Microsoft Research, based in Cambridge, Massachusetts, USA, has begun developing and prototyping next-generation cloud storage systems. Datacenter Dynamics recalls that it is this division that is behind the Project Silica project, within which media were created that could store data for 10 thousand years.
The description of one of the vacancies states that researchers are reconsidering the scope of work, switching from creating media for archival storage to more “hot” solutions. In particular, it is mentioned that recent advances in AI and machine learning have fundamentally changed the scope of what is possible in many areas, including the development of new materials, the creation of scientific simulations, signal processing, etc. The candidate is required to have a PhD in computer science or computer engineering or equivalent experience in R&D in the areas of storage, operating systems, networking, distributed systems, etc.
Image source: Microsoft
New approaches are redefining the capabilities of cloud storage in the era of energy-constrained data centers—the demand for low-cost, resilient data storage has never been greater. As part of the new project, researchers intend to introduce innovations at the level of media and methods of recording and reading information. The project will require “inventing new research methods” to quickly assess how changes in the physical layer will affect key storage performance metrics.
Building an efficient system from scratch also involves solving more traditional problems – assessing workloads, scheduling data to be written and read in an optimal way, efficiently placing information on storage media, and error correction to protect data from errors and erasure. The head of the project is Richard Black, who previously worked on the Pelican initiative – under his leadership the company tried to create the cheapest array of disk drives.