Reproducing Network Research
by Lisa Yan and Nick McKeown
With help from Sachin Katti, Keith Winstein, Nikhil Handigol, Brandon Heller, and Bob Lantz
This document gives a suggested format for a graduate networking class project on reproducing network research. More resources are available on the sidebar.
Feel free to make improvements, suggestions, and edits to this wiki page as you try out this project in your own classroom.
Why this project?
There is a wide range of graduate networking class projects, and this is just one of them. This project is one way to give students the experience of doing networking research within the time and resource constraints of the classroom.
Reproducing Network Research is a project that follows a similar approach to introductory science in high school and university: reinforce classroom learning by repeating well-known experiments in the lab. Our goal is for students to obtain a detailed, in-depth understanding of a significant paper, its key ideas, and its key results.
Most of the learning experience comes from the process of recreating and setting up the experiment. We have observed that students learn a huge amount when their experiments yield different results from the original research: they must discern if there are unstated assumptions in or inaccuracies in their own results or the published results. Furthermore, students who spend a lot of time studying and repeating a published experiment tend to ask meta questions: Why did the researchers pose such a problem? Why did they use or build a particular prototype or simulator? Why did they collect a specific set of results? All of these questions contribute to a fundamentally deeper exploration of the research process than the experience of simply reading the paper.
You can read more about our experiences in our CCR paper, which was awarded Best of CCR at SIGCOMM 2017.
Read the assignment page
To get started, go to the assignment page or click on any link in the sidebar.
This project is open to everyone.