Xingda Wei (魏星达)
Institute of parallel and distributed systems
Shanghai Jiaotong University
Email: wxdwfc (at) sjtu (dot) edu (dot) cn
I am a (tenure track) Assistant Professor in the institute of parallel and distributed systems (IPADS) at Shanghai Jiaotong University (SJTU).
My research focuses on computer systems and distributed systems,
particularly how to design systems with emergency hardware technologies and design systems with machine learning technologies.
My research focuses on building fast and reliable distributed systems. I primarily think about the following question: how can we leverage the heterogeneous hardware features in modern datacenters and powerful machine learning technologies to build the next-generation distributed systems? Specifically, I focus on the following topics:
How to co-design systems with modern datacenter networking? Modern datacenter fast interconnects like RDMA and SmartNIC offer extremely powerful in-networking computing capabilities. How can we fully utilize the power of these networking features? My related projects include:
Machine learning for systems. The advances in machine learning (ML) technologies also bring huge optimization space for computer systems. Besides providing powerful heuristics, the ML methods can change the complexity of traditional algorithms (e.g., sorting and indexing). My related projects include:
Systems with storage class memory. Storage class memory, aka, NVM has finally come to the public market. Besides high performance, its memory-like interfaces give us opportunities to re-think the design of traditional storage systems. My related projects include:
- Fast remote persistent memory: Characterizing and Optimizing Remote Persistent Memory with RDMA and NVM, ATC'21
Next-generation cloud computing infrastructure. Serverless computing is a promising future direction for future cloud computing; it has the benefits of auto-scaling and pay-as-you-go, just to name a few. However, current systems have huge performance costs when supporting the above features. How can we build the next-generation serverless system?
- Fast RDMA-based Ordered Key-Value Store using Remote Learned Cache, Xingda Wei and Rong Chen and Haibo Chen, 2020, In 14th USENIX Symposium on Operating Systems Design and Implementation.
- Deconstructing RDMA-enabled distributed transactions: Hybrid is better! Xingda Wei and Zhiyuan Dong and Rong Chen and Haibo Chen, 2018, In 13th USENIX Symposium on Operating Systems Design and Implementation.
- Fast in-memory transaction processing using RDMA and HTM. Xingda Wei and Jiaxin Shi and Yanzhe Chen and Rong Chen and Haibo Chen, 2015, In Proceedings of the 25th Symposium on Operating Systems Principles.
- Unifying Timestamp with Transaction Ordering for MVCC with Decentralized Scalar Timestamp, Xingda Wei and Rong Chen and Haibo Chen and Zhaoguo Wang and Zhenhan Gong and Binyu Zang, 2021, In 18th USENIX Symposium on Networked Systems Design and Implementation.
- Replication-driven Live Reconfiguration for Fast Distributed Transaction Processing. Xingda Wei and Sijie Shen and Rong Chen and Haibo Chen, 2017, In 2017 USENIX Annual Technical Conference,
- Fast and general distributed transactions using RDMA and HTM. Yanzhe Chen and Xingda Wei and Jiaxin Shi and Rong Chen and Haibo Chen, 2016, In Proceedings of the Eleventh European Conference on Computer Systems.
- Computer system engineering, 2021, instructor
- Advanced distributed system 2020, chief TA
- Advanced distributed system 2019, chief TA