The Future of Data Storage

With the advent of cloud computing, we’ve become accustomed to having access to our online data whenever and wherever we need it. But despite whatever comforting feelings may be imparted by its name, “The Cloud’s” existance is much more ephemeral than it is ethereal. The fact is all the data in the world – from your favorite news site to your online music library – is stored the same way; it’s all imprinted, burned, or stamped into strips or coils of magnetic material. And that material won’t last forever.

It may not be for many years, but alas it is true – one day all our hard drives will fail; our CDs, DVDs, and USB drives will no longer be readable; our cloud data will cease to exist. Data archivists and cloud server farms have protocols in place for preserving data against degradation and the steady march of time, but as humanity continues to generate more and more data at a geometric rate, at some point – maybe sooner than we expect – our data storage demands will outpace our current storage technology. This month, let’s learn about what makes long-term data storage such a challenge, and then see what some research teams are doing about it.

CHALLENGES

Storage Capacity

The staggering amount of data generated by all of humanity each year is measured in zettabytes (one zettabyte equals one billion terrabytes) and continues to grow exponentially. It’s estimated that the storage demands of the entire world will nearly exceed hundreds of zettabytes by the year 2025. Storage solutions of the future need to be able to keep up with this ever-inflating number in such a way that is cost effective and easy to maintain.

Speed

The speed at which we can reliably read and write data also needs to keep pace with the sheer amount of data needing to be read or written. Speed requirements will continue to balloon as average file sizes, network speeds, and potential simultaneous users all increase as well.

Longevity

Hard drives today typically last between 4-6 years, while solid-state drives last a bit longer at 10 years. That’s fine for the average consumer’s storage requirements, but as far as preserving data for future generations that’s not going to cut it. Surprisingly, even the most up-to-date longterm storage solutions currently in use can only expect 15-30 years of reliability, after which point the data needs to be recopied to keep it safe. A primary goal in adopting a future storage solution will be finding a medium that is truly reliable, durable, and long-lasting.

Energy Efficiency

It currently takes a huge amount of energy to store, access, catalog, and maintain large hordes of data. As the world’s power consumption continues to grow wildly out of control, energy efficiency will need to be considered as a key component of any future storage solution.

The Future

So with these challenges to be met in finding a solution, what will the future of data storage look like? Here are a few examples of amazing technologies being researched as we speak that might one day provide an answer:

Storing Data in Glass

One of Microsoft’s many tech research projects is exploring the idea of storing data inside quartz glass. Dubbed “Project Silica”, the idea is that data can be literally etched into a glass platter in layers using a highly sophisticated laser and then read from that point on using normal light. This has several advantages over traditional means of storing data:

1. Quartz glass is low cost and highly durable. It doesn’t degrade over time like the magnetic material used in current storage media, and it’s completely EMF proof.

2. Quartz glass is considered WORM (Write Once, Read Many) media. Once data is written, it is a permanent part of the glass and cannot be erased or tampered with. This is ensured because reading the data is done using regular light which does not have enough power to alter the data.

3. Once data is written, it can sit in an idle state for the entirety of its lifetime using absolutely no power.

4. A single glass platter the size of a DVD can be used to store as much as 7TB of data.

5. Retention lifetime could easily exceed tens to hundreds of thousands of years.

Storing Data in DNA

It may sound like science fiction, but the concept of storing encoded data in a DNA strand has in fact already been proven. DNA is composed of four base ingredients with initials A, C, G, and T which can be arranged into groups of three called “codons” (AGT, CTT, AAG, etc.). Rather than storing data in binary (0’s and 1’s), data is transcribed into codons and then spliced into a synthetic DNA strand. To retrieve the data, the DNA strand is sequenced in a lab to find the inserted codons which are then decoded back into readable form.

The technology is still very much in its early stages, but researchers in different parts of the world are working to improve it such that it may one day become the de facto standard, seeing as it has several immediate advantages:

1. The storage capacity of a 1-gram DNA strand is over 200 petabytes (or 200,000 TB). With such a high storage density, you could potentially store all the digital information in the entire world in a single shoe box.

2. DNA has a half-life of 500 years, which means the data could last for centuries if stored properly in the right environment.

3. If a few centuries isn’t long enough, scientists have already shown that if the DNA is inserted into a bacterium, it can be replicated through the bacterium’s progeny as it reproduces – up to 100 generations without error!

4. Given that certain organisms can thrive in the most hostile environments on the planet (and even in space), it’s conceivable that DNA data could be carried and replicated down through generations that can survive extremes of heat, cold, pressure, and radiation.