Research On Load Balancing Of C-RAN Based On Software Defined Network

C-RAN(Centralized,Cooperative,Cloud RAN)is one of the standards for implementing software definition and green communication technology in the fifth generation wireless network(5G).It is an efficient processing architecture with real-time cloud computing and collaborative radio capabilities.However,it faces some challenges: the interconnection and inter-working of Base Band Unit(BBU)in the base band processing pool and the problem of sharing and efficient management of BBU virtual machine resource.But,with the development of Software Defined Network(SDN)technology based on the separation of control and forwarding,there is a new direction in the research of C-RAN network.In this context,the load balancing problem of C-RAN based on SDN technology is studied in this thesis.Firstly,a C-RAN load balancing architecture based on SDN technology is studied in this thesis.The SDN controller is added to the BBU pool to realize the interconnection of BBU virtual machines and the sharing and efficient management of network resources.This thesis elaborates the operating mechanism of the architecture,and analyzes the information exchange based on the OpenFlow protocol.At the last,this thesis focuses on the design method of the SDN load balancing controller.Then,the scalability of the architecture is analyzed.Secondly,based on the above architecture,a joint load balancing algorithm that takes into account the Quality of Service(QoS)and the remaining capacity of each BBU virtual machine is presented in this thesis to solve the problem of load balancing among BBUs.The joint load balancing algorithm mainly includes QoS optimization algorithm and load balancing algorithm between BBUs.First,the load difference of candidate BBU is graded,and then different algorithms are selected according to the difference level.If the difference is small,a multi-attribute QoS evaluation method based on TOPSIS(Technique for Order Preference by Similarity to an Ideal Solution)is designed.Otherwise,a load balancing algorithm between BBUs based on utility theory is designed.Simulation results show that the algorithm reduces the call blocking rate and improves the system throughput.Finally,a BBU-RRH(Remote Radio Head)load balancing algorithm based on Discrete Particle Swarm optimization(DPSO)is proposed to solve the problem of wasteof spectrum and computing resources caused by static BBU-RRH configuration in traditional C-RAN networks.The dynamic load balancing of BBU-RRH is formulated as a linear integer optimization problem.Then,the load balancing optimization objective function is constructed based on load fairness index,average load index and handover index,and the improved DPSO algorithm is used as optimization algorithm.Because of the BBU-RRH configuration is a discrete optimization problem,there are some incomprehension in the solution space.Therefore,the greedy operator is added to the DPSO algorithm.At the same time,in order to avoid the algorithm falling into the state of local optimal solution,the proportion of particle learning factor is adjusted in the algorithm.Simulation results show that the algorithm can reduce the number of required BBUs and the number of blocked users.