Using a secure element to protect the users’ profiles generated by web search engines

Web search engines (WSEs) are valuable tools that are widely used to find specific information in the World Wide Web. Recently, they have increased search result relevance by personalizing them according to the users’ interests. Nevertheless, WSEs also pose an important privacy threat, as they profile users by storing and analyzing their previous search data. To address this privacy problem, current solutions propose new mechanisms that add a significant computation and communication overhead, and/or lack personalized search results. In this paper we present a server-side web search model that serves personalized search results while preserving the privacy of the users. In this model, a mechanism called the secure element (SE) acts as an intermediary between the web search engine and end users. The secure element forwards queries from the users to the WSE and later re-ranks search results according to the user’s previous search behavior. All communication between the users and the secure element is encrypted to prevent eavesdropping. A privacy analysis shows that the scheme effectively protects users from being profiled by WSEs or external attackers.