Efficient Transmission of Live Multi-View Video in LTE-A Cellular Networks using D2D Cooperation

The increase in video traffic is causing considerable load on the already-limited cellular network resources. This poses a new landscape of challenges to service and network providers in distributing bandwidth-hungry services such as Multi-View Video (MVV). A paradigm to circumvent this challenge is to allow direct communication between closely located User Equipment (UE) without data routing via the radio and core network components. Device-to-Device (D2D) capable networks promise scalability and performance improvements in utilizing scarce resources. In this work, a mode selection strategy at the Base Station (BS) is proposed. It considers segment availability information at the UE to decide whether the Mobile User (MU) receives dedicated resources via a traditional Content Delivery Network (CDN) or reuse existing resources via the D2D communications channel. A novel architecture whereby the BS in the Radio Access Network (RAN) has video buffers, with caching maps which are aware of the video preferences of users in cell sites is analyzed. This presents the foundation to build a D2D-enhanced live MVV streaming system with no additional infrastructure for mobile devices. Simulation results provide guidelines for maximizing view and user diversity as well as bandwidth scalability in Long-Term Evolution (LTE) networks in order to achieve high system resource utilization whilst mitigating load on the network.