This project targets the theoretical foundations of the next generation of large scale distributed storage systems, which are currently undergoing explosive growth because they form a critical component of cloud computing. This growth has been driven primarily by a system-building approach, and the theory is underdeveloped and lagging behind.
This project develops the mathematical foundations as well as practical engineering constructions related to meeting the data-integrity challenges of next-generation distributed storage systems, and focuses on: (i) Energy-aware Distributed Storage Network Codes aimed at optimizing system energy tradeoffs across the dimensions of storage cost, network bandwidth, disk access cost, and encoding/decoding complexity; (ii) Secure Distributed Storage Capacity with regard to both active and passive attacks; and (iii) Latency-minimized Distributed Storage Systems that minimize user-experienced latency through codes by exploring the complex interplay between coding and queuing delay.
