CSE 561: Graduate Computer Networking

Spring 2001

Instructor:Venkat Padmanabhan, Microsoft Research
(To email me use padmanab@cs.washington.edu, not my Microsoft address!)
TA: Andrew Whitaker


Announcements

May 11: The mid-term exam is scheduled for 3:30-5:30 PM on May 17 in Sieg 134.
Apr 16: I have swapped the readings scheduled for April 20 and 23. We will discuss router mechanisms on April 20 and turn to applications on April 23.
Mar 24: The class mailing list is now operational. Please add yourself to the list by following these instructions. Also, note that I've made a few changes to the reading list and review schedule.
Mar 22: No lecture on Monday, March 26 (I'll be out of town). The first lecture will be on Friday, March 30.


Administrivia

Lectures: MF 12:00-1:20PM, EE1 003
Office hours: MF 1:30-2:00 PM, Sieg 325E (as needed)
TA office hours: Th 2:30-3:30 PM, Sieg 433
Optional textbook: Computer Networks: A Systems Approach (Second Edition), Larry Peterson and Bruce Davie
Class mailing list: instructions, archive
Reading list
Lecture notes


Overview

The growth of the Internet over the past couple of decades has been nothing short of phenomenal. Today people are even talking of extending the Internet to "wire up" motes of "smart" dust! In this graduate course, we will focus on understanding the fundamental design principles that have contributed to the Internet's remarkable scalability and robustness. We will also discuss the challenges facing the Internet today and consider interesting new research directions.

The course has two main components. The first component involves reading and discussing 25-30 research papers, including classic papers written by some of the early designers of the Internet and the inventors of technologies such as the Ethernet. We will devote several lectures to emerging areas such as wireless and mobile networking. The goal is both to come up to speed on the networking research literature and to learn how to critique research work (think of yourself as a future conference program committee member!). The second component of the class involves defining and working on a networking research project, preferably in small groups. The goal is to do high-quality work that could eventually lead to a publication at a top networking conference.

Prerequisites: We will assume that students have a basic knowledge of computer networking. We will not cover basic material in the lectures. If you have taken an undergraduate networking class (e.g., CSE 461 at UW or a similar course elsewhere), then you're all set. If you haven't, please come talk to me. There are several excellent and easy-to-read networking textbooks that you could read to come up to speed. (The book we'd recommend is the one by Peterson and Davie listed below.) Networking is a very broad field. There is opportunity to do an interesting and relevant research project whether your background is in operating systems or in graph theory!


Grading



Reading List and Reviews

The reading list for the course includes 26 "core" papers. You are expected to have read the core papers for each lecture before the lecture. Also, you'll need to turn in a review for each paper marked with a "review due" tag. In addition, we've listed several "optional" papers that are relevant to the lecture topics. You are not required to either read or review these papers. But if you are interested in learning more about a specific topic, you would want to read them.

The focus of the review should be critiquing rather than summarizing the paper. What we'd expect in a review is an overview of the paper in 2-3 sentences and followed by what you think the main contributions and key shortcomings of the paper are (a bulletized list would be perfectly fine). You may get bonus points for contrasting the paper with other related ones from the reading list or elsewhere. Your review for a paper should not exceed one page (and can certainly be shorter than a full page!). It is fine to discuss the paper with other students, but you must write your review individually. You'll need to turn in a hardcopy of your review at the start of each lecture. You can choose to post your review to the class mailing list after the lecture to get some discussion going.


Mid-term Exam

The mid-term exam is scheduled for 3:30-5:30 PM on May 17 in Sieg 134. This will be an open-book exam -- you are free to bring along any books, papers or notes that you'd like. The exam will include a mix of design questions, which will ask you to propose solutions to networking design problems, and shorter questions, which will ask you to comment on design decisions made in the papers that we discussed in class or in related systems in existence. Here are a couple of sample questions, one of each kind.

Project

The goal is to define and work on a research project that is challenging yet doable in the short timeframe of the course. It would be ideal to work in groups of 2 or 3 students, but that's not a requirement. We would encourage you to think of project ideas on your own, but we'll distribute a list of possible projects in class in case you are looking for ideas. I'll also be meeting with each group individually several times during the quarter to help guide the projects.

If you are onto something interesting, it would certainly be an option (and one that we would definitely encourage) to continue working on it beyond the quarter and possibly publishing the work at a top conference. But remember, we can only evaluate you on what you accomplish during the quarter! Plan your project schedule accordingly.