| C-RAN is a green wireless access network architecture based on cooperative radio,centralized process and real-time cloud architecture proposed by China Mobile.In order to solve the serious problems of resource wasting and inter-cell interference caused by the deployments of high-ensity base stations,A centralized multi-core architecture has emerged such as C-RAN.In the LTE system,the traditional single-core architecture has been unable to meet the high-speed,low-latency requirements of the network,and the Modules of LTE protocol stack has disadvantages of relatively low scalability and high complexity.How to optimize the existing protocol stack architecture which aims to increase the independence of various layer protocols and improve data processing capabilities has become an urgent problem to be solved in the LTE protocol stack architecture.The thesis is based on the “C-RAN Low-Power General-Purpose Processor Platform Development” project of the Institute of Computing Technology,Chinese Academy of Sciences.In order to solve the above problems,this thesis makes an in-depth analysis of the multi-core protocol stack architecture,and presents a multi-core architecture optimization scheme for C-RAN LTE protocol stack.The new architectural has the following merits: It has the functional isolation of the components,and the internal structure can be replaced and reconfigured after the independent encapsulation.It also can completely decouple the communication between modules,which greatly improves the processing capability of data services.In the above-mentioned multi-core protocol stack architecture scheme,the parallel data distribution system plays a crucial role.After in-depth analysis,comparison and research,this thesis gives an reformative data distribution system,and carries out detailed analysis and research on it,finally implements it.In terms of functions,the system can implement the communication functions of the protocol stack by centrally processing all data in protocol stack,and it can implement statistical multiplexing of computing resources based on online dynamic expansion,reduction,and migration of computing resources.In terms of performance,the system can use parallel processing of data to increase the data transmission rate,and use data distribution strategy and load balancing strategy to improve the performance.Finally,the above protocol stack architecture and data distribution system are tested.The integration test proves that the system can be compatible with other modules and run stably,and it can achieve statistical multiplexing of computing resources.The performance test shows that the performance of the multi-core parallel protocol stack architecture is superior to the traditional single core architecture in terms of the peak data processing rate and data processing delay.When the number of users is 64,the total performance gains up to 30%. |
