Service Feature Oriented Cross Domain Resource Allocation In Non-Terrestrial Networks

In recent years,the terrestrial cellular network has developed rapidly.With the official commercial use of 5G,the increase in the number of users and mobile devices,and the emergence and vigorous development of emerging industries,the data traffic in the network has increased sharply.This puts forward higher requirements for the stability and communication ability of the current mobile communication network,which is a powerful challenge for the development of mobile communication network.The satellite network and ground network are integrated and cooperated.On the basis of realizing the full coverage of communication services,the communication network architecture increases the available resources in the communication network in a disguised manner to cope with the surge of traffic.As one of the research focuses of satellite ground integration network,the research and development of non-terrestrial network have attracted much attention.In the process of satellite ground integration of non-terrestrial network,the management,allocation and scheduling of various cross domain resources is one of the keys to promote the optimization of system performance.Based on the characteristics of various services in non-terrestrial networks,this paper designs a network slice resource allocation method suitable for non-terrestrial networks,and designs a resource scheduling method with both slice level and user level to meet the needs of various services.This paper carries out innovative and effective cross domain resource allocation.The main research contents and scientific research innovations of this paper include the following points:1.The OPNET simulation software is used to build a non-terrestrial network simulation platform with CU(Centralized Unit)and DU(Distributed Unit)separated,in which the cross domain resource management algorithm is deployed to complete the simulation and obtain the simulation results.It mainly completes the following functional modules of the satellite node in the platform as the distributed unit(DU):network slice deployment module,cache resource allocation module,user scheduling module,etc.The ground centralized unit(CU)distributes the resource requested by the satellite DU,and has the resource scheduling function module of the satellite DU.The ground user node completes the following functional modules:sending resource request,receiving scheduling from DU,calculating signal-to-noise ratio and throughput,carrying out channel quality feedback(CQI).This platform models and simulates the protocol plane functions of non-terrestrial network CU and DU,and can realize the completed communication process.2.A cross domain resource allocation algorithm based on network slice is proposed.The satellite caches some resources in advance to mitigate the impact of high delay of feeder link and meet the QoS requirements of different types of services at the same time.Design a network slice deployment method suitable for the characteristics of nonterrestrial network with the separation of CU and DU.By allocating the cache resources of satellite DU,avoid the concentration of network resources loaded by the feeder link;Considering that the non-delay sensitive services in the non-terrestrial network have low demand for early caching,the rental cost of satellite DU when requesting services from the ground CU in real time is modeled.Facing the QoS characteristics of business,the QoS optimization objectives of different types of slices are customized and modeled.The system utility function composed of customized QoS satisfaction,resource utilization and resource rental cost of different slices is taken as the optimization objective.The objective function is solved by KKT conditions to obtain the optimal network slice deployment strategy of each slice type and improve the QoS satisfaction of each type of business.Simulation results show that compared with the traditional algorithm of no network slice deployment and simultaneous interpreting of resource allocation,this algorithm improves the QoS satisfaction of all kinds of services by 5%,30%and 40%respectively,and has better resource allocation effect.3.A business feature oriented dual priority resource scheduling strategy is proposed.After the deployment of network slices,the allocation of cache resources and wireless bandwidth resources,the algorithm is deployed based on the RB pre allocated scheduling model.Design the joint scheduling algorithm of slice level and user level,add the influence of the priority of resource allocation on the feeder link on the system into the modeling factor,and model and reflect the difference of business characteristics.When calculating the priority weight factor,do the joint design of double priority,and integrate the QoS satisfaction of various slices,the weight of each satellite DU for each kind of slice,the user’s rate,user’s QoS satisfaction and other factors can flexibly adjust their weights according to the actual situation after the deployment of network slices,as an important calculation factor of scheduling order.Simulation results show that compared with the PF resource scheduling algorithm,M-LWDF resource scheduling algorithm and DPF resource scheduling algorithm,simultaneous interpreting algorithm improves the system throughput,system utility and user fairness index by 50%,15%and 11% respectively.