Project
The project is a central component of CS 445. Your goal is to pose and investigate a research question that requires Internet-scale experimentation or measurement.
This can involve:
- Extending an existing system or dataset
- Designing a new measurement
- Running active experiments
- Using natural experiments or quasi-experimental design
- Characterizing an infrastructure component (DNS, CDN, ASes, IXPs, cables, outages, etc.)
- Applying measurement to understand an Internet phenomenon
Projects are done in groups of 2–3 students.
Deliverables & Timeline
Week 1 — Form a group
Look for teammates with complementary skills (measurement, data analysis, systems, etc.).
Week 2 — Meet with instructor
Short discussion of your topic direction, feasibility, and data needs.
Week 3 — Project Proposal (Canvas)
Your proposal should answer the core questions in the Heilmeier Catechism:
- What problem are you solving?
- Why is it important?
- What’s your approach?
- What will you measure or build?
- How will you evaluate success?
- What are the risks and unknowns?
- What are the milestones?
Week 5 — Midterm Presentation & Report
4 slides:
- Project name & team members
- Revised goals + key concepts
- Issues addressed & issues pending
- Updated milestones (with timeline)
Short midterm writeup (1–2 pages) explaining progress and risks.
Last Class — Final Presentation
10–12 minute talk + Q&A.
Finals Week — Final Report
Due Monday of finals week.
Final Report Format
Format follows the HotNets style (6 pages, references excluded).
Template: https://conferences.sigcomm.org/hotnets/2023/hotnets23-template.zip
Recommended structure:
- Abstract — What you did and what you learned
- Problem Statement — What question you investigated
- Prior Work — What’s been done and why it isn’t enough
- Approach — Measurement/experimental design, datasets, instrumentation
- Results — Findings, figures, analysis
- Lessons Learned — Surprises, mistakes, limitations
- Future Work — What remains open
- Conclusions
Appendices or external webpages may be used to document details.
Resources & Data Sources
Some examples — you are not limited to these:
- RIPE Atlas — global active measurement platform
- ZMap ecosystem — scanning, probing, hitlists
- Public DNS resolvers and ECS behavior
- CDN request/response measurements
- Traceroutes + submarine cable inference (Calypso)
- Historical outages and censorship events
- AS-level topology (CAIDA, RouteViews, RIPE RIS)
- IP geolocation datasets (IPinfo, MaxMind, iGDB)
If your project needs access to restricted datasets (e.g., M-Lab, APNIC, private logs) talk with the instructor early.
Expectations
A good project will:
- Ask a question clearly grounded in the Internet
- Use a method appropriate to real-world constraints
- Demonstrate creativity in design or analysis
- Provide reproducibility as far as possible
- Produce takeaways beyond a narrow dataset
A great project will:
- Identify something that matters to Internet operators, researchers, or users
- Develop a framework others could reuse
- Produce at least one non-trivial insight supported by evidence