"Farewell to Servers: Hardware, Software, and Network Approaches towards Datacenter Resource Disaggregation

"Farewell to Servers: Hardware, Software, and Network Approaches towards Datacenter Resource Disaggregation
Monday, October 15, 2018 - 3:00pm
Gates 463A
Yiying Zhang (Purdue University)
Abstract / Description: 

Datacenters have been using a "monolithic" server model for decades, where each server has a motherboard that hosts a set of hardware devices such as processors and memory chips. This monolithic architecture is easy to deploy but cannot offer efficient resource packing, hardware elasticity, failure isolation, or good support for heterogeneity. Going forward, I believe that we should rethink the decade-long server-centric model.

Our answer is to break monolithic servers into distributed, network-attached hardware components that can each manage its own resource and can fail independently. For the past three years, my lab has been working on such datacenter "resource disaggregation" at system software, networking, and hardware levels. In this talk, I will discuss challenges and our efforts in building a disaggregated datacenter. Specifically, I will focus on two systems: LegoOS, a new distributed operating system designed for datacenter resource disaggregation (OSDI'18 best paper), and LITE, a kernel indirection tier that improves RDMA scalability and usability (SOSP'17).


Yiying Zhang is an assistant professor in the School of Electrical and Computer Engineering at Purdue University. Her research interests span operating systems, distributed systems, datacenter networking, computer architecture, and security, with a focus on building software, hardware, and networking systems for next-generation datacenters. Her lab is among the few groups in the world that build new OSes and full-stack, cross-layer systems. She won the best paper award at OSDI'18 and has published at and served on the program committees of top systems conferences such as SOSP, OSDI, and ASPLOS. Her work has also attracted various industry and academia attentions. Yiying received her Ph.D. from the Department of Computer Sciences at the University of Wisconsin-Madison under the supervision of Andrea and Remzi Arpaci-Dusseau and worked as a postdoctoral scholar at the University of California, San Diego before joining Purdue.