through errors. I’d be surprised if that isn’t the case. Sacrificing 1-2% of the storage space for better reliability is an obvious trade-off.
Also LTO is rather expensive, way out of range for the home archivist. Discs tend to be much cheaper! Hopefully this is the case for these as well.
Not sure of that, but there are discs which are made for archival purposes:
https://en.wikipedia.org/wiki/M-DISC#Durability_claims
Optical discs are not one thing, they’re a variety of different technologies. This particular one is predicted to last many decades.
The best medium for long term data retention is etching words in granite.
Stamping glyphs into clay seems to be pretty decent, too. If you don’t have access to granite. Especially in this economy.
Depends on what material they switch to. They could be using some new polymer or something, hence the longer lifespan.
True. It also depends on manufacturer quality controls and end client storage environment. I just recall that being a big selling point for DVD and CDs only for some people to find unreadable garbage after a decade. Or at least not easily readable garbage.
*Some optical discs. Others not so much, it’s not an inevitability… M-DISC, introduced in 2009, has a rating with proper storage of one THOUSAND years. They are even readable and writable by most regular DVD/Blueray drives!
https://en.wikipedia.org/wiki/M-DISC
I absolutely cannot wait for these new discs to be available!
These aren’t going to be leaving the lab, so I wouldn’t get my hopes up.
I wouldn’t be so sure, LTO is big business. A niche market sure, but a multi-million dollar market nonetheless.
That was my question about this. It can store a lot of data, great! But will the media last 10 years or more? For real long term storage it needs to last decades.
From the article:
Thank you! I’ll admit I hadn’t read the article.
You’re welcome.