SDN dynamic flow scheduling algorithm based on discrete particle swarm

Abstract

Aiming at the problem of unreasonable traffic path allocation in data center networks, which easily leads to large flow conflicts, a software defined network (SDN) traffic scheduling algorithm based on discrete particle swarm is proposed. The algorithm redefines the solution search process within the particle swarm, aims to optimize network performance, dynamically allocates the optimal path for the traffic in the data center to reduce large flow conflicts; and introduces Metropolis to design diversified optimal scheduling schemes to ensure the reasonable scheduling of data center traffic. The experimental results of comparison and verification with other traffic scheduling algorithms show that the algorithm improves network quality, reduces large flow delay, and achieves better load balancing.