CSR: Medium: Collaborative Research: Workload-Aware Storage Architectures for Optimal Performance and Energy Efficiency

The most significant performance and energy bottlenecks in a computer are often caused by the storage system, because the gap between storage device and CPU speeds is greater than in any other part of the machine. Big data and new storage media only make things worse, because today's systems are still optimized for legacy workloads and hard disks. The team at Stony Brook University, Harvard University, and Harvey Mudd College has shown that large systems are poorly optimized, resulting in waste that increases computing costs, slows scientific progress, and jeopardizes the nation's energy independence. First, the team is examining modern workloads running on a variety of platforms, including individual computers, large compute farms, and a next-generation infrastructure, such as Stony Brook's Reality Deck, a massive gigapixel visualization facility. These workloads produce combined performance and energy traces that are being released to the community. Second, the team is applying techniques such as statistical feature extraction, Hidden Markov Modeling, data-mining, and conditional likelihood maximization to analyze these data sets and traces. The Reality Deck is used to visualize the resulting multi-dimensional performance/energy data sets. The team's analyses reveal fundamental phenomena and principles that inform future designs. Third, the findings from the first two efforts are being combined to develop new storage architectures that best balance performance and energy under different workloads when used with modern devices, such as solid-state drives (SSDs), phase-change memories, etc. The designs leverage the team's work on storage-optimized algorithms, multi-tier storage, and new optimized data structures.