Cross-Layered OS Design for Managing Heterogeneity in Modern Datacenters

Abstract:

Modern datacenters increasingly rely on heterogeneous architectures to manage data growth, enhance performance, optimize resource utilization, and reduce energy costs. In this talk, I will present our work on operating systems designed to address extreme memory and storage heterogeneity. I will emphasize the importance of a cross-layered OS design philosophy that distributes responsibilities across runtimes, firmware, and hardware controllers. This approach promotes collaborative data processing and enhances scalability, performance, and resource efficiency while maintaining the data reliability and isolation guarantees of traditional monolithic OS designs. To demonstrate the potential of this cross-layered approach, I will discuss its application in managing near-storage accelerators and redesigning OS architectures for storage heterogeneity. Additionally, I will explore cross-layered management of both fast and slow memory technologies. Finally, I will provide a brief overview of our ongoing work in developing cross-layered systems for sustainability and improving natural hazard detection systems.

About the Speaker:

Sudarsun Kannan is an Assistant Professor in the Computer Science Department at Rutgers University. His research focuses on operating system design and its intersection with computer architecture, distributed systems, and high-performance computing (HPC) systems. His work has been published in top venues such as ASPLOS, OSDI, and FAST, and he has received best paper awards at SOSP and ASPLOS, along with the Google Research Scholar award. He co-chaired the HotStorage'22 workshop and serves as an Associate Editor for ACM Transactions on Storage. Before joining Rutgers, he was a postdoctoral research associate at Wisconsin-Madison and graduated with an M.S. and Ph.D. from Georgia Tech.