About Weitao

Feel free to contact me at anytime for any reason

我的中文名叫王伟涛,是一名美国莱斯大学的在读博士生。我的导师是T. S. Eugene Ng教授。同时我也是谷歌公司的学生研究员。

我目前的研究方向主要在设计应用与系统相结合的高性能计算平台,包括拥塞控制算法,云应用规划,可编程硬件,异质性系统,分布式系统,数据中心网络,以及专注于优化人工智能应用的系统。

I will be on academic job market from Oct 2023.

学术论文

你也可以查看我的谷歌学术个人页面.

  • Augmented Queue: A Scalable In-Network Abstraction for Data Center Network Sharing
  • Poseidon: Efficient, Robust, and Practical Datacenter CC via Deployable INT
    • Weitao Wang, Masoud Moshref, Yuliang Li, Gautam Kumar, T. S. Eugene Ng, Neal Cardwell, Nandita Dukkipati
    • NSDI 2023 [pdf] [code] [slides] [video]
    • Typo: Page 6: “The larger m is, the slower the rate of update will be” -> “The smaller m is, the slower the rate of update will be”
    • Typo: Page 7: “the target delay is D.” -> “the observed delay is D.”
  • RDC: Relieving Data Center Network Congestion with Topological Reconfigurability at the Edge
  • SpiderMon: Harnessing Wait-For Relations for Performance Debugging with Programmable Switches
  • Shufflecast: An Optical, Data-rate Agnostic and Low-Power Multicast Architecture for Next-Generation Compute Clusters
    • Sushovan Das, Afsaneh Rahbar, Xinyu Crystal Wu, Zhuang Wang, Weitao Wang, Ang Chen, T. S. Eugene Ng
    • ToN 2022 [pdf]
  • MXDAG: A Hybrid Abstraction for Cluster Applications
  • Abstractions for Reconfigurable Hybrid Network Update and A Consistent Update Approach
  • Towards All-optical Circuit-switched Network Cores: The Case for Mitigating Traffic Skewness at the Edge
  • Grasp the Root Causes in the Data Plane: Diagnosing Latency Problems with SpiderMon
  • Say No to Rack Boundaries: Towards A Reconfigurable Pod-Centric DCN Architecture

Contact

Email: wtwang AT rice.edu

Hi, this is Weitao Wang. I am a 5th-year Ph.D. student at Rice University, advised by Prof. T. S. Eugene Ng. I also work closely with Prof. Ang Chen, and a group of great researchmates: Sushovan Das, Xinyu Crystal Wu, and Zhuang Wang. I am also a student researcher at Google NetInfra.
    My newest research focuses on decentralizing centralized algorithms. This is a unique research direction, making use of the telemetry information from a distributed environment to achieve global optimal or pareto optimal state, which previously can only be approached by centralized algorithms.
    I grew up in Anhui, China. And before joining Rice, I received my B.S. degree in Computer Science from Shanghai Jiao Tong Univeristy in 2018.
    Full CV is here.

I am currently on academic job market.

Publications

You can also browse my Google Scholar profile.

  • Aurelia: CXL Fabric with Tentacle
  • Augmented Queue: A Scalable In-Network Abstraction for Data Center Network Sharing
  • Poseidon: Efficient, Robust, and Practical Datacenter CC via Deployable INT
    • Weitao Wang, Masoud Moshref, Yuliang Li, Gautam Kumar, T. S. Eugene Ng, Neal Cardwell, Nandita Dukkipati
    • NSDI 2023 [pdf] [code] [slides] [video]
    • Typo: Page 6: “The larger m is, the slower the rate of update will be” -> “The smaller m is, the slower the rate of update will be”
    • Typo: Page 7: “the target delay is D.” -> “the observed delay is D.”
  • RDC: Relieving Data Center Network Congestion with Topological Reconfigurability at the Edge
  • SpiderMon: Harnessing Wait-For Relations for Performance Debugging with Programmable Switches
  • Shufflecast: An Optical, Data-rate Agnostic and Low-Power Multicast Architecture for Next-Generation Compute Clusters
    • Sushovan Das, Afsaneh Rahbar, Xinyu Crystal Wu, Zhuang Wang, Weitao Wang, Ang Chen, T. S. Eugene Ng
    • ToN 2022 [pdf]
  • MXDAG: A Hybrid Abstraction for Cluster Applications
  • Abstractions for Reconfigurable Hybrid Network Update and A Consistent Update Approach
  • Towards All-optical Circuit-switched Network Cores: The Case for Mitigating Traffic Skewness at the Edge
  • Grasp the Root Causes in the Data Plane: Diagnosing Latency Problems with SpiderMon
  • Say No to Rack Boundaries: Towards A Reconfigurable Pod-Centric DCN Architecture

Contact

Email: wtwang AT rice.edu
You can also leave a message in the following Disqus.