EventMe: Location-Based Event Content Distribution Through Human Centric Device-to-Device Communications

Fengrui Shi,Zhijin Qin,Julie A. McCann
DOI: https://doi.org/10.1109/icc.2018.8422703
2018-01-01
Abstract:Location-based information dissemination has become increasingly popular in the recent years. Extensive research work has been done on the matching of interested parties to event information via publish/subscribe systems. However, the rich content types of such location-specific data, especially when the data are presented in multimedia form, requires efficient methods with low cost to transfer the content to the subscribers. In this paper, the potential of utilising human centric device-to-device (D2D) communications to disseminate location-based event content is investigated. The human centric D2D data dissemination process is formulated as a task assignment problem, which can be modelled as a Integer Quadratically Constrained Quadratic Programming (IQCQP) problem. Since the IQCQP problem is in general NP-hard, a sub- optimal polynomial framework named EventMe is proposed, which is able to compute a solution with guaranteed lower bounds on data distribution capacity in terms of throughput. Through extensive evaluation using several real world datasets, it has shown that EventMe is able to improve the network throughput by 100%-500% compared to baseline methods. A prototype is developed and shows that it is practical to implement EventMe on mobile devices by generating minimal control data overhead.
What problem does this paper attempt to address?