Presentations: May 30 and June 4, in class Report Due: Saturday, June 8, 2024 at 6:30PM
You've spent the past 7 weeks striving to replicate a result from a prior publication. Now it's time to report on your results and share them!
Your group will give a 10 minute presentation in class. We'll randomly assign groups to the 2 class meetings. If you want to swap days, you'll need to find a group to swap with. To swap, a member from each group should email the staff list about the swap.
You should expect that a good part of your 10 minutes is questions and discussion. So you should plan on presenting for 5-6 minutes. You can structure your presentation in any way you want. It should provide the following information (recall that most people in the class know nothing about your project!):
What paper did you focus on?
What is the relevant intellectual contribution of this paper that you ended up focusing on?
What is the result or claim you sought to replicate, what does it mean, and why this one?
What did you do to replicate the result?
What were your results?
What do you conclude?
5 minutes is not much time: you should keep the number of slides small. And practice! Since it only takes 5 minutes, the improvement you'll see in your presentation from practice, given how important this is to your grade, is absolutely time well spent. Time yourself.
For example, one structure might be:
Slide 1: Paper and key idea
Slide 2: Graph and methodology
Slide 3: Results
Slide 4: Observations
Your final report is a blog post on Reproducing Networking Research. Read a few prior posts and find some that you think are good work you'd like to emulate; you can use their structure as a starting point for your own post. It should be in the format of posts from 2012 or 2013, such as these
Why Flow Completion Time is the Right Metric for Congestion Control
DCell: A Scalable and Fault-Tolerant Network Structure for Data Centers
Your report should include:
a short introduction that explains the paper, its contributions, and their importance,
the result or claim you chose and why,
the methodology described in the paper,
the methodology you used (and if it diverged, why),
the results you obtained, alongside the original results in a way that lets the reader usefully compare them (e.g. match graph axes if appropriate),
a discussion of your process or details in replicating the result, or a discussion of why you think your results do not agree,
anything else you've learned that you think would be useful to others trying to understand the paper in context.
Please send an email to cs244-spr2324-staff@lists.stanford.edu, with the subject “Team <NAME> Slides”, attaching your slides as a PDF. Send this by the time when your blog post is due.
To create a blog post, follow the instructions on the site; your CAs are the contact for creating an account.