The Research Of The Baseband Pool Resource Management Model And Related Algorithm Based On Cloud Radio Access Network System

C-RAN, A new generation of mobile communication network, which is based on centralized processing, wireless access network architecture of collaborative radio and real-time cloud computing architecture, divides the traditional distributed base station into a baseband processing unit and a remote radio unit and centrally manages all the baseband processing units, which are connected with all the remote radio units by high bandwidth and low latency optical transmission network. The disadvantages of the high cost of construction and operation, low spectrum utilization rate, and the difficulty in expansion and upgrade in tranditional base station are improved by C-RAN which can meet the growing needs of users. The resource management, resource configuration and related algorithm in the baseband pool of C-RAN are gaining increasing attention. In order to improve the utilization rate of resources and the operating income and to reduce the overall energy consumption of the system and total delay of task, the design of the virtual machine in the baseband pool and virtual machine resource management of C-RAN system will be studied in this dissertation, which establishs a efficient management model and the improvement of corresponding algorithm, under the premise of ensuring the prescribed standards of service quality. The main contribution of this dissertation:(1) The baseband processing part of C-RAN architecture has been mainly studied in this dissertation. Using the characteristics of wireless real-time communication and virtualization technology, the virtual machine configuration method of baseband pool has been explored. Based on the constraints of resources and hard real-time that the protocol required, the optimization model of the virtual machine configuration has been set up for improving system resource utilization and reducing the total task processing time. The technology of mobile sliding window has been used to enhance robustness, to make the algorithm to achieve faster convergence speed and better value. The reasonablility of virtual machine configuration scheme and the superiority of the improved algorithm has been verified by using experiment and simulation. A more excellent virtual machine configuration scheme is obtained.(2) The character of asking service for customers has been analyzed and the mixed resource allocation model has been proposed. For fixed resource part in this model, aiming at reducing the system energy consumption with specific customer request load, the virtual machine cluster management model, which is matched with customer demand, has been established. With different user request arrival rate, energy consumption optimization-based fixed clustering scheme, which matches the user request, can be designed. Additionaly, "precocity" problem of the existing genetic algorithm has been improved by improving the selection operator and crossover operator. Faster convergence speed and better convergence value has been also obtained. The superiority of the algorithm has been verified by experiment simulation.(3) In this dissertation, the resource allocation model for achieving maxmum basedband pool resource income has been established and shared resource capacity configuration of the total request load for different customers has been assured by introducing the concept of operating income. When the fixed cluster resources is not enough, shared resources dynamically configurate virtual machines for all standard user request and accomplish related task processing. It can both ensure the user service quality and the baseband pool resources utilization.In this dissertation, the baseband pool configuration of the virtual machine, the virtual machine cluster management and shared resource pool configuration are researched, but due to time and other factors, this research also has many deficiencies to continue in-depth study. Dynamic scheduling of virtual machine cluster and virtual machine resources will be researched in the future study, under the premise of the various types of business demand in a real-time communication environment.