Research on computing resource allocation strategy for cloud radio access network based on demand forecasting

Cloud radio access network(CRAN) is a new network architecture commonly used in 5G network. It uses network function virtualization and software defined network technology to support end-to-end network slicing so that access network can share the same infrastructure. A computing resource allocation scheme based on virtualized network function(VNF) computing resource demand forecasting was proposed for cloud wireless network supporting end-to-end network slicing. Data-driven operation and maintenance framework established by the control plane collects network in- formation and performs data processing to predict the demand for computing traffic in a period of time, which used the discrete particle swarm optimization based on best-fit-decreasing to dynamically and dynamically allocate CRAN com- puting resources, reducing the migration energy consumption and migration times of VNF.