Dissertation Research: Data-Intensive Computation Offloading
This project provides the first parallel disk system where large parts of data and I/O processing are offloaded to multi-core processors embedded in disk drives. The computation offloading technology is employed to improve parallel disk I/O bandwidth by greatly reducing data movement between disk subsystems and computing hosts. The proposed techniques and mechanisms are highly adaptive to dynamic workloads with both large and small disk requests, making modern parallel disk systems leverage multi-core processors to scale up to arbitrary size.
Research Interests:
parallel and distributed I/O, storage networks, processor embedded active storage, multi-core processors and software techniques for I/O-intensive applications