Research On Key Technologies In C-RAN Based 5G Network

In order to meet the rapid growth of data stream in 5G network,Cloud Radio Access Network(C-RAN),which can reduce the system cost significantly,is a potential system architecture investigated by more and more researchers.The processing parts,also known as Baseband Units(BBUs),are centralized to form a BBU pool in C-RAN system.The BBU pool is connected with the distributed Remote Radio Heads(RRHs)through optical fronthaul.With the employment of cloud computing technology,C-RAN system enj oys reduced system cost,enhanced resource utilization ratio and high-level network flexibility.However,how to develop effective evaluation methodologies and optimization algorithmes has brought big challenges for the development of C-RAN.In this paper,the evaluation method and resource optimization problems are investigated.Firstly,Software Defined Radio(SDR),which is the key technology to form the BBU pool in C-RAN is studied.A SDR testbed is set up and tested for further research.Then a SDR based C-RAN system model is proposed according to the testbed architecture.BBU load balancing is carried out under that system assumption.Finally,the C-RAN system model is extended by introducing network slicing functions.Radio resource allocation for multi-service co-existing scenarios under the new system model is also investigated.The main contribution and novelty of this paper are as follows.1.Research and deployment of SDR testbedSDR,which is employed to build the BBU pool and carry out the load balancing process,plays an important role in the actual deployment of C-RAN.Based on the recent survey about SDR,OpenAirInterface(OAI)is employed to build the experimental testbed.The function structure and mechanism of OAI are then analyzed.Furthermore,a cross-layer optimization use case is deployed to optimize the actual video streaming.Finally,experimental results show that OAI is a whole realtime protocol stack,which can be used to build and evaluate the system model of SDR based C-RAN.Meanwhile,with the help of SDR,cross-layer optimization can be easily carried out in C-RAN.It will be an interesting idea to optimize the application service by adjusting the radio resource through cross-layer interfaces.2.BBU load balancing in C-RANThe main feature of C-RAN is that the processing parts,which cost most of the power in C-RAN,are centralized in the BBU pool.In that case,the resource utilization ratio in the BBU pool plays an important part in optimizing the system performance.In order to make better use of the computing resource in the BBU pool,it is necessary to design better resource allocation scheme for the BBUs.A system model of SDR based C-RAN is proposed in section III according to the testbed introduced in section II.Then the mechanisms of SDR,Virtual Machine(VM)live migration and data transmitting are investigated.Radio resource waste,computation resource waste and switching cost are analyzed and weighted to form a comprehensive evaluation function.Based on the evaluation function,a dynamical BBU load balancing scheme is proposed in section IV.Finally,simulations are carried out to validate the evalutation method and optimization scheme.It is shown by the simulation results that the proposed evaluation method can evaluate the system performance effectively and the proposed optimization scheme can keep the BBUs working in a proper load with low switching cost.3.Packet scheduling in network slicing C-RANA network slicing C-RAN system is proposed and analyzed in section V based on the cross-layer use case introduced in section II.In the slicing C-RAN system,ultra Reliable Low Latency Communication(uRLLC),enhanced Mobile Broadband(eMBB)and massive Machine Type Communication(mMTC)share the same physical network.Then the packet transmitting mechanism is analyzed to propose a Quality of Service(QoS)evaluation method.Reinforcement learning(RL)tools are used to optimize the slice scheduling scheme so as to maximize the total QoS of the three services.Simulation results show that the RL optimization scheme can satisify the services' different QoS requirements as much as possible with limited radio resource.