The amount of data that exists in the world is growing at such a rapid rate that it doubles every two years, according to a a 2011 EMC Digital Universe study. It’s an impressive thought, but it’s also quite an alarming one too, because it means we might soon find ourselves in a situation where we’ve got more data than we know what to do with.

A new paper co-authored by the University of Washington and Microsoft Research highlights these concerns, citing a forecast from International Data Corp. (IDC) which predicts the total amount of digital data in the world will exceed 16 zettabytes by the end of next year. “Alarmingly, the exponential [data] growth rate easily exceeds our ability to store it, even when accounting for forecast improvements in storage technologies,” the authors of the paper said.

The problem arises from how we store digital data. The majority of the world’s data currently resides in archival storage, where the most dense medium available is good ol’ fashioned tape. As old as it is, tape offers a maximum density of around 10 GB per cubic millimeter, far greater than flash or disk, which is why it’s still so widely used today.

DNA: The future of storage?

But if EMC’s prediction pans out, it means the world’s data will exceed 32 zettabytes within just three years, and 64 zettabytes within just five years. Clearly, storing that data is going to become a problem, and soon. So what’s to be done?

Luckily for us, there is one storage medium that’s already been proven to work, and not only that, it promises a massive storage density of around one exabyte per cubic millimeter, which is eight orders of magnitude greater than tape. We’re talking about DNA.

DNA storage doesn’t just have the advantage of much greater density though. It also addresses the problem of longevity. Tape has a lifetime of 30 years at the most before its integrity begins to corrode, while spinning disks have a lifetime of just three to five years. In contrast, DNA has a half-life of greater than 500 years, the paper claims.

Of course, people have thought about using DNA as a storage medium before, but big problems need to be overcome before that’s possible. Synthesizing DNA in the first place has already been solved by the biotech industry, but sequencing it to extract the stored data is a much bigger obstacle.

Making DNA storage viable

Now, the paper’s authors have proposed two new ideas that may finally make DNA storage viable, by improving the integrity of stored data from current error rates of about one percent per necleotide, and also by devising a way to access individual bits of stored data in a sequence randomly (current methods involve sequencing the entire DNA pool to be able to access a single byte of data stored within it).

To solve these two problems, the authors propose a DNA storage system architecture that consists of a DNA synthesizer, a DNA sequencer, and a storage container. The DNA synthesizer is needed to encode data to be stored, the DNA sequencer reads DNA sequences to convert and access data, and the container stores the DNA in a pool.

The system solves the problem of redundancy without having any impact on storage density because it offers something called “controllable redundancy” that allows the reliability and density for each type of data to be specified.

As for the problem of random access, the authors suggest this can be solved by using a technique called Polymerase Chain Reaction, which molecular biologists use to isolate specific regions of a DNA sequence. The technique “amplifies” a piece of DNA by repeatedly heating and cooling it, enabling accelerated reads and access to specific data without sequencing the entire DNA pool.

DNA storage is still impractical for now, but the researchers say that with the “impending limit of silicon technology” and the rate at which the biotech industry is advancing DNA sequencing and synthesis, the idea is something computer architects should seriously look into.

Image credit: PublicDomainPictures via pixabay