Research On Server Load Balancing Strategy Of SDN-based Cloud Data Center

Traditional network architecture has problems such as low resource utilization and difficult network flexible management,SDN decouples the control plane and the data plane,the network is dynamically managed and controlled by deploying strategies to controller through open programmable interface,so as to achieve the effect of software-defined network.At present,SDN architecture is widely used in cloud data center,which realizes the flexible management and adjustment of resources,but the ever-increasing requests make server cluster bear more load,and distributing load in balanced way is important for maintaining excellent performance in cloud data centers.Combined with the characteristics of the SDN architecture,this thesis discusses the server load balance problem of cloud data center,designs a server load balance strategy based on Open Flow switch,uses the wildcard mechanism to set multi-address directed flow tables and single-address directed flow tables,and change requests direction to server by adjusting the flow table entries,so as to achieve the server load balancing,and optimize the use of network resources.The main innovations of this thesis are as follows:(1)A server load balance measurement algorithm SLBM is proposed.Controller collects the application layer data of servers through SNMP technology as the basis for calculating the overall load and load imbalance.Set the overall load threshold in consideration of load fluctuations,and dynamically adjust the data collection cycle accordingly.When the overall load and load imbalance degree exceed the threshold,TCP and UDP traffic load of the transport layer is calculated according to the statistical data of the multi-address directed flow table to provide different granular data support for the load redirection algorithm.(2)An ant colony algorithm based load redirection algorithm ACO-LR is proposed.Service requests are partitioned and mapped through multi-address directed flow tables.Based on the transport layer load,controller uses ant colony algorithm to solve the load redirection scheme.During load migration,controller uses single-address directed flow tables to ensure orderly forwarding of traffic at different stages,thereby achieving safe and reliable load redirection.(3)Combining actual scenarios of the cloud data center,design the server immigration and emigration mechanism a to ensure scalability and robustness of the cloud data center server cluster.Controller monitors the load data and operating status of servers through SNMP technology,dynamically adjusts the number of running servers according to the server cluster load,monitor the abnormal server and migration the faulty equipment,so as to achieve the requirements of high performance and energy saving of cloud data center.(4)The SDN simulation environment is built to verify the feasibility of the proposed strategy,and the effectiveness and compare the effectiveness and network overhead with other load balancing strategies in six typical cloud data center scenarios.The experimental results show that the proposed strategy can reduce the resource consumption and show better load balancing effect.