Michael Alan Chang

I am a Ph.D. student in the CS department at UC Berkeley, where I work with Scott Shenker. I research in the areas of cloud computing, systems for distributed machine learning, and computer networks. In particular, I am interested in making it easy for anyone to deploy, orchestrate, and manage complex distributed services; these services range from web applications to distributed training of neural networks.

Previously, I was an undergraduate at Princeton University, where I worked on SDN, network resiliency, and network analytics with Jennifer Rexford, Laurent Vanbever, and Theo Benson.

I am supported by a NSF GRFP Fellowship.


Contact

EMAIL: machang (at) cs (dot) berkeley (dot) edu

UNDERGRADS

At this point, I am only seeking undergrads (primarily sophomores and juniors) who can commit to research this semester. If you are interested, please complete an application at this link.

INDUSTRY

My research is very focused on finding practical solutions for cloud computing problems. I am constantly seeking to work with companies to identify new research challenges and the practicality of the systems I develop. Please email me if you are interested in chatting more about this.

Publications
and Projects

PUBLICATIONS

Network Evolution for DNNs
Michael Alan Chang, Aurojit Panda, Domenic Bottini, Lisa Jian, Pranay Kumar, Scott Shenker
SysML 2018

Throttlebot: Performance without Insight
Michael Alan Chang, Aurojit Panda, Yuan-Cheng Tsai, Hantao Wang, Scott Shenker
Arxiv Pre-print

Supercharge me: Boost Router Convergence with SDN
[Sigcomm Abstract], [Arxiv Full-length]
Michael Alan Chang, Thomas Holterbach, Markus Happe, Laurent Vanbever

Chaos Monkey: Increasing SDN Reliability through Systematic Network Destruction
Michael Alan Chang, Brendan Tschaen, Theo Benson, Laurent Vanbever
Sigcomm 2015 (Demo)

Destroying Networks for fun (and profit)
Nick Shelly, Brendan Tschaen, Klaus-Tycho Forster, Michael Chang, Theo Benson, Laurent Vanbever
HotNets 2015


Other Projects

Equivalence Class Snapshots in the Data Plane: A Measurement Framework for Network Analysis and Performance Error Diagnosis
Advised by Jennifer Rexford and Srinivas Narayana
Spring 2016

Computation and Design of Locally connected Internet Exhange Points
Advised by Sanjeev Arora and Nick Feamster
Fall 2016

Website Traffic Analysis: Bursting into Burst level countermeasures
Advised by Prateek Mittal
Spring 2014


Issued Patents

Peer-to-peer Home Automation Management
Michael Alan Chang
US 8190275, Filed June 22, 2009

Intelligent Gateway for Heterogeneous Peer-to-peer home automation networks
Michael Alan Chang
US 8086757, Filed March 23, 2010

Undergrads
Advised

Hantao Wang
Class of 2020
Projects: Throttlebot, Verified Triggers
Research Tenure: May 2017 - present

Yuan Cheng (Anson) Tsai
Class of 2020
Projects: Throttlebot, Verified Triggers
Research Tenure: May 2017 - present

Junkeun Yi
Class of 2021
Projects: Verified Triggers
Research Tenure: May 2018 - present

Gavin Kliger
Class of 2021
Projects: Verified Triggers
Research Tenure: May 2018 - present

Emerson Hsieh
Class of 2021
Projects: Verified Triggers
Research Tenure: May 2018 - present

Michelle Hwang
Class of 2020
Projects: Verified Triggers
Research Tenure: December 2017 - June 2018

Amit Talreja
Class of 2020
Projects: Throttlebot
Research Tenure: May 2017 - October 2017

Alex Fang
Class of 2019
Projects: Throttlebot
Research Tenure: October 2017 - present

Domenic Bottini
Class of 2018
Projects: Network Evolution for DNN Training
Research Tenure: November 2017 - present

Lisa Jian
Class of 2018
Projects: Network Evolution for DNN Training
Research Tenure: October 2017 - present

Pranay Kumar
Class of 2018
Projects: Network Evolution for DNN Training
Research Tenure: October 2017 - present