Course Project Requirements

The course project in CPSC 641 requires students to demonstrate an in-depth understanding of some important aspect of high-speed networks and/or network protocol performance. Your project accounts for 50% of your final grade for the course, and thus should represent 4-6 weeks of thoughtful independent research effort. Projects are to be done individually.

The project should take the form of a research paper, similar to those found in the published networking literature. (Of course, there is no requirement that your paper be publishable!) The research paper should present your own (novel) research results on a relevant network performance problem. Results may be obtained analytically, through simulation, or experimentally through measurement of an existing system or implementation. The paper should be 12-15 pages in length, including abstract, figures, tables, and bibliography. Use a reasonable word processing package, a readable font size, and single-column formatting.

Some background reading is always required to do a good research paper on a given topic, and should be reflected in your bibliography. Note, however, that it is NOT sufficient to do just a survey paper. Creativity, originality, and your own contribution are also required. This may involve applying previously known approaches to new network scenarios, or applying new approaches to previously studied scenarios, or new approaches to new problems entirely.

Deadlines

The timetable for project requirements is as follows:

• by Thursday, February 28 (4:00pm): Hand in a one-page project proposal in hardcopy form. The proposal should clearly identify the topic being addressed and your proposed approach to the problem. This proposal will be my record of what you are working on. Informal feedback will be provided on the scope and suitability of your proposed project, though the project proposal itself will not be marked. Deviations from the proposed project at a later stage are still possible, if discussed with me first.
• by Friday, April 12 (4:00pm): Hand in the completed research paper to me in hardcopy form. Projects involving significant implementation effort can be accompanied by a scheduled demo, if appropriate.

Project Ideas

Project topics are to be mutually agreed upon between you and me. Topics of your own choosing are preferred, but if you need ideas, I can offer several suggestions, and point you to relevant literature.

Several possible ideas for project topics are listed below:

• anomaly detection
• DASH-based video streaming
• D2L session characterization and modeling
• energy-aware networking
• extremely parallel data transfers
• longitudinal analysis of email traffic
• malware attacks (DoS, DNS)
• modern Web page structure
• multi-path TCP
• Netflix CDN design/optimization
• network traffic visualization
• Office 365 email traffic characterization
• online social network analysis
• QoS scheduling in cellular networks
• redundant traffic elimination
• stretched exponential models
• traffic matrix estimation
• Web server access log analysis
• wireless sensor networks

Here are examples of actual project titles from students in this course over the past few offerings:

• Bottleneck Analysis in Multi-core Web Servers
• Broadcast Schemes for Video-on-Demand
• Cloud Access Using Smartphones: A Performance Study
• Fairness in Speed Scaling Systems
• Online Advertising Systems
• Performance of High Speed TCP Variants
• Power-Aware Clustering in Wireless Sensor Networks
• Quality of Service Support for WLAN Traffic
• TCP Performance over ZigBee Networks
• TPC-W Benchmarking of Multi-core Web Servers
• Virtual Key Rings in Wireless Sensor Networks

In any event, put your thinking caps on, and discuss your idea(s) with me as soon as possible.