CS 396M Advanced Networking Protocols

Spring 2017   Unique No. 52380

Instructor: Simon S. Lam

The term paper is allocated 35% of your overall course grade of which : (i) 20% is for the written paper submitted, and (ii) 15% is for your presentation and in-class participation during student presentations. 

Attendance. You are required to attend all classes scheduled for student presentations in the latter part of the semester. If you do not attend any of these classes without prior approval by the instructor, 1% of your final course grade will be deducted for each class missed.

Term paper deadlines:

1. February 28, 2017 - Submit at the beginning of class a single page proposal (hard copy) with a tentative title, a list of three or more papers that you will study for writing your term paper, and a summary of less than 250 words.  If you submission is late, 1% of your final course grade will be deducted for each calendar day it is late.

2. April 6, 2017 - Submit at the beginning of class a hard copy of your term paper.  If your submission is late, 1% of final course grade will be deducted for each calendar day it is late. Also email a soft copy (pdf file) of your term paper to me by 10 p.m. for my class folder.


About your term paper

Your term paper will be an individual project.  The maximum length of your paper is 6 pages (single-column) or 3 pages (double-column) using a font size of 11 pt. If you choose the double-column format, it is recommended that you use an ACM or IEEE template.  Each student must write his/her own term paper with his/her own sentences, analytical thinking, and ideas.  Copying and pasting from someone else's paper, published or not, will constitute plagiarism. The use of someone else's phrases or sentences should be attributed and shown within quotation marks.  Also, your paper should be written only for this class and not be used to satisfy the requirement of another past or present class.


You can choose one of two options for your term paper

Option 1 (This is the straightforward case and most students are expected to choose this option)

Begin by looking over the table of contents of the following conference proceedings for the past four years (2013-2016):  SIGCOMM, ICNP, NSDI, IMC, and CoNext.  Select a paper that inspires you to work on a problem of interest to you and relevant to this class.  (This inspiring paper must be from one of the conferences listed above so that your chosen topic is relevant to this class.) From its list of references, or using Google Scholar, find at least two related papers for you to study and understand the topic very well. Some of the related papers can be from other publication venues, including technical reports published by research labs and academic departments.

(You can get free downloads of papers in ACM and IEEE conference proceedings via the UT-Austin Libraries website:  http://www.lib.utexas.edu/indexes/index.php .)

Guidelines - writing a term paper for Option 1

The objective of your writing is to motivate and explain a technical problem and the key ideas of its solution method(s) to other computer scientists who do not know much about the problem.  Your paper should be a narrative in your own words providing insights from reading the selected and related papers and thinking about the issues and ideas therein.  Note that if you only summarize/paraphrase the selected and related papers, your term paper would not receive a very good grade.

You can reuse some illustrations and figures in papers if they are necessary for your narrative.  You must cite the source of each instance of reuse and be very selective.  The following list is intended to help you organize your thinking.  There is no need for you to address every item on the list.

Option 2 (Few students are expected to choose this option)

On your own, read the paper [YL13] in the reading list about a new idea, Atomic Predicates. (The instructor will not cover this paper until the last week of March.) The paper presents a new theory and algorithm for classifying strings of bits into equivalence classes for a given set of filters. Try to think about how this new theory and algorithm can be applied to solve problems in other problem domains, e.g., finding equivalence classes of data in very large datasets in other disciplines to substantially reduce both computation time and/or storage space.  If you can, let your thoughts wander away to problem domains far away from networks (in particular) and computer science (in general).

I have no guideline on how to write this term paper.  However, if you do not have a good application in mind by February 21, play it safe and choose Option 1.