Research Of VANET’s Routing Management Based On Cloud Platform

Vehicular Ad Hoc Networks (VANET),which is regarded as one specific appllication of Mobile Ad Hoc Networks (MANET), is oriented to vehicular communication.The prospect of its applications is in emergency management,driver assistance,traffic information sharing and it shows good prospects for development.For the self-organizing network VANET with no control center, fully distributed architecture and dynamic topology, there exist more difficulties in real-time applications,network management and QoS assurance.VANET has characteristics including narrow bandwidth and topology rapidly changing, and traditional routing protocol does not apply to it. A kind of car networking path scheduling system based on cloud platform with high stability and scalability was investigated. Specific contents are as follows:1.Architecture, structural model and features of self-organizing network VANET was studied. Key technology of VANET including Dedicated Short Range Communication (DSRC), Wireless Access in theVehicular Environment (WAVE), IEEE 802.11p、IEEE 1609.4and IEEE 1609.3.2.For GPSR routing protocol in a multichannel environment, greedy forwarding and neighboring forwarding selected route is not necessarily the path with optimal performance. Multi-channel WAVE and GPSR routing protocol were combined, based on cross-layer design,to study multi-channel routing protocol MGFRP.3.Hierarchical structure,with the better scalability than the planar structure,was usually constructed using clustering algorithm. When signal strength of the cluster head was only considered during the node joining the cluster, the cluster showed poorer stability. In the paper, was put forward based on ant colony algorithm,and distance as well as Relative velocity direction was synthetically considered. Furthermore, the superiority of LEACH-RSD was verified by the simulation results.4.Three service models, five features of cloud computing and four cloud deployment strategies of cloud platform were studied. In addition, a private cloud platform was built using Openstack.5.MGFRP was simulated employing NS2, whose results were further analyzed by AWK script. The results showed that MGFRP had a smaller delay compared with traditional GPSR, DSDV, DSR.