| With the rapid development of wireless communication and embedded technology,the diversity of IoT business types brings many requirements of mobility.For geographically dispersed mobile IoT devices,their mobility leads to a prominent dynamic imbalance of network traffic,which is mainly manifested in consistency of network state and traffic load balancing.How to coordinate and manage mobile devices in the scenario of IoT,to ensure the rapid dynamic adjustment of mobile traffic and the flexibility of data exchange in different geographical areas,is an urgent problem to be solved.Combined with the characteristics of Software Defined Network(SDN),which supports the separation of control plane from forwarding plane.In addition,a consistent overlay structure based distributed control mechanism is proposed to manage the IoT network in mobile scenarios.The main contents are as follows:(1)This paper has designed an effective framework of mobility management.Firstly,for the city-scale Internet of Things,multiple controllers are used to divide the network space into multiple network partitions for management.A distributed hash function based overlay structure is proposed to map the network partitions into nodes to implement distributed control.The overlay structure has efficient consistency maintenance in mobile scenarios.Then,based on this coverage structure,a cross-regional flow handover method is designed,which can effectively reduce the handover delay and improve the network throughput.Finally,scalable control and fault tolerance approaches are designed to maintain scalability and stability of distributed network effectively.(2)Based on the network calculus theory,an analysis model of performance for SDN is designed.The analysis model can accurately estimate the network status by using the accurate measurement method provided by SDN controller.Besides,the queue models of network performance are used to further analyze the performances of the OpenFlow switch and the SDN controller,thereby providing a more refined analysis method for the network cyberspace state in terms of multiple network performance indicators.Then,a dynamic mapping algorithm based on double hash function is proposed,which can adaptively alleviate the traffic highlighting problem caused by static mapping between switches and controllers.(3)This paper extends the functionality of the open source controller Floodlight and designs a prototype of a distributed SDN system that includes the proposed mobility management framework,performance analysis model,and load balancing functional modules.The simulation experiment of setting the moving scene is carried out to verify the experimental results,which show that the system prototype is effective. |