DNA: Nature's data center
- By Matt Leonard
- Feb 28, 2018
Data centers are booming right now, gearing up to handle compute, network and storage needs for a more connected world. But they already have large footprints, and as the collection of data continues to accelerate, those facilities will have to expand as well. So the Intelligence Advanced Research Projects Activity is looking for storage media that take up significantly less space than traditional technologies.
The fix, the agency thinks, could be nature’s preferred storage solution: DNA.
IARPA is looking for molecular information storage technologies (MIST) that can hold massive amounts of data within a small footprint for long periods without degradation, and that is far enough along in the research process to be commercially viable within 10 years. DNA could check all those boxes, according to a broad agency announcement released this week.
Molecular information storage has been a theoretical idea for decades, but it has hit some important milestones recently. In 2016, researchers at University of Washington and Microsoft Research were able to encode four digital images onto DNA and then successfully retrieve them without any data loss.
If these molecular storage methods are perfected, they could deliver information storage densities 10 million times greater than current techniques. They also could keep data intact for hundreds of years, greatly reducing operation and maintenance costs, according to IARPA.
IARPA met with stakeholders throughout 2016 and 2017, but now it wants to make the technology a reality, so it’s planning on spinning up the MIST program.
“The fundamental aim of the MIST program is to develop deployable storage technologies that can eventually scale into the exabyte regime and beyond with reduced physical footprint, power and cost requirements relative to conventional storage technologies,” the solicitation reads.
The agency is looking for solutions from a number of fields -- including chemistry, molecular biology and microfluidics -- that could help bring MIST to life. Its goal is to be able to store and retrieve data from a MIST storage solution and have an operating system that helps facilitate the process.
The MIST program, which will consist of two two-year phases, is expected to begin later this year with a two-day kick-off workshop in the Washington, D.C., area.
Responses to the solicitation are due by March 15.
Read the full BAA here.
Matt Leonard is a former reporter for GCN.