Trust-Based and Privacy-Preserving Fine-Grained Data Retrieval Scheme For MSNs Enahoro Oriero ∗ , Khaled Rabieh ∗ , Mohamed Mahmoud ∗ , Muhammad Ismail † , Erchin Serpedin ‡ , and Khalid Qaraqe † ∗ Department of Electrical & Computer Engineering, Tennessee Tech University, Cookeville, TN, USA † Electrical and Computer Engineering, Texas A&M University at Qatar, Doha, Qatar ‡ Electrical and Computer Engineering, Texas A&M University, College Station, USA Abstract—In this paper, we propose a trust-based and privacy- preserving fine-grained data retrieval scheme for mobile social networks (MSNs). The scheme enables users to create a log of trusted users who store (or are interested in) some topics related to a subject of interest. A subject is a broad term that can cover many fine-grained topics. In creating logs, we leverage friends-of- friends relationships and transferrable trust concept. Each user trusts its friends and the friends of friends. If a friend is not interested in a subject, he can help his friend in creating the log by linking the friend to his friends without knowing the subject to preserve privacy. In order to reduce the storage and computation overhead, we use Bloom filters to store the topics. A distinctive feature in our scheme is that it can query users who possess a fine-grained topic, rather than querying users who are interested in the broad subject but they may not have the specific topic of interest. We analyze the security and privacy of our scheme and evaluate the communication and computation overhead. Index Terms—Trust; privacy preservation; Mobile Social Net- works; and Data Retrieval. I. I NTRODUCTION With the rapid increase in the number of mobile devices, the applications of Mobile Social Networks (MSNs) are becoming very popular. MSNs allow users to discover and interact with existing and potential friends [1]. They involve real- time communications to enable information sharing and social interaction amongst users. Many promising applications of MSNs require exchanging information on a subject of interest. Users who are interested in a subject do not necessarily possess the same topics of interest. That is why it is not sufficient to know the users’ subjects but also their topics of interest. In this paper, the term “subject” refers to the broad category of interest, like “Soccer”, “Politics”, etc., while the term “topic” refers to a sub-category under the subject. Usually topics are many and fine-grained, e.g., the subject “Soccer” can have topics like “World Cup 1998”, “Spanish League 2015”, etc. Our scheme target three applications called chatting, file sharing, and Web page pre-fetching. In chatting applications, users need to contact other users about a topic of interest. For instance, if a user wants to ask someone who has experience in repairing cars, it should first look for someone who is knowledgeable in the subject “Car Repair”. Then, it may need to know fine-grained information about users’ experience like “Air Conditions Repair” or “Changing Tires”. These are fine grained topics under the subject “Car Repair”. In file sharing applications, each user stores a group of files, e.g., for songs, pictures, etc., and users can request them. In such applications, if the subject is “Songs”, topics can be the names of the stored songs. In Web page pre-fetching applications, when a mobile device is connected to Wi-Fi, it downloads the Web pages that will be most likely requested by the user, so that the downloaded pages will be used when the mobile device does not have Internet connectivity. However, it is possible that a user needs a Web page that has not been downloaded. In this case, one can contact nearby friends to request the Web page. For instance, in this application, if the subject is “Politics”, a topic is the requested Web page. However, security and privacy issues in these applications are becoming a real concern. In this paper, we propose a trust-based and privacy- preserving data retrieval scheme For MSNs. The scheme addresses security and privacy issues such as how to securely identify a trusted user who is interested in a subject, how to prevent others who are not interested in a requested subject from knowing the subject, and how to securely communicate and exchange information with a friend who possesses a topic on the subject of interest. First, we discuss a trust-based and privacy-preserving log creation scheme to enable users to build a log of trusted users who store some topics related to a subject of interest. In order to preserve privacy, if a user is not interested in a subject, he/she cannot know the requested subject by the scheme. Also, distrusted users cannot know the subject even if they are interested in the subject. In creating the log, we leverage friends-of-friends relationships and transferrable trust concept. Each user trusts its friends and the recommended friends by his/her friends. When a user looks for users who are interested in a subject, it does not only search in its list of friends but also the friends of friends. If a friend is not interested in the subject, he/she can help his/her friend by linking the friend to his/her friends without knowing the requested subject. The scenario can be extended to include more friends of friends. The motivation here is that collecting topics only from direct friends may not be enough and users may need a wide range of topics. Our scheme enables the users to share symmetric keys with friends of friends. It uses Bloom filters to reduce the storage and computation overhead of the users’ topics. This is important because the number of topics may be large. After creating the log, it can be used to request a topic from the users who possess it. The data retrieval in our scheme is fine- grained because users exchange fine-grained topics, rather than querying users who have the subject of interest, but they may not have the topic of interest. We analyze the security and