Strategy Of Data Center Network Traffic Scheduling Based On ACO

With the continuous development of cloud computing and virtualization,the types of services and data volumes in data center network are growing rapidly.At the same time,users put forward higher requirements for service reliability and delay.How to improve network throughput and guarantee the quality of service of users through traffic scheduling under different load conditions of the data center network is an urgent problem to be solved.Therefore,the thesis analyzes the traffic scheduling strategies of data center network.The main research contents are as follows:1.The existing strategies of link load balancing in data center network often ignore the complexity of the network and the quality of service of the flows when calculating the transmission path.Therefore,the thesis proposes a strategy of link load balancing based on ant colony optimization,which aims to find the best transmission path for the flow under the constraints of path length,load and delay.Aim to find a path at low load more accurately,the strategy redefines the heuristic function based on the large-flow numer on the link and the real-time load.At the same time,in order to reduce the delay,the strategy determines the best path based on delay and real-time load.The results of the study are that the proposed strategy improves the link utilization and network throughput,and reduces delay and delay jitter to some extent,compared with ECMP and Hedera strategies.2.However,when the network load increases,strategies dedicated to load balancing tend to cause bandwidth fragmentation results in the network throughput decrease.Therefore,the thesis proposes a traffic scheduling strategy for bandwidth fragmentation minimization and QoS guarantee.When the overall load of the network is lower than the pre-set threshold,the strategy establishes the weight function for each candidate path according to the free bandwidth of the path and the application requirements of the small flows,and the forwarding path is selected based on the weight function value by roulette algorithm.When the overall load of the network exceeds the threshold,in order to make full use of the link bandwidth resource,the strategy redefines the heuristic function according to the remaining bandwidth of the path,then select the best path according to the remaining bandwidth fluctuation information of the path,finally get the best transmission path after multiple iterations.The results of the study are that when the network load is high,the proposed strategy reduces the packet loss rate and delay of small flows,and ensures the network throughput,compared with ECMP and Hedera strategies.