Biography

I am a final-year Ph.D. student in the Department of Computer Science at Duke University, advised by Prof. Danyang Zhuo. Before Duke, I received my bachelor’s degree from Peking University.

I will join NVIDIA to continue my research on RDMA networks in December 2024.

Research

As a member of Duke Systems Group, my research interests include datacenter networks and AI infrastructure, with a focus on RDMA and Infiniband networks. Specifically, I lead the project of HostBench, a test framework for RDMA/Infiniband software/hardware stack. The ambition of this project is to assist cloud providers and hardware vendors in building more reliable, efficient, and secure next generations of RDMA networks. Our current publications/projects, Husky (NSDI ‘23) and Collie (NSDI ‘22), are part of this initiative. We have already seen real-world impact from our work, as vendors have released new CVEs and firmware upgrades in response to the issues we identified. We are now working on generalizing this framework to large scale RDMA networks and diverse types of heterogeneous hardware to support future AI infrastructure.

Publications

HAL: Hardware-assisted Load Balancing for Energy-efficient SNIC-host Cooperative Computing
Jinghan Huang, Jiaqi Lou, Srikar Vanavasam, Xinhao Kong, Houxiang Ji, Ipoom Jeong, Eun Kyung Lee, Danyang Zhuo, Nam Sung Kim
ISCA 2024
[pdf]

Hardware-assisted RDMA Performance Isolation for Public Clouds
Jiaqi Lou*, Xinhao Kong*, Jinghan Huang, Wei Bai, Nam Sung Kim, Danyang Zhuo (*equal contributions)
NSDI 2024
[pdf] [slides] [video]

Towards a Manageable Intra-Host Network
Xinhao Kong, Jiaqi Lou, Wei Bai, Nam Sung Kim, Danyang Zhuo.
HotOS 2023
[pdf]

Remote Procedure Call as a Managed System Service
Jingrong Chen, Yongji Wu, Shihan Lin, Yechen Xu, Xinhao Kong, Thomas Anderson, Matthew Lentz, Xiaowei Yang, Danyang Zhuo.
NSDI 2023
[pdf] [slides] [code] [video]

Understanding RDMA Microarchitecture Resources for Performance Isolation
Xinhao Kong, Jingrong Chen, Wei Bai, Yechen Xu, Mahmoud Elhaddad, Shachar Raindel, Jitendra Padhye, Alvin R. Lebeck, Danyang Zhuo.
NSDI 2023
[pdf] [slides] [code] [video]

Collie: Finding Performance Anomalies in RDMA Subsystems
Xinhao Kong, Yibo Zhu, Huaping Zhou, Zhuo Jiang, Jianxi Ye, Chuanxiong Guo, Danyang Zhuo.
NSDI 2022
[pdf] [slides] [code] [video]

Invited Talks

Towards Reliable and Predictable RDMA Networks
- NVIDIA Networking Software Architecture Group, March 2024

Towards a Manageable Intra-Host Network
- HotOS 2023, June 2023

Understanding RDMA Microarchitecture Resources for Performance Isolation
- USENIX NSDI 2023, April 2023
- Microsoft Research and Microsoft Azure, August 2022

Collie: Finding Performance Anomalies in RDMA Subsystems
- Harvard Cloud & Network System Group, May 2022
- USENIX NSDI 2022, April 2022
- Student Lightning Talk @Google Networking Research Summit 2022, March 2022
- Microsoft Research and Microsoft Azure, September 2021
- Bytedance System & Networking Group, May 2021

Awards / Honors

Duke CS Best Preliminary Examination Award, 2024
USENIX NSDI Student Grant, 2024
Duke Outstanding Research in Progress Award, 2023
Duke Outstanding Teaching Assistant Award, 2023
USENIX NSDI Student Grant, 2023
Duke Computer Science PhD Fellowship, 2022-2023
USENIX NSDI Student Grant, 2022
Duke Computer Science PhD Fellowship, 2021-2022
Peking University Outstanding Graduate Award, 2020