How would molecular storage work?

Disk drives and flash use nanotechnology to build higher density storage. But why not simply encode data directly into molecules? We could cram a billion gigabytes in a single rack. Here's how.

In Advances in Macromolecular Data Storage Masud Mansuripur, a professor at the University of Arizona, proposes a molecular storage system using a tech mashup of gene sequencing, microfluidics, electrophoresis, scanning tunneling microscopes and polymer engineering to create storage of staggering density. The NSA is probably working on this right now.

Read this...

Storage in 2014: An overview

As data usage continues to grow exponentially, IT managers will need to orchestrate multiple kinds of storage — including flash, hard disk and tape — in a way that optimises capacity, performance, cost and power consumption.

Read More

Molecular storage isn't a new idea. Every so often someone announces they've encoded data into, say, DNA as a lab demo, but Prof. Mansuripur has been working on this concept for over a decade and may have a commercial concept.


Unlike current storage, where data locations are fixed, molecular storage has floating data and fixed read/write locations. The molecules are transported around the storage by microfluidics plumbing.

The macromolecules are constructed of 2 or more bases, strung together in long - ≈1cm - sequences. The datacules - my coinage - are created at a write station and then moved to parking spots before being summoned to read stations.


The hardware - R/W stations, plumbing, parking spots - would be built into a 10 micron thick device that could be stacked to achieve density. The datacules would be moved electrically through electrophoretic transfer.

Atomically sharp needles of a scanning tunneling microscope write and read the datacules. Starting with a blank macromolecule made up of base A, the STM modifies the bases to encode binary data.

The resulting datacule is read by passing it through a nanopore membrane read head. This idea is borrowed from the technology of gene sequencing, whose success is a wonder of our age. You read the molecule as you would a strand of DNA.

The address lines tell the system where the molecules are parked. The microfluidics system moves the chosen molecule to the read/write station.  

Here's a diagram from the paper:

Screen Shot 2014-09-28 at 10.30.06 PM


The Storage Bits take

This is a bulk storage concept. It won't be fast enough to run databases. But there's a huge market for archive storage that this technology could serve.

Read this

Make your cloud safer: How to enable two-factor authentication for the most popular cloud services

Step-by-step instructions to help you tighten security and dramatically reduce the risk that crucial cloud services will be compromised. If you use a Microsoft or Google account, Office 365, Dropbox, Facebook, or Twitter, keep reading.

Read More

The pieces needed to realize this concept exist today. But integrating the technology from such diverse fields will require a deep and broad technical team and a gigabucket of money. But what if we could store an exabyte in a single rack, at low cost and faster-than-tape access times?

The normal 5-5-5 development cycle - 5 years from theory to lab demo; 5 more years to prototype; and another 5 years to reach the market - may be optimistic. But who doubts that the cloud suppliers would snap it up?

If drive vendors can't figure out how to mass produce patterned media and heat-assisted recording heads, perhaps molecular storage will take us to the next level in mass storage.

Courteous comments welcome, of course. Using multiple chemical bases would allow much denser encoding. Hexadecimal storage, anyone?