General

Instructor:	Chen-Nee Chuah
	3125 Kemper Hall
Lectures:	Mon/Wed 4:10-5:30pm, 1132 Bainer
Office Hours:	Mon/Wed 3-4pm
	or by appointment

Prerequisite

EEC273 or ECS 252 or equivalent

Overview

Internet measurement is a relatively new field, but is playing a key-role in providing crucial, fundamental knowledge about Internet infrastructure (e.g., topologies, routing, etc.), traffic, and performance (e.g., loss, delay, bandwidth, etc.). This course will discuss the design requirement and challenges in measuring and modeling the Internet, the existing measurement methodologies (how/where/when to measure), and the latest measurement results. We will also explore how Internet measurements enable the design and enhancement of a large set of applications, including: optimizing peer-to-peer file-sharing system, capacity planning and traffic engineering, network management and trouble-shooting, detecting network abuse and intrusions, etc. We will study the fundamental algorithmic principles related to these topics, address the limits of algorithmic practicability, and discuss when heuristics should be employed.

Goals

The objective of this course is to expose the students to the state-of-the-art Internet measurements research. The students will learn several statistical tools (e.g., multiple time-scale analysis, linear algebra, stochastic processes) for the analysis and modeling of network measurements, Internet measurement techniques (e.g., active vs. passive measurements), and different software tools (e.g., zebra route listener). Through reading and critique of published papers in this area, the students will learn how to define/formulate a research problem, choose a specific approach, and complete a measurements-related research project.

Course Structures:

The class will NOT be solely lecture-based. Students will be asked to lead discussions on selected papers. Expect a lot of interactions in class. There will be a number of guest lectures given by researchers/experts in the related fields.

Please refer to the tentative lecture schedule for the course syllabus.

Grading

• Assignments/Paper reviews: For paper reviews, you are required to provide a short summary (one or two paragraphs) and a review of the strengths and weaknesses of the paper. The goal of these reviews is to help you synthetise the main ideas and concepts presented in each paper. Assignments may include a) collecting, visualizing, and analyzing empirical network measurement data; b) designing network experiments for evaluation purposes; c) testing hypotheses/inferences statistically; d) modeling system performance and characterizing design tradeoffs.

• Topic Presentation: Students will be asked to lead paper discussions in class. Each student will choose a topic among those listed in the schedule. There will be one or two main paper on that topic that the whole class should read, but to cover that topic well, the student who leads the discussion will typically have to read 5-10 research papers in that specific area. Each student will have to go through the slides with the instructor prior to the corresponding presentation. The length of each presentation is approximately 50-55 minutes, leaving 25-30 minutes for questions and discussions.

• Project: Students work individually or in small groups on a course project. The projects will complement and extend the lecture material. The project may include design, modeling, simulation or performance analysis of a new measurement technique. See the project page for more details.

Textbooks:

None required.

References:

• Selected conference/journal papers and other materials from the References.

Related Conferences:

• ACM Internet Measurement Conference (IMC)
• ACM SIGMETRICS/Performance
• Passive and Active Measurement ocnference: PAM

Acknowledgment