Feeds

back to article Hitachi claims glass data storage will last millions of years

Hitachi is showing off a storage system using quartz glass that it claims will retain data for hundred of millions of years. Company researchers displayed the storage unit, consisting of a sliver of glass 2cm square and 2mm thick, which can hold 40MB of data per square inch, about the same as a standard CD. The data is written …

COMMENTS

This topic is closed for new posts.

Page:

Silver badge
Mushroom

a cover showing Earth's location

Helpful for the Vogons when they plan their newest hyperspace bypass.

6
0
Bronze badge

I think they dropped one.

Not as if I'm throwing stones.

1
0
Silver badge

Prior art

I'm sure I remember seeing this in the superman films.

3
0
Bronze badge
Coat

HAL's memory modules?

See title

3
0

This post has been deleted by a moderator

Silver badge
Alert

While the data may still be readable for that long, one has to wonder if the encoding will last as long. After all, reafing a binary 65 won't mean much if the ASCII code system got lost along the way.

1
0
Anonymous Coward

ASCII code system got lost along the way...

The system used will be chosen for being straightforward, one assumes. Even if it isn't, future archaeologists should be able to decipher it by looking for patterns in the data.

It seems to me that the harder challenge is flagging these plates as actually containing information. You can hardly expect READ.ME to stay unchanged for millennia.

0
0
Silver badge

You're forgetting the power of backward compatibility !

If the DoD was involved we would still be publishing data in Trilobite compatible formats

0
0
Bronze badge
Boffin

A nice shiny thing for a pendant... (Re: ASCII code system got lost along the way...)

" flagging these plates as actually containing information"

Yes, that is probably the biggest problem if the technical knowledge falls down to medieval level even briefly. These, like the surviving CD/DVD/Blueray disks make good material for jewellery... and magnetic tapes would get woven into fabrics.

But if one is optimist, the readme problem could be solved by carving a binary->ASCII table on stones left near the archives of these glass records. These last thousands of years unless deliberately destroyed. A small pictorial dictionary could also be provided the same way, to help make some sense of the words after some microscope-wielding future archeologist has managed to transcribe dots into letters.

0
0
Bronze badge

Re: A nice shiny thing for a pendant... (ASCII code system got lost along the way...)

"and magnetic tapes would get woven into fabrics."

Not the softest underwear....

0
0
Happy

Re: A nice shiny thing for a pendant... (ASCII code system got lost along the way...)

lol @ SG reference.

0
0
Silver badge

Re: A nice shiny thing for a pendant... (ASCII code system got lost along the way...)

>But if one is optimist, the readme problem could be solved by carving a binary->ASCII table on stones left near the archives of these glass records

A good start, but I can't help feeling that any storage site would be missed, for the million other monuments to our vanity that we have built. Why look for glass sheets when you have the remains of cathedrals, shopping malls and power stations to pick through?

Arthur C Clarke's answer was to have a technological hurdle- 'you've got to make it too the moon'. We might consider parts of Antarctica, too, such as those extremely dry valleys if not the actual South Pole. Left there would be instructions for finding more accessible, larger and duplicated data stashes elsewhere in the world. If you advertise these stashes too readily, you run the risk of 'grave robbers' (or cultural revolutionaries!) getting there before scholars. Perhaps mixing the plates with radioactive waste might be a way of both advertising them, and deterring the wrong people at the same time.

In reality, I think the applications are going to be a bit more domestic- people today looking to leave family records and photos to their grandchildren. You could also imagine these integrated into gravestones, so that they can be read with a portable device once the moss has been wiped off. If we can the data to last a few hundred years, later generations can work on it lasting thousands.

1
1
Anonymous Coward

re: binary 65

It's not so difficult. In the case where the data are intended to be read many (hundreds of) years down the road, say to be deposited in a time capsule, then people can design the encoding to be as readable as possible. There are some standard tricks that can be borrowed from basic data transmission protocols: stop bits frame bytes (letters) and perhaps rudimentary blocks with checksumming. A document designed for long-term storage would also have some preamble material which can be used as a guide for understanding the rest of the document. Since you mention ASCII, we could include pictorial representations of the letters alongside the binary-encoded forms, for example. Similarly, a rudimentary vocabulary might also be included, perhaps also augmented with pictorial forms. Or perhaps the document would eschew English entirely and use a set of hieroglyphics (encoded in binary from the included pictorial dictionary). One fascinating possibility, I think, would be to encode information at different scales. At the highest level, a pictorial representation of a lens could direct the savvy reader on how to construct a tool for reading the rest of the text, for example.

Even if the document wasn't created with the intention for it being read many years hence, if it encodes language strings (rather than just program data, say), then there are standard analytical techniques that can be brought to bear. Provided the future reader has a modicum of mathematical ability and (presumably) the ability to transcribe data off the glass, given enough text it should be possible to discern lexical features (basically letters and words), syntactic elements (distinguish verbs, nouns, tenses, pronouns, etc.) and from there start working on the semantics of the message. I think that the task of finding a message that's worth reading would likely be much more difficult than the task of decoding a useful message, should it be found.

Overall, I wouldn't worry about ASCII (or EBCDIC or whatever) being a barrier to future understanding. There are two much bigger problems: how to ensure that we have a worthwhile message to transmit (and that it can be found) and that there will actually be anyone around to read it. On that scale, worrying about ASCII is pointless.

Bonus link: http://en.wikipedia.org/wiki/Into_Eternity_(film)

0
0
WTF?

retain data for hundred of millions of years...

How on earth do they test something like that? It sounds good but that is about as far as it goes.

1
0
Silver badge

Re: retain data for hundred of millions of years...

This isn't a new material they have developed, its 'just' quartz. As such, it is the same material that has passed details of the dinosaurs down to us... hundreds of millions of years after the event.

Plastic optical media is untested over mere decades, magnetic media too easily corrupted by magnets and in any case gets eaten by mould.

http://en.wikipedia.org/wiki/Fused_quartz

1
0
Boffin

Re: retain data for hundred of millions of years...

The speed with which quartz (which is SiO2) will diffuse and degrade is entirely predictable. Furthermore, the relationship of that speed to the absolute temperature is also known. That is why they raised the temperature to 1000C (1273K) for two hours. That two hours at that temperature caused the same level of degradation as an enormous amount of time at room temperature.

2
0

Inconceivable?

You keep using that word. I don't think it means what you think it means.

I'd hate to be the backup guy, having to swap out 2cm^2 disks that hold 40MB a piece.

5
0
Devil

Re: Inconceivable?

ouch, ah, ouch, sh*t, aaarraaaghhh.

0
0
Anonymous Coward

Re: Inconceivable?

Inconceivable = my daughter, until I gave her mother the test tube and syringe!

0
0

Won't last 100's of millions of years

Glass crystallizes. It's called "devitrification" It's commonly seen in obsidian which is volcanic glass. Or to put it another way, it just wants to be a crystal again. Glass is a metastable state.

It will outlast plastic media by thousands of years. But 100's of millions is fluff.

0
7

Re: Won't last 100's of millions of years

http://en.wikipedia.org/wiki/Devitrification

Appears to be clear glass by the photos, which is the least likely to devitrify.

http://en.wikipedia.org/wiki/Fused_quartz

It sounds like this stuff is pretty damn stable. Unless it gets buried under immense pressure, or ground back in to sand, it's apt to be around for some time.

2
0
Silver badge

Re: Won't last 100's of millions of years

"Glass crystallizes. It's called "devitrification" It's commonly seen in obsidian which is volcanic glass. Or to put it another way, it just wants to be a crystal again. Glass is a metastable state."

*jaw drops*

Did you really think that was something these researchers didn't know?

2
0
Silver badge

quartz isn't silica glass

I suspect they could also do the same thing in transparent sapphire (which is quite cheap)

2
0
Trollface

Re: quartz isn't silica glass

Superman's ancestors stored data that way!...

1
0
Silver badge

Re: quartz isn't silica glass

>I suspect they could also do the same thing in transparent sapphire (which is quite cheap)

I smell a new gimmick for overpriced wristwatches coming on.... "You never actually own a Patek Philippe.... you merely look after it for the next generation." 40Mb is plenty to record a family tree and a few photos, the data stored on the watch's crystal.

0
0
Coat

meltdown...

i remember reading once that glass is really a liquid...so your data would just turn into a puddle...that couldn't be good!...

0
2
Anonymous Coward

Re: meltdown...

That's a myth.

I believe the origin of the myth is old plate glass windows which tend to be thicker at the bottom than the top but this is an effect of the manufacturing process rather than because it has "flowed" over the years. There are some cases where the glass has been installed upside down and the thicker edge is actually at the top.

4
0
Silver badge

Re: meltdown...

Somewhat a myth but with basis in reality. Glass is considered an amorphous solid. IOW, it has no real structure to its solid form, such that it closely resembles its fluid form. So unlike crystalline solids like quartz and sapphire, it IS more vulnerable to damage and distortion under the right circumstances, given its inherently low strength and toughness.

0
0

praos

Flowing of glass is a myth. Glass crystallizes, by only at high temps, if not of special composition (enamels). Natural glasses (obsidian, tectites) last for ages. Of all glasses the fused quartz is the most stable, due to its high softening point; in fact the main role of all other components in glass is to depress its softening point to workable temperatures. Don't worry, quartz if for ever.

0
0

millions of years data storage

of course will be used for porn.

if humans evolve enough, will it be illegal to watch porn of humans from millions of years ago? what if man evolves an extra penis or something? then all porn of men with one penis will be extreme porn and jacqui smith will have to make it illegal. thats what i thought anyway.

1
1
Silver badge

Re: millions of years data storage

Well, would you want to watch Neanderthal pr0n?

Hint: They didn't look like Raquel Welch.

0
0
Anonymous Coward

Re: millions of years data storage

What if you're into Hirsute porn?

0
0
Go

Massive storage

When is this going to be available in 10ZB (Zettabyte)? I need more space.

0
0
Bronze badge

I remember when...

you had to use a chisel on a rock.

It seems to me that the longest surviving storage medium to date is DNA. The best way to keep data alive is to keep reproducing it with error correction.

4
0
Silver badge

Re: I remember when...

"I remember when... you had to use a chisel on a rock."

Ahh yes, debugging Fortran was always fun.

6
0
Silver badge
Coat

Re: I remember when...

You had chisels! I used to dream of chisels.

Back in my day we used to code by hitting rocks with other rocks. And we were *grateful*!

4
0
Silver badge

Re: I remember when...

It has been said that all the things around us now are "either good at not being destroyed, or good at being made", with rocks and DNA being examples of each.

0
0
Silver badge

Re: I remember when...

There's just one problem with DNA: Evolution - which means the "error correction" you refer to has a way of completely altering the data over multiple generations.

See, if I consider my DNA as information handed down to me by my father, who got it from his father, who got it from his father, and so on back to the Cretaceous, for example, then we find that the DNA is describing something along the lines of a cynodont, which is a rather different creature to a human being. It certainly isn't the same information by any stretch, despite said cynodont being my own direct-line male ancestor. (Although I find the idea of calling a tiny ratlike creature "great-great-something granddad" rather amusing myself!)

Life isn't a good preserver of information. It seems to me the ancients got it right the first time around - carving things in stone seems to be the way to go!

2
0

Re: I remember when...

You had rocks to hit other rocks ?

In my day we had to chew the rocks to inscribe letters and counted ourselves lucky if we had any left after the first sentence !

2
0
Boffin

Re: I remember when...

That's not necessarily right, it seems that DNA could be a reliable carrier of info.

Paul Davies suggested, quote from a brief Google search, "that some sections of junk DNA seem to be markedly resistant to change, and have remained identical in humans, rats, mice, chickens and dogs for at least 300 million years". Biologists like Richard Dawkins seem to agree this is generally plausible.

People overestimate how well artificial things last. There was a TV series about what would happen to the world if every human on the planet dropped dead, which drove home the point in rather dramatic visual style. The fact is it wouldn't take nature long to wipe out all traces of us.

Stuff like this glass data - buried, crushed, eroded to sand, in a geological instant. Even if you tried to protect it with sophisticated bunkers, or put it in orbit, or on the moon - you're still only just delaying the inevitable.

But if life evolves, always driven to survive as it is, then it is actually the ideal vehicle of legacy. Information could be genetically engineered into us, or other life forms whose evolutionary trees are judged more likely to continue. Bacteria for example.

The likes of Paul Davies suggest that it could be a likely place to look for messages encoded for *us*, by past visiting ETs for example (and that this is actually a more likely scenario than scanning the heavens for radio waves which seem less and less likely to be out there, even if the galaxy is relatively abundant with life). If you are ET and you visit earth, the chances are astronomically against it happening to be the year 2012. If you arrived 300 million years ago, how would you communicate with any yet-to-evolve intelligent species? Maybe leave a message that couldn't disappear or be easily vandalized, and the reading and understanding of which would depend on highly developed intelligence and technology? How would you do that?

1
0
Bronze badge
Boffin

Messages in a cosmic bottle

I think it would be easier for space-faring ETs to carve their messages on the Moon. Say in 100 meter wide symbols, with enough redundancy that a few meteor hits would not destroy all of them. Perhaps scrawl a few copies of the same text on different lunar "mares". They probably would have the science to compute how much duplication and how wide and deep the engraving has to be to ensure a good probablility of survival for the 300 million years.

0
0
Big Brother

Re: Messages in a cosmic bottle

Or a big black obelisk, of course.

It may be that any dependency on geology is unreliable. After all the moon had a rough time of it. As an alien futurologist you may consider it unpredictable that even massive earthworks on an apparently stable surface could really endure. It may also have to be considered conceivable that intelligent life may never develop an interest in the moon's surface and never bother to go find your message, whereas DNA encoding would almost certainly cross the radar at some point as the medical benefits of the research would all lead to it.

Finally, because of the time scale, you have to consider all sorts of whacked out scenarios - I intentionally mentioned vandalism. A message on the moon will only be discovered by a select few, by those with the technological power to do it. You may have to think about the idea that an authoritarian power of your intended recipients may not like the idea of said message, and not only keep it from their population, but destroy it, meaning that no future civilizations will find it either. Or, a third party alien group may come across it, and also destroy it for whatever reason. But, something that is an integral part of every life form on the planet forever is a little harder to snuff out. Sooner or later, whether the next generation or the next species, somebody else will have another opportunity.

0
0
Bronze badge
Happy

Re: Messages in a cosmic bottle

>"A message on the moon will only be discovered by a select few, by those with the technological power to do it."

Well, make it large enough to be seen by a good telescope, of the kind that could have been built in the 19. century. The hypothetical authoritarian regime would have a hard time keeping that hidden (in one country maybe, but not all over the globe).

0
0
Gold badge
Boffin

No mention of data transfer rate.

Reading this stuff could take a while.

And no mention of the actual *process* that puts these bits into the quartz in the first place. " lasering dots on the glass" is frankly pretty vague. Is it *very* near the surface (they say they can do 4 layers so far) or can it fill the *whole* volume, which would give a pretty good packing density. There are a lot of layers with say a 10micrometre spacing between them.

BTW "glass" specifically means a non-crystalline super cooled fluid with no long range order (like most thermoplastics, but not thermo setting resins like Bakelite) . Silica, SiO2 is *very* crystalline and large crystals of it are routinely grown in high pressure (c400atm) water..It might have the *appearance* of glass, but it's not. It *could* be sputtered onto a flexible tape, which would be an interesting storage medium. That would not be crystalline. A reel of quartz tape (no plastic layers) could be flexible and have a long life time (But probably not a 100 million years).

0
0
Anonymous Coward

Re: No mention of data transfer rate.

"Fused quartz or fused silica is glass containing high-purity silica in amorphous (non-crystalline) form. It is different from traditional glasses, in that there are no ingredients added to lower the melt temperature."

They didn't mention how the made the dots... as they were talking about layers, I assumed they were using two lasers whose combined power at their intersection point was enough to mark the silica glass internally, stereolithography.

0
0

Reminds me of 'On the Beach'

This has a disquieting similarity to Neville Shute's chilling book "On the beach" (written circa 1950) which describes a world where mankind slowly exterminates itself by unleashing a global nuclear war. The last remaining country (Australia) encases pages of the Encyclopedia Britannica in glass, presumably in the hope it will survive till another human species appears and can read the assembled knowledge. I guess 100M years might be enough.

0
0
Facepalm

You're doing this on purpose aren't you

". . . 2cm square and 2mm thick . . ." immediately followed with ". . . 40MB of data per square inch . . ."

1
0
Anonymous Coward

well.. a 100Million years from now..

.. they'd see my cache of porn, delicately saved.. for future reference of course.

0
0
Silver badge
Headmaster

Re: well.. a 100Million years from now..

While "cache" is technically correct, the preferred term is "stash".

0
0
Silver badge
Headmaster

Re: well.. a 100Million years from now..

A porn cache is a subset of porn from your porn stash that you keep on hand for low latency access.

4
0

Page:

This topic is closed for new posts.