What If Big Data Is Too Big? A Radical Solution May Be In DNA

[ErikTheNorwegian]

Big

[dr.diesel]

can last indefinitely.

Very dense yes, but not forever.  DNA has a known half life, around 500 years IIRC.

[dr.diesel]

Where as a CD/DVD appears to have a life time limit of a year at times it seems

Nature knows best, a hell of a lot more than we ever have.

Optical media is manufactured at the absolute lowest price point, they all suck and are never properly stored.  I've been using spinning drives for years now as backups, swapping them out for newer every few.

Nature is very clever but not perfect, we can learn from both sides for-sure.

[ovnr]

Yes, DNA may have splendid information density, but retrieving the data is a pain in the ass.

Also, the article is horribly bad (surprise!). They've stored 700 TB? Maybe, if you count "replicating the same 10M basepair sequence over and over" as a valid way of doing it. Also, making it about books is hilarious. Assuming 500 kB/book (books are compressible you know - most epub books I've got around are around 500 kB, incl. cover image), the whole 70 million book archive would take a staggering... 33 TB. Oh lol. I've got room for 17 TB on my boxes (incl. RAID6 redundancy on most of it), and with today's disks you could store it in a single desktop computer (9 4TB drives). Not exactly difficult, at least compared to digitizing 70 million books...


I'll just chalk this one down to "reporter hears about revolutionary new storage method, writes glowing article about it, nothing further happens, and the world moves on". Remember holographic disks (HVD)?

[dr.diesel]

DNA can be sett up to replicate, and if its damaged, it can be correted/restored.. so a type of CRC is buildt inn..

True, but then you have to wrap it up in a living machine.  Then since it's organic and living you have to add an immune system to fight off the endless supply to stuff trying to eat it.  All of this ultra complexity for it to just replicate wrong 70 years down the road and get cancer.

[amyk]

AFAIK the "bit error rate" of DNA is extremely high, not so much an issue in living things where it's frequently rebuilt but for data storage and archival as the article suggests, I wouldn't trust its robustness. NAND flash is already bad enough...

[dr.diesel]

No, sorry, DNA used like this, dont need a living machine to be stored, just in a nautral substance that could be almost anny kind of material.

If not kept living, then you're back to the half life problem.  As stated this is a horrible data storage method, it will never be used outside of normal life.

[dr.diesel]

Never say never, they ruled out both the car and trains as usefull for any transort, and IBM said that 64K with memory would do for anything that we could have use for in the future.. 

I can't argue with that my friend. 

However, life, DNA, etc is likely just too fragile, if you really need your message to last etch it in Bismuth-209, which has a half life of 4.6×10^19 years.