Network Slicing Based Computing And Communication Resource Allocation For Cloud Radio Access Network

With the development of terminal technologies such as wireless technology and mobile Internet,mobile traffic on the wireless access network has exploded,and a variety of business demand has been spawned,such as driverless vehicles,telemedicine,etc.Faced with more diverse and differentiated business needs,the network slicing is chosen as the core of the network architecture in the actual development of 5G.Network slicing uses virtualization technology to share multiple logical networks on the same physical infrastructure,and network slicing services are deployed and upgraded through software.The network slicing resource allocation is on-demand,and it can reduce the network's Capital Expenditure and Operating Expense.The deployment of network slicing not only provides the basis for the introduction of artificial intelligence and big data on the wireless network,but also makes its resource allocation more complicated.How to combine big data technology to design efficient network slice resource allocation,reduce cost and improve operation has become an urgent problem to be solved.This thesis studies the allocation of computing resources and communication resources for network slicing under the cloud radio access network infrastructure.Firstly,this thesis studies the virtualization technology of network slicing and designs a data-driven network slicing operation and maintenance framework.Through the network framework,data information in the access network is collected,and a single-step time series prediction and a multi-step time series model are established.Secondly,considering the dynamics of slice resource allocation,in order to prevent service fluctuations caused by excessive dynamic allocation of computing resources,this thesis designs a computing resource allocation scheme based on the traffic information which predicting the traffic for a period of time in the future.The designed resource scheduling algorithm reduces the fluctuation of each service under the premise of ensuring the utilization of computing resources.Finally,under the heterogeneous cloud radio access network infrastructure,network slicing communication resource allocation in the case of multiple services coexistence.In this scenario,different user utility functions are established according to the communication resource requirements of different services,and an optimization problem is solved.Base station sleep,user access,channel selection,and power allocation issues are considered in the design of the communication resource allocation algorithm.At the same time,it is considered that the sleep of the base station cannot be turned on and off too frequently according to the real-time user demand.In the design of the sleep strategy of the base station,the statistical traffic of the next time period is predicted to guide the sleep of the base station.The communication resource scheduling algorithm can improve resource utilization when the utility of the user is satisfied.In summary,this thesis carries out the network slice computing resource and communication resource allocation scheme design for the cloud radio access network infrastructure.While extending the network slicing architecture,it provides a solid theoretical foundation for the combination of 5G network and big data.