Desktop version

Home arrow Engineering arrow Modeling and Optimization for Mobile Social Networks

Security Services

The MSNs, which can easily be constituted by smartphone users in a local area, has become a hopeful social networking platform to enable media sharing, social gaming, group chat, and interaction among users nearby [21]. Through WiFi, Bluetooth, and device-to-device communications, connections are established among users and an opportunistic network is formed for a long or temporary period (e.g., several hours). For instance, rich interaction opportunities are created for students in a campus area, residents in an urban neighborhood, tourists visiting or a scenic site or museum, and customers in a shopping mall. Without worrying about the wireless data charge as well as Internet access, MSNs allow users’ interactions anytime and anywhere. A recent report from comScore indicates that Twitter users take over 86 % of time on mobile devices rather than desktop in the United States. And, for Instagram users, the percentage is 98 %. In a word, with the rapidity, efficiency, and pervasion of MSNs [22], users can have high QoE to obtain social services (e.g., crowdsourcing) [23, 24].

An investigation by Nexgate reveals that during only the first half of 2013, spams over social media have risen by about 355 %. They are fast disseminated in MSNs as every 1 of 200 social media posts is considered as spam. Spam filtering has attracted much attention in both industry and academic. Several services have been designed on the basis of blacklist to block spammers [25] or whitelist to permit legitimate senders. Filtering can be realized through content inspection by matching the keyword with the packets [26] or leveraging machine learning techniques [27] to detect spams. Relevant characteristics and social graph are also studied for spam filtering [28, 29].

Most of services which are implemented by a credible authority or centralized server need the historical information to detect spams. If there is no credible and centralized servers and lack of historical information in MSNs, spammers are more likely to be undetected [30]. A distributed filtering servcie can help MSN users to personalize the spam filters, and send these spam filters to others filter holders to allow them to filter spams efficiently. Nevertheless, for the spam filtering service in MSNs, there are still some challenges. Firstly, taking into account both the filtering accuracy and distribution costs, how to distribute filters is a challenge. Secondly, some filter creators’ sensitive information, such as lifestyles, physical situation and preferences [26,31], may be remained in the distributed filters. The privacy and security problems should be considered. In addition, some malicious attackers may forge the original filters to block some useful information, how to resist these malicious attackers is the third challenge.

< Prev   CONTENTS   Source   Next >

Related topics