Graduate Communication Networks
The final project will be more extensive and research oriented,
and you will work in groups of 2-4:
You will decide the type and
subject of the project, but for now think in terms of the following:
- Research topic survey . You can survey any of the topics we
covered in class (we only read a few papers in each), or you can
survey a topic that we didn't cover at all but you're interesting in
learning more about (e.g., analytic modeling of TCP/IP performance).
- Reproduce research results in published papers (e.g., using ns).
- Network trace analysis. CAIDA
has an extensive
list of project ideas for network analysis projects. Any one of
these projects would be a great final project for the class. Working
on one these projects has two advantages as well: (1) they already
have the trace data and libraries for manipulating it, so you only
have to spend time implementing the analysis tools, and (2) the CAIDA
folks are a short walk away and incredibly helpful (they want to know
the answers to the questions, too!).
- Implement a protocol or tool (e.g., using Alpine)
- Analyze a content delivery network (e.g. Akamai) as a black box.
The authors of
"The Measured Performance of Content Distribution Networks" were
trying to determine if CDNs are successful in reducing the latency of
downloading a small image. One possible project is to try to replicate
their results. Their work can be extended in several ways -- instead
of looking at latency of small image download you could analyze whether
CDNs do a better job at streaming that the content providers themselves.
- A project of your own devising...
You will write a report describing your project and present it to
the class during the final exam period: