Opportunities for power optimizations in residential Internet traffic
This project focuses on exploring the dataset 0 collected by the authors of 1, which captures Internet service usage and traffic traces from a large residential area in Columbia, NY, USA.
Specifically, this project will analyse the temporality and nature of home Internet traffic. It aims to clarify what types of applications are sending/receiving Internet traffic, when, and how much congestion is happening over time.
There are multiple sub-questions to explore, including:
Understanding what class of traffic/applications are making up the mix of residential traffic, and how does that change over the day/week. More specifically, what fraction of the traffic is time sensitive (e.g., linked to live streaming applications) and when does it take place? This will be used to quantify the potential energy savings from sporadically turning off home router hardware, or operating more power proportional hardware.
What is the volume and nature of web traffic? How does the dataset view compares with the view reported by web domain popularity lists? The higher-level objective is to assess the relevance of power/sustainability optimizations suggested by the W3C 2, for example.
How much of the traffic is machine-to-machine, be it either benign (e.g., bots) or malicious (e.g., port scan)?
References
- Off-Campus Residential Traffic From Columbia University, https://wimnet.github.io/CUResidential/
- Shuyue Yu, Thomas Koch, Ilgar Mammadov, Hangpu Cao, Gil Zussman, and Ethan Katz-Bassett. 2025. Internet Service Usage and Delivery As Seen From a Residential Network. Proceedings of the ACM on Measurement and Analysis of Computing Systems 9, 2, Article 41 (June 2025), 30 pages. https://doi.org/10.1145/3727133
- Web Sustainability Guidelines (WSG), https://www.w3.org/TR/web-sustainability-guidelines/
Supervisors
