Systems Seminar

EPFL IC Systems Seminar

Building Low-latency and Scalable Computing Systems Using Non-Volatile Memory



Abstract

The recent availability of Intel/Micron’s 3D Xpoint to the market marks the transition of non-volatile memory (NVM) from research artifact to real product. While NVM shares similarities like persistency and byte-addressability with DRAM, it has a different performance characteristic with DRAM and changes the persistency boundary of computer memory hierarchy. In this talk, I will first present a characterization on the performance of NVM and then presents a series of work that exploits NVM to build low-latency and scalable computing systems, including file systems, Java virtual machines and transaction-processing systems. Finally, I will discuss the challenges and opportunities ahead and outline some future directions.

Bio

Haibo Chen is a Professor at the School of Software, Shanghai Jiao Tong University, where he co-founds and leads the Institute of Parallel and Distributed Systems (IPADS). He is currently on sabbatical at Huawei, where he founds and directs the OS Kernel Lab and serves as the Chief Scientist for OS. Haibo’s main research interests are building scalable and dependable systems software, by leveraging cross-layering approaches spanning computer hardware, system virtualization and operating systems. His research work has been widely adopted by open-source community like Linux and Open JDK as well as commercial products deployed over hundreds of billions of devices, resulted in numerous publications in top systems conference like SOSP/OSDI, and won several best papers from EuroSys, APSys and ICPP. He currently chairs ACM SIGOPS ChinaSys, serves on the program committees of SOSP 2019, CCS 2019, IEEE S&P 2020, ASPLOS 2020 and the editorial boards of ACM Transactions on Storage and Communications of the ACM.