Internet surfing entails exchanging numerous HTTP requests between clients and servers. Attached with each request is a string containing plenty of information about the client called User-Agent string. There have been numerous researches that showed the privacy and security risks of User-Agent strings. In this work we aim to minimize those risks by directing users towards using less exposing browsers. In doing so, we propose and test a new method for calculating an exposure score for web browsers based mainly on their User-Agent strings besides other attributes. We then apply this method on a dataset of over a million browsers to infer and compare their exposure scores. Finally, we present a demo web tool that shows how the resulting scores can be leveraged to aid end users. Our results reveal a variation in the exposure scores of these browsers, which creates prospect for safe Internet surfing.
|Title of host publication||Proceedings of the Future Technologies Conference|
|Editors||K. Arai , S. Kapoor , R. Bhatia |
|Place of Publication||Cham|
|Publication status||Published - 31-Oct-2020|
|Name||Advances in Intelligent Systems and Computing|