Metamorphic and Analytical Guessing of Internet Communications (MAGIC)

Despite several years of research towards improving user experience with the Web, users remain dissatisfied. Although, Content Providers make contracts with Content Delivery Networks (CDNs) and Online Advertisement providers (OAPs) to speed up webpage load times and improve relevancy of content being shown the website, respectively; however both CDNs and OAPs remain unaware as to what users might be interested in requesting next on the Web. Therefore, CDNs often strive to prefetch data which prevents further reductions in webpage load times, and OAPs display advertisements based on stale an incomplete Web browsing histories collected for a given user. We introduce Metamorphic and Analytical Guessing of Internet Communications (MAGIC) , a technique to perform a thorough search of website relationships using DNS logs. MAGIC identifies websites that are requested more often in a particular order, thus allowing CDNs and OAPs to accurately predict users' next browsing interests based on which website they have previously visited. MAGIC is effective even when the user leaves a website or installs a browser plugin to prevent websites from tracking browsing history. We make MAGIC available to the community as an online tool to calculate popularity of a given list of website names, using the DNS logs collected from a large university campus network.