ABSTRACT
Proximity services intentionally exploit the geographical position of mobile devices, particularly the geographical proximity and social relationships of mobile users, to enable various applications, including commercial advertisement, communication offloading, and mobile social networking in proximity [1]. Existing technologies used to serve the proximity awareness can be broadly divided into over-the-top (OTT), and device-to-device (D2D) (peer-to-peer [P2P]) solutions. In the OTT model, the centralized server (usually located in the cloud) receives periodic location updates from user mobile devices (using GPS), and determines proximity based on location updates and interests. The constant location updates not only result in significant battery impact because of GPS power consumption and the periodic establishment of cellular connections, but also causes serious privacy problem. Moreover, OTT approaches may incur undesired network overheads and latency for discovery and communication. Different from OTT, D2D schemes forego centralized processing in identifying relevancy matches, instead of autonomously determining relevance at the device level by transmitting and monitoring for relevant attributes. This approach offers crucial privacy benefits. In addition, keeping discovery on the device rather than in the cloud will allow for user-level controls over what are shared.
