Abstract

Due to the upsurge of context-aware and proximity aware applications, device-to-device (D2D) enabled mobile cloud (MC) emerges as next step towards future 5G system. There are many applications for such MC based architecture but mobile data offloading is one of the most prominent one especially for ultra dense wireless networks. The proposed system exploits the short range links to establish a cluster based network between the nearby devices, adapts according to environment and uses various cooperation strategies to obtain efficient utilization of resources. We proposed a novel architecture of MC in which the total coverage area of a eNB is divided into several logical regions (clusters). Furthermore, UEs in the cluster are classified into Primary Cluster Head (PCH), Secondary Cluster Head (SCH) and Standard UEs (UEs). Each cluster is managed by selected PCH and SCH. An algorithm is proposed for the selection of PCH and SCH which is based on signal-to-interference-plus-noise (SINR) and residual energy of UEs. Finally each PCH and SCH distributes data in their respective regions by efficiently utilizing D2D links. Simulation results demonstrate that the proposed D2D-enabled MC based approach yields significantly better gains in terms of data rate and energy efficiency as compared to the classical cellular approach.