Zadok and Kaufman Receive NSF Award: Storage Architecture Improvements on the Way


Stony Brook, NY- January, 2014

Computer Science Associate Professor and lead Principal Investigator Erez Zadok and Distinguished Computer Science Professor and Chairman Arie Kaufman at Stony Brook University, along with Harvard University and Harvey Mudd College, received a million dollar National Science Foundation (NSF) Collaborative Research Award for their proposal, Workload-Aware Storage Architectures for Optimal Performance and Energy Efficiency.

This collaborative research project is based on the theory that the most significant performance and energy bottlenecks in a computer are caused by the storage system and are largely due to the gap between storage device and CPU speeds. Big data and new storage media worsen the situation because today's systems are still optimized for legacy workloads and hard disks. The teams at Stony Brook University, Harvard University, and Harvey Mudd College have shown that large systems are poorly optimized, resulting in waste that increases computing costs, slows scientific progress, and jeopardizes the nation's energy independence.

“Large computing systems are expanding rapidly to process the ever growing data-set sizes, but they are also consuming more energy at alarming rates, often due to delays in the data storage subsystem. Working with top researchers nationwide, we hope to make one more critical step towards our nation’s energy independence and advancing the U.S. scientific capabilities,” said Erez Zadok.

With developing optimal performance and energy efficiency solutions top-of-mind, during the first phase of this three-year project, the team will examine modern workloads running on a variety of platforms, including individual computers, large compute farms, and a next-generation infrastructure, such as Stony Brook University's Reality Deck, a massive gigapixel visualization facility.

Second, the team will apply statistical feature extraction, Hidden Markov Modeling, data-mining, and conditional likelihood maximization to analyze datasets and traces. The Reality Deck will be used to visualize the resulting multi-dimensional performance/energy datasets. The team's analyses will reveal fundamental phenomena and principles that inform future storage architecture designs.

The findings from the first two efforts will be used to develop new storage architectures that balance performance and energy under different workloads when used with modern devices, such as solid-state drives (SSDs) and phase-change memories. The designs will leverage the team's work on storage-optimized algorithms, multi-tier storage, and new optimized data structures.

“Working alongside Margo Seltzer at Harvard and Geoff Kuenning at Harvey Mudd to develop solutions to storage architecture challenges is an opportunity to share resources and to collaborate with exceptional faculty, where the whole is greater than the sum of its parts. We are thrilled that the importance of this research is recognized by NSF with a joint grant of one million dollars,” said Arie Kaufman.

About Computer Science Department at Stony Brook University
The Department of Computer Science at SBU is ranked among the top 20 computer science departments in the nation by the National Research Council. Interdisciplinary collaboration and research recognition elevated the Department to be the largest and one of the strongest unit at SBU. The Department boasts internationally renowned faculty who have made significant contributions in computer systems, visual computing, cybersecurity, networking, intelligent computing, algorithms, and verification.