| The ground mobile communication system has entered the 5G era,and higher coverage is required in large-scale Internet of things(IoT)business scenarios,while satellite communication makes up for the lack of ground mobile communication due to its wide coverage and good communication quality.The integrated satellite and 5G terrestrial networks can connect multidimensional space of air,sky,earth and sea,and realize global three-dimensional coverage.With the rapid development of mobile communication and the increasing demand of user service,due to the different communication system between ground and satellite,each ground base station is equipped with proprietary physical hardware equipment and the resource efficiency of the equipment is low,which leads to the surge of base station quantity and the increase of base station construction cost.With the help of virtualization and cloud technology,the baseband unit is virtualized and managed centrally to form a virtual baseband pool,which can not only realize the sharing of physical resources,improve the utilization rate of server resources,but also manage the virtual resources of the baseband pool flexibly,so that the resource load is balanced.Too high or too low load of servers in the BBU pool will affect the performance of the virtual baseband unit and increase the total energy consumption and maintenance costs of the physical equipments.A dual threshold adaptive dynamic migration strategy is proposed around the resource state of the server to reduce baseband pool energy consumption and service level agreement violations.This strategy sets the upper and lower threshold of server resource utilization,and combines the Kalman prediction algorithm to judge the relationship between resource utilization and threshold range of computing nodes to determine whether to trigger dynamic migration.If the migration is triggered,the strategy first selects the virtual machine to be migrated according to the minimum number of migrations and the migration time standard,then comprehensively determines the target node based on the load situation,and finally realizes the dynamic migration.In this paper,the proposed strategy is simulated on Cloudsim and compared with the existing strategies.The simulation results show that the proposed strategy can effectively reduce the number of migrations and time on the basis of reducing energy consumption and service level agreement violations.Under the background of 5G satellite-ground integration,this paper implements the construction of an OpenStack-based baseband pool virtual resource management and control platform.Multiple virtual machines are virtualized on the server through KVM and the virtual baseband unit functions deployed on the virtual machine,to achieve the sharing of the underlying hardware resources of physical devices.Centralized management of virtual resources to build a virtual baseband pool and implement the control of the baseband pool with the help of OpenStack.Through the evaluation of the management and control platform,the scalability,effectiveness and stability of the platform are verified.Finally,this paper develops and implements the proposed dynamic migration strategy on the OpenStack platform.The actual operation results show that this strategy can replace manual operation and automatically carry out dynamic migration to complete dynamic scheduling of virtual resources and reduce unnecessary migration. |