Research On Resources Optimization Algorithms For Cloud Radio Over Fiber Access Network

Cloud radio access network (CRAN) enabled by optical transport technology has been recently proposed as a next-generation mobile access network. It separates the legacy base station into remote heads(RRH) and centralized baseband unit (BBU) pools. "Any to any"connections between RRHs and BBUs can be realized through a flexible optical fronthaul network. However, due to the heterogeneity of network resources, how to integrate optical resources and computing resources to achieve a good scheduling and optimization is a difficult problem faced by optical access network. In this thesis, to solve these problems, we focused on design and implementation of algorithm. Some products have been achieved as follows:Firstly, for the energy consumption of BBU pool under tidal traffic scenario, an energy-efficient BBU aggregation was proposed by dynamically sleeping and awaking BBU cards. We proposed the mapping relation between optical resources and computing resources. For the problem of the joint optimization of optical resources and computing resources, we proposed the BBU capacity model and transform the BBU aggregation into an evolved two-dimensional bin-packing problem. In order to minimize the number of active BBU cards in the BBU pool, we developed an integer linear programming model and two heuristic algorithms: Load Reallocation with Best Fit Decreasing (LR) and Load adjustment with greedy searching (LA). Simulation comparisons were performed among different cell sites (i.e.,macro cell, micro cell, pico cell and heterogeneous cell) and network coverage. Numerical evaluations showed that BBU aggregation can achieve significant energy saving especially for small cell scenario.Secondly, the heterogeneity of various resources in the radio over fiber access network leads to the difficulty of network control, and drives the need of intelligent control more and more urgent. Software defined networking (SDN) is a solution of controlling multi-resources. Based on SDN, we developed an experimental platform which can control the optical resources and computing resources in radio over fiber access network flexibly. In this platform, we developed the northbound interface,realized the algorithm for BBU aggregation in controlling module and extended the OpenFlow protocol. The experimental results showed that the LR algorithm can also achieve good energy-saving result in our platform and our platform can control the equipment in optical access radio access network.