A Bug’s Life vs. A Byte’s Life: Systems Approaches to Enhance Data Integrity & Usability

Abstract: Data storage is a fundamental need of modern society. Numerous sciences and business rely on data to gain insights and depend on the underlying storage systems to enable data-driven solutions. Nevertheless, as the data volume and system scale keep growing, maintaining data integrity and deriving value from data becomes increasingly challenging. In this talk, I will first introduce our efforts on improving the robustness of Linux file systems, which is the foundation of many practical systems. Next, I will discuss our work on large-scale storage systems that empower top supercomputers for scientific computations. Third, I will present a provenance framework to make data more usable for domain scientists. Finally, I will touch on other on-going projects including a city-scale data infrastructure for wireless and rural America. 

Bio: Mai Zheng is the Harpole Pentair Assistant Professor in the Department of Electrical and Computer Engineering at Iowa State University. He leads the Data Storage Lab to investigate problems of real-world systems that affect data integrity, usability, and other desirable properties. His work appeared at major storage systems venues (e.g., FAST, SYSTOR, TOS, HotStorage), supercomputing venues (e.g., ICS, HPDC, TPDS), and other renowned venues (e.g., OSDI, ASPLOS, USENIX ATC). He received the NSF CAREER award and Best Paper Honorable Mention Awards at FAST and HotStorage. Mai received his PhD degree in computer science from The Ohio State University in 2015. For more information, please visit https://www.ece.iastate.edu/~mai/.