Bandwidth Resource Management Of Low-orbit Satellite Internet Based On Software-defined Multicast

Since the 4G,video sharing has become a popular social method in contemporary society with the significant increase in bandwidth capacity of the Internet.Compared with inefficient unicast,one-to-many multicast technology is more efficient,which can save a lot of bandwidth and reduce network load.However,traditional terrestrial optical fiber networks are difficult to achieve efficient and low-latency multicast due to crossoperator deployment and complex network equipment.With the development of electronic technology,large-scale commercial low-orbit satellite Internets have gradually emerged due to significant cost reduction in satellite launching.In the software-defined satellite network(SDSN)constructed by deploying a softwaredefined network architecture in the satellite network,not only can efficient softwaredefined multicast technology be realized,but also advanced bandwidth resource management can be implemented.Therefore,taking video distribution as an application,this thesis discusses the problem of network bandwidth resource management in the satellite Internet for the purpose of optimizing network throughput.The satellite network is a heterogeneous network composed of the inter-satellite network and the satellite-terrestrial network.The following effective bandwidth resource management algorithms are respectively proposed for these two parts of the network:In the inter-satellite network,the bandwidth resource management algorithm proposed in this thesis mainly realizes the optimization of throughput in two ways.In the stage of planning obstacle-avoiding route,the algorithm plans routes with sufficient bandwidth for all multicast group by routing migration,and minimizes total bandwidth consumption;In the stage of allocating the congested link bandwidth,in order to achieve efficient bandwidth resource allocation for multicast with high-reusability routing,the algorithm allocates bandwidth of congested links based on the branches of the routing tree associated with the multicast flow.The results of experiments show that,compared with the control experiment,the bandwidth resource management algorithm can optimize the throughput by more than 50%,and reduce network congestion.In the satellite-terrestrial network,for the satellite-to-ground multicast of multibeam satellites enhanced by frequency division multiplexing technology,this thesis builds a cellular cluster model based on the seven-color frequency reuse strategy and proposes a bandwidth allocation based on beam clusters.Considering the bandwidth resource competition among beams in the same cluster,the algorithm can calculate an excellent beam coverage scheme,and then simplify the problem variables to a smallscale core set to get the bandwidth resource allocation scheme,which can realize flexible bandwidth resource scheduling among beams.Compared to the control experiment,the experiments show that the throughput is improved by about 40% at most.