Energy-aware data placement strategy for SSD-assisted streaming video servers

This paper is motivated by the tremendous energy consumption of streaming video servers and the energy-economic nature of solid-state drives. An energy-aware data placement strategy with its associated scheduling policy is proposed for streaming video servers equipped with solid-state drives (SSDs). A skewed energy-aware load over SSDs and disks is adopted with a duplication policy to deal with popular and unpopular video files so that disks could spin down as much as possible. With the concepts of batch processing and patching integrated with the scheduling policy, the proposed strategy could further reduce the bandwidth load and minimize the number of active disks.