| Devices-to-devices (D2D) communications is one of the most important technology to improve spectrum efficiency, system throughput and energy efficiency. D2D communications is also termed as proximity services (ProSe) and has been included in 3GPP Release 12. As a typical technology in 5G mobile communication systems, D2D communications is considered promising and will be widely used in the future. However,the unification of the collaborative data distribution mechanism multi-hop D2D networks is challenging since the wireless links are time-varying and the services in the networks are varied. Considering the complex architecture of multi-hop D2D communication systems, we should analyze different user behaviors in various application scenarios and then formulate the corresponding data distribution model and caching management strategies for cellular traffic offloading. Typically, user mobility is one of the most critical factors affecting the transmission performance because multi-hop D2D networks work with the characteristics of delay tolerant networks (DTN).Most of the existing data distribution algorithms in multi-hop D2D networks ignore the participation of cellular networks and seldom consider the condition that the network is instable. Similar to DTN, multi-hop D2D networks has a low user density and cannot always guarantee the connection of any two user nodes. Therefore, in this work, we will propose a data distribution algorithm considering the assistance of the base station based on user mobility in multi-hop D2D networks. Based on the existing work, a multi-hop D2D-based data distribution algorithm is proposed in this paper. By considering the assistance of the BS and the impact of user mobility on the data distribution scheme, the selection of multi-hop data distribution path is performed based on user mobility, e.g., the contact duration, inter-contact time and contact frequency of the communication between user nodes. Then the shortest path between any two nodes is calculated, and several nodes that can connect to the most user nodes in the network are selected to act as agent nodes. Efficient data distribution can be achieved by exploiting these agent nodes, since they can obtain the requested data blocks from the BS and disseminate data among mobile users quickly and widely, further leading to increase delivery success probability and reduce transmission delay.On the other hand, when considering the distributed collaborative caching schemes in multi-hop D2D networks, most of the existing works design caching strategies have limitations. Some of these works did not take user mobility and large-scale networks into account, and some are limited to one-hop D2D networks. Currently, the researches about the distributed storage in multi-hop D2D networks remain open. Given the complicated network structure and limited transmission resources in multi-hop D2D networks with hotspot, this work mainly targets the design of the distributed data storage schemes including caching content selection and update by considering factors including user mobility and user interest.Based on the encounter probability and interest similarity among mobile users, we can obtain the linear programming optimal solution to the objective function and the hotspots caching content according to the optimal solution. The interest of nodes will be updated based on encounter probability so that the caching content in the hotspot will also be updated.Numerical results have shown that the scheme proposed in this work can effectively increase delivery success probability and reduced transmission delay. The data distribution algorithm and distributed data storage schemes can improve the performance of multi-hop D2D networks. |
PDF Full Text Request