Idle thought: ~30 shipping containers full of 512GB microSD cards could store all of the world's data.https://twitter.com/marcan42/status/1005721534533550081?s=19 …
-
-
Replying to @marcan421 reply 0 retweets 1 like
-
Replying to @carb0nyle
Yes, that's what I was reminded of. I just re-ran the numbers with current microSD card densities.
1 reply 0 retweets 0 likes -
Replying to @marcan42 @carb0nyle
I'm wondering now what the storage density of DNA is.
1 reply 0 retweets 0 likes -
215 petabytes per gram, apparently https://en.m.wikipedia.org/wiki/DNA_digital_data_storage … Crap, someone let the swarm of the internet out again.
1 reply 0 retweets 1 like -
Replying to @benjymous @marcan42
The concept of storing information in DNA is indeed very tempting. The calculation might be correct, however this assumes a single copy of the information. PCR is extremely sensitive but It would be risky to keep only one copy as DNA tend to degrade / hydrolyse
1 reply 0 retweets 0 likes -
Replying to @carb0nyle @benjymous
and even if sequencing cost decreased by order of magnitude (and continue to) is still to expensive for storage (and synthesis is also expensive) but it's still a promising field
1 reply 0 retweets 0 likes -
Replying to @carb0nyle @benjymous
The "single copy" issue is irrelevant. These are all order-of-magnitude estimates anyway. As long as your degradation rate isn't abysmal (e.g. >50% of your data goes away), the overhead of using an appropriate error detection and erasure coding scheme is well below error margins.
3 replies 0 retweets 0 likes -
Replying to @marcan42 @benjymous
It's a very exciting topic and if you know about good papers treating about error correction for DNA storage I'll be interested to read them
1 reply 0 retweets 0 likes
I don't know anything specific to DNA, but in general look up things like Reed-Solomon and Shingled Erasure Codes. These techniques are fairly agnostic to the actual underlying storage mechanism. They're used in everything from CDs to cluster storage to satellite comms.
-
-
This kind of approach is unlikely to evolve in nature, because it's basically complex mathematics, but it's trivial to do once you put a computer in front of the storage. So you don't have to worry too much about the specifities of DNA storage.
1 reply 0 retweets 0 likes -
The same goes for bias. Transmission and storage mechanisms already have this problem, and the solution is whitening, which is basically scrambling the data to be statistically uniform, thus avoiding any excess incidence of biased sequences.
1 reply 0 retweets 0 likes - Show replies
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.