Video streaming over IoV using IP multicast

Abstract

Video traffic over Internet of vehicles is increasing day by day to provide passengers on-board infotainment services. A video streaming service over Internet of vehicles involves design challenges of vehicular gateway selection, and vehicle to vehicle collaboration. Vehicular gateway selection is challenging due to high vehicle mobility and time-varying link conditions. Existing works on vehicular gateway selection focus on either service cost minimization, or quality of experience. To achieve a balance between these two aspects, we propose a novel vehicular gateway selection protocol that minimizes streaming service cost, keeping end-to-end delivery delay as a constraint. In our work, the vehicular gateway selection problem is formulated centrally as the minimum set-covering problem, and is solved by the greedy approximation method. In existing works, vehicle to vehicle collaboration happens in the application layer ignoring the presence of network layer nodes which do not run any proprietary application. To make the vehicle to vehicle collaboration in the network layer, we propose a novel multicast protocol SS-CAST specifically designed for streaming service over vehicular networks. Compared to existing protocols, the simulation results show VSIM as the clear winner for minimizing service cost while giving a competitive performance in the quality of experience metrics.