Want your data to live forever? Bin those SSDs, and adopt DNA….

Deoxyribonucleic Acid? Sounds odd….. Yep, 'that' DNA. And yep, it does sound a bit strange. In reality, it's less odd, and more devilishly cunning.


Today's post provides a short "future tech" overview, similar to my Quantum Internet blog from earlier in the year.

I want long term storage, really long term

Great. You can get that today using our old friend the public cloud of course. For instance,

AWS' Glacier is listed as being designed to retain data for "decades". Those wily folk at Azure are even more specific, offering a "99 year" retention policy. Enough to satisfy even the pickiest of compliance officers.

In the real world, the underlying hardware won't last that anywhere near that long. Stone tablets last for millennia, paper might last for 500 years or more, yet poor old digital technologies are unlikely to last for even 50 years. (Obviously the cloud providers above provide technology to replicate and move the data around – abstracting you away from any hardware failures).

There are some legitimate demands for storage that lasts longer than is offered by digital technologies today, for example the UNESCO programme to document the 'memory of the world'. Imagine if we could actually store all current known data.

So, what's the best storage medium found in the world today? DNA, of course. What other medium have you seen store data for over 700,000 years?

But storage is all about 1s and 0s?

Not any more… Now it also includes A, C, T, G. For those that remember their biology from school, these are the core nucleotides which make up DNA – the very fibre of our cloudy beings.

Two digits or 4 characters. Notionally, they're very similar for the purposes of encoding data. And earlier this year a team of smart folk in Switzerland did exactly that:

"Dr Grass said that his team has converted 83 kilobytes of text from the medieval Swiss Federal Charter of 1291 and the Methods of Archimedes from the 10th Century into the digital code of DNA based on sequences of four chemical building blocks, the nucleotides A, C, T and G."

So, there's just DNA roaming loose about the lab? Well, no. This is where is gets clever. In an attempt to preserve the DNA, it's encased in 'glass' (to be precise it's'silica spheres with a diameter of roughly 150 nanometres'). Sadly it's not encased in amber, Jurassic Park fans…..

Any other benefits apart from long term retention?

It's dense. Very, very dense. So dense in fact, one could potentially store as much as 455 Exabytes of data per gram of DNA. For anyone wondering, that could comfortably encode all the words ever spoken on Earth……In one gram.

(If you're baffled by the Exabyte unit of measure, it flows as follows: TeraByte -> PetaByte -> ExaByte. See here for more info.)

Coming soon to PCWorld?

Erm, no.

As is oft the case with emerging tech it's prohibitively expensive – it cost roughly $2,000 for the 83KB above – and at the time of writing is lacking basic storage functions like random access to the data.

This is all possible to overcome though – the Swiss team made a huge breakthrough this year with error correction by adapting familiar maths to this new medium.

Keep Innovating

Regular readers will know one of the primary things I love about public cloud computing is innovation. The rate of change is truly incredible, and you and I can take advantage of that easily via service based delivery.

For me it encapsulates something that's great about humanity: the drive to keep pushing further, to be better tomorrow than you are today.

Storage has been well and truly commoditised by the public cloud. Where else can one look for innovation? Well, thanks to the spectacular work from Dr. Grass & his team, here's one avenue for continued future innovation.

Come on then cloud providers, who's going to be the first to provide long term DNA storage?

Further reading:

Deoxyribonucleic Acid? Sounds odd…..