Measurement, Analysis, and System Implementation of Internet Proxy Servers

Abstract

The rapid growth of the Internet has caused serious security, privacy, and performance issues for users. To mitigate these challenges, proxy servers are widely adopted to provide enhanced protection, anonymity, and bridging. In this thesis, we put our emphasis on the measurement, analysis, and implementation of Internet proxy servers. We collected detailed information on 1681 proxy servers across 90 countries. Our data indicate that most Internet proxy servers are unstable with relatively high latency. Finding and adopting the right proxy servers is no easier than locating a needle in a haystack. To help the users pinpoint their desired proxy servers, we developed a system to collect real-time proxy server data from the Internet. This system will check proxy servers' location, availability, and performance, giving users the most updated information. We also applied a BitTorrent-like protocol to provide better synchronization across the data collection servers. Our evaluation indicates that utilizing a BitTorrent-like protocol on uTorrent to synchronize data between a single host machine and four virtual machines (VMs) improves data synchronization performance. The involvement of multiple peers leads to a reduction in latency during the synchronization process. This reduction in latency can be attributed to the parallelized and efficient distribution of data among peers. Overall, our evaluation demonstrates that leveraging a BitTorrent-like protocol in this setup enhances the data synchronization process, improving latency and performance.