In recent years, intelligent transportation systems are a vital part of the development of smart cities. In intelligent transportation systems, it is necessary for each vehicle to know the states of all other running vehicles within a range, depends on the road speed limit. The states include location, travelling direction, speed, and possibly other useful information. Knowing the states of surrounding vehicles is critical in high-risk locations, such as traffic intersections and roundabouts. The development of 5G and its technologies, such as massive multiple-input multiple-output antenna, enables a new communication model, termed vehicle-to-infrastructure edge network (V2IEN), to achieve the above communication goal. In this paper, we study the optimization of the data throughput of the proposed V2IENs in 5G with the Time Division Duplex scheme. We show that due to the nature of information flow in the proposed V2IENs, the uplink and downlink traffic loads are asymmetric. This asymmetric characteristic allows maximizing the degrees of freedom of the V2IENs by optimizing the time-slot resources for uplink and downlink transmission. In this work, we present the sum-DoF capacity of the V2IENs with proof. We demonstrate that a significant DoF gain is achieved for the V2IENs by carefully allocate the system transmission time resources. We further propose an iterative algorithm for network sum-rate maximization via the optimization of the vehicle and base station precoders. Numerical results demonstrate that a careful design of the precoders at the base station and at the vehicles can considerably improve the V2IENs performance.