Research On Load Balancing Algorithm Of Air-space-ground Integrated Network Based On SDN In Emergency State

With the continuous development of modern society and the continuous changes of climate and environment,major emergencies such as earthquakes,floods and other natural disasters often occur in our country,resulting in heavy casualties and economic losses.In the event of a major emergency,it is the top priority to restore communications in the disaster area and communicate the situation in the disaster area in a timely manner.Therefore,a temporary solution is needed to quickly solve the large-scale continuous and stable communication problem in the disaster area under emergency conditions,and save more lives within the "golden 72 hours" of rescue.This paper starts from the multiple constraints of emergency state on communication in disaster areas.Based on the SDN(Software Defined Network)air-space-ground integrated architecture,taking advantage of the low dependence on ground facilities and flexible configuration of the architecture,a set of load balancing algorithm based on GCN(Graph Convolutional Network)algorithm is studied and designed.The algorithm scheme combines the traffic prediction load of each node in the SDN network architecture and other parameters of multiple nodes that may be affected by the emergency state,which solves the problem that the communication network in the disaster area is difficult to quickly cover the disaster area in a large area during the emergency state disaster relief process.The long-term sustainability and reliability of the communication network in the disaster area are insufficient.The main work and innovations of this paper are as follows:(1)Based on the emergency state conditions,this paper proposes a training process for the GCN algorithm model based on the node data of the SDN network architecture and the principle of the GCN algorithm,which can predict the traffic load level of each node in the SDN network architecture in advance,which is convenient Make adjustments to the nodes as early as possible.The algorithm customizes the relevant processing logic for these factors,so that the SDN network architecture can still ensure the long-term continuity and stability of the overall communication under the condition of multiple unfavorable factors in the emergency state.(2)Based on the traffic prediction load of each node of the SDN network architecture,and comprehensively analyze the multiple constraints of the emergency state,this paper proposes a comprehensive analysis of other parameters of SDN nodes(such as hardware status,etc.),SDN node based on the traffic prediction load.A load balancing algorithm for multiple factors such as point weights that are affected by the emergency state.The algorithm customizes the relevant processing logic for these factors,so that the SDN network architecture can still ensure the long-term continuity and stability of the overall communication under the condition of multiple unfavorable factors in the emergency state.(3)In this experiment,the mininet software is used to simulate the SDN network architecture,the performance testing tool Cbnech is used to verify the traffic load,and multiple hosts and virtual machines with different configurations are used to verify the load balancing scheme.The experimental simulation results show that by adjusting many parameters of the algorithm,the recognition accuracy of the GCN algorithm process designed in this paper reaches 85%,which proves that it can effectively predict the traffic load of each node of the SDN network architecture in an emergency state.The load balancing experiment proves that the load balancing algorithm flow based on the GCN algorithm designed in this paper can properly handle the load balancing support problem of the SDN air-space-ground integrated network architecture under the conditions of multiple unfavorable factors in the emergency state,so that it can keep the disaster area Continuous and stable communication capability.