Xiaodong Zhang (The Ohio State University): Data Management: Interactions with Computer Architecture and Systems

Abstract:  We have entered a data-centric computing era, characterized by the coexistence of diverse parallel and specialized hardware accelerators along with general-purpose processors. In this ecosystem, minimizing data movement has become a critical priority for the design of both systems and applications. Over the years, the CPU-centric ecosystem has evolved into a one-size-fits-all environment, supporting a wide variety of applications. However, its efficiency in performance, computational power, and energy consumption has steadily declined, making the general-purpose computing model increasingly unsustainable for the rapidly growing demand of data analytics and machine learning applications. In this presentation, I will explore the constraints and obstacles inherent in our current computing ecosystem. I will also provide case studies to support the evolution of computer hardware and software for high-performance data processing, featuring advanced hardware components such as GPUs, RDMA, and other relevant technologies. All associated algorithms and software implementations are open source, with some having been integrated into production systems. The system infrastructure transition for data-centric workloads also challenges our college computer science education. In this context, I will briefly introduce a new textbook of mine, which shares the same title of this presentation, and was published by Cambridge University Press last year.

Bio: Xiaodong Zhang is a University Distinguished Scholar and the Robert M. Critchfield Professor in Engineering at the Ohio State University. His research interests focus on data management in computer and distributed systems. Driven by a commitment to translate his academic research solutions into cutting-edge technology, he has made continuous efforts in advancing the design and implementation of several major production systems. He was recognized by the 2020 ACM Microarchitecture Test of Time Award for his contributions on memory architecture design and the 2024 VLDB Test of Time Award for an initial development of open-source spatial data processing systems on large-scale clusters. He received his Ph.D. in Computer Science from University of Colorado at Boulder, where he was honored with a Distinguished Engineering Alumni Award in 2011. He received the Education Leadership Award from the Lutron Foundation for chairing the Department of Computer Science and Engineering at Ohio State from 2006 to 2018. He is a Fellow of the ACM, and a Fellow of the IEEE.