Abstract:With the rapid development of urbanization, enormous amounts of vehicular services have been emerging and challenge both the architectures and protocols of the Internet of Vehicles. The high-speed mobility features of nodes in the vehicular networks changes the network topology frequently, resulting in low routing efficiency, and higher packet loss. In this article, we utilize software-defined networking (SDN) technology to decouple the network control plane from the data forwarding plane, and divide the vehicular networks into three functional layers: data, control, application layers. Based on the proposed network architecture, we propose an adaptive traffic engineering (TE) mechanism to guarantee the V2V continuous traffic in vehicular networks with high-speed mobile vehicles or dynamic network topology. In particular, the proposed TE is based on a proposed active network measurement mechanism under the assistance of the centralized management ability of the SDN technique. The proposed active network measurement approach is a greedy approach where the next hop determination for the measurement packet takes multiple link reliability factors (e.g., the delay, the length, the packet error rate, the neighbors, etc.) into account. Then, we utilize the artificial bee colony (ABC) algorithm to optimize the TE mechanism that can be deployed and executed in the SDN controller. By the proposed TE mechanism, multiple candidate end-to-end paths can be concurrently measured, and the optimal data forwarding path can be adaptively switched. Simulation results demonstrate that our approach performs better than some recent research outcomes, especially in the aspect of performing reliable data forwarding (almost 5% better than the compared objects).

Adaptive and Low-cost Traffic Engineering based on Traffic Matrix Classification

Short-term Traffic Flow Prediction at Isolated Intersections Based on Parallel Multi-Task Learning

Traffic Matrix Estimation: A Neural Network Approach with Extended Input and Expectation Maximization Iteration

Traffic Matrix Prediction Based on Deep Learning for Dynamic Traffic Engineering

Traffic engineering in hybrid SDN networks with multiple traffic matrices

Asynchronous multi-class traffic management in wide area networks

MATE: when Multi-Agent Deep Reinforcement Learning Meets Traffic Engineering in Multi-Domain Networks

A Multi-agent Reinforcement Learning Perspective on Distributed Traffic Engineering

A Decentralized Model Predictive Traffic Signal Control Method with Fixed Phase Sequence for Urban Networks.

Distributed and Adaptive Traffic Engineering with Deep Reinforcement Learning

RedTE: Mitigating Subsecond Traffic Bursts with Real-time and Distributed Traffic Engineering

Dynamic hybrid routing: Achieve load balancing for changing traffic demands

Semi-Oblivious Traffic Engineering: The Road Not Taken

Flow-level and Efficient Traffic Engineering in Conventional Routing Systems

Enabling efficient routing for traffic engineering in SDN with Deep Reinforcement Learning

Flow-level Traffic Engineering in Conventional Networks with Hop-by-Hop Routing

Adaptive Traffic Engineering Based on Active Network Measurement Towards Software Defined Internet of Vehicles

Achieving Near-Optimal Traffic Engineering Using a Distributed Algorithm in Hybrid SDN

Achieving Optimal Traffic Engineering Using a Generalized Routing Framework

TrafficMCTS: A Closed-Loop Traffic Flow Generation Framework with Group-Based Monte Carlo Tree Search

Learning Based Methods for Traffic Matrix Estimation From Link Measurements