Multi-factor Load Balancing Algorithm for 5G Power Network Data Center

Abstract

In 5g era, data center is the core facility of load data exchange, and its internal network traffic shows explosive growth. Aiming at the coexistence of bandwidth sensitive traffic and delay sensitive traffic in power data center network, this paper proposes MFLB (multiple factor load balancing) algorithm. Considering the multi-dimensional statistical characteristics of links, the algorithm integrates six parameters and four dimensions of hops, path delay, average link delay, maximum link delay, maximum residual bandwidth and minimum residual bandwidth to filter paths. The simulation environment is built based on mininet. By comparing with the existing algorithms, the performance of MFLB algorithm in average throughput, standardized total throughput and link delay is significantly improved.