Abstract
With the rapid development of information technology, software defined network (SDN) has been widely used in data centers, and has become an important trend in the development of the new generation of networks. Flow scheduling is one of the core problems in large-scale distributed networks. Aiming at the problems of low bandwidth utilization and high delay in network flow scheduling in high-load data centers, a flow scheduling method based on improved ant colony algorithm, DI-ACO, was proposed. In this method, we construct a flow scheduling model, and a differentiated path pheromone initialization method based on link load and design a dynamic update mechanism of pheromone volatility coefficient. Experimental results in typical topologies show that, compared with ECMP and other existing algorithms, DI-ACO can effectively improve the average bandwidth utilization of network links by 7.6% to 30%, and effectively reduce the round-trip delay by 7.2% to 39.5%.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call