In this video from the 2016 HPC Advisory Council Switzerland Conference, Miro Lehocky from Seagate presents: .
“As the storage systems used on today’s supercomputers are increasing significantly in size, the issue of power and cooling, reliability and serviceability are becoming a cause for concern. Where storage in the past usually constituted less than 10% of the total power and cooling budgets, projections indicate that these systems may require up to one third of future budgets. Storage solutions need to start behaving like compute where each cycle is important and systems are designed for maximal efficiency. Traditionally, storage have been using brute force rather than intelligent design to deliver the required throughputs but the current trend is to design balanced systems with full utilization of the back-end storage and other related components. These new systems need to use fine grained power control all the way down to individual disk drives as well as tools for continuous monitoring and management of these systems. In addition, the storage solutions of tomorrow needs to support multiple tiers including backend archiving systems supported by HSM as well multiple file systems if required. This presentation is intended to provide a short update of where Seagate HPC storage is today.”