DNA Storage Edges Closer To Production Readiness
Posted on Wednesday, March 27, 2019
DNA storage has been covered on the CDS blog several times in the past because it presents a potential solution to the problem of storage demand versus capacity. In 2016, scientists made their first successful tests using DNA as a storage medium.
A year later and Microsoft was doubling down, purchasing a further 10 million strands of synthetic DNA from Twist Bioscience. But now in 2019, Microsoft’s genetic storage project has moved up a gear as they unveil their own system for synthesizing DNA.
An important milestone
In the past DNA storage has been lab-based, requiring specialist genetic engineers and equipment to perform the necessary encode/decode functions. Data has been successfully stored and retrieved from DNA strands, but the process has been slow and manually intensive.
This week Microsoft has unveiled the world’s first automated DNA read and write platform, dramatically reducing the complexity and cost of synthesizing genetic material for data storage. The system is still proof-of-concept, but it brings the possibility of a DNA-powered data storage platform even closer.
The new automated system is still relatively slow, taking nearly 21 hours to encode and decode a five-letter word. It has also failed to reach the 1GB storage densities achieved in previous tests using lab-synthesized DNA strands, storing nothing more than the text string “hello” for demonstration purposes. But this new automated system remains an important development, not least because it helps to reduce the overall cost of using genetic storage.
According to Microsoft, the entire automated encode/decode system costs just $10,000. Researchers believe that this cost will continue to fall, reaching around $3,500 once existing sensors and actuators are refined or removed. This would make Microsoft’s desktop synthesizer considerably cheaper than a traditional all-flash array offering similar (theoretical) capacities.
For now, the biggest challenge remains latency. Until scientists are able to speed up the synthetization process, DNA is unlikely to be suitable for any business deployment, apart from long-term archiving, perhaps.
We will continue to watch Microsoft’s progress with interest.
Which Server Components Are Most Likely To Fail?
Meeting stringent uptime SLAs requires planning for component failures– so which are most likely to need replacement?
3 Steps to Truly Secure Data Destruction
Secure data destruction from old disk arrays is not only good business practice, it is also a regulatory requirement. Here’s how to ensure unwanted data cannot be recovered.