Crowdsourced Measurements of Internet Traffic Manipulation
Event Type
DescriptionIn its half century lifespan, the internet has revolutionized politics, economics, social structures, and information as we know it. Various bodies have interests in controlling data flow, such as governments, internet service providers, content distribution networks, and content providers. These entities enforce their interests with tactics that have potential to affect efficiency and user experience. We seek to determine the control tactics implemented on real networks and how these tactics affect the user.

No testbed could match the medley of paradigms present in consumer networks, and so it is from the public internet from which we seek to gather data. We propose to do so by distributing experiment software via a crowdsourcing platform and then analyzing the measurements generated from the participants’ computers. Ethical consideration has been given to the valuation of participants’ labor, liability when searching for certain content, and privacy.

We are currently in the process of building the experiment software. The software will record some details about the participant’s computer, such as the capacity of their network interface card, the AS’ executor, the connection provider, and the average bandwidth speeds available to the participant. It will then run a body of network tests against domains of varying popularity and contentiousness to survey metrics such as request round-trip time, network hosts traversed in transit, and transmission capacity. These data will be analyzed for patterns in relation to the nature of the queried domain’s content, business relationships between connection providers and AS executors, and more.