Design And Implementation Of Load Balancing System In The Cloud Platform

As a kind of distributed computing model, cloud computing is a business evolution of computational grid. In cloud computing, improving resource utilization rate and load balance degree is an eternal topic. In order to achieve this goal, this paper designs and implements a load balancing system in cloud platformIn view of the position of load balance subsystem in the cloud platform, this paper presents functional requirements, makes overall design, and analyzes three major functional modules, such as load collection and monitoring module, strategy controller, load adjusting module. Module function, design scheme, key issues and implementation plan are described in each module. Load collection and monitoring module is responsible for data collecting and processing. Strategy controller is designed for load strategy implementation and dynamic adjustment. Load adjustment module focuses on overload judgment and migration.To better meet the system requirements, strategy controllerdesigns a new dynamic load balancing algorithm—Exponential Smoothing forecast Based Weighted Least-Connection,which is named ESBWLC. Load adjusting module achieves overload control function through the optimized combination of simulated annealing and genetic algorithm. According to efficient load balancing strategy and overload migration mechanism, the load balancing subsystem achieves load balancing of cloud platform to a great extentFinally, the load balancing system is tested both on functional testing, and performance testing of ESBWLC and improved simulated annealing genetic algorithm, to verify the feasibilityof system design. Then according to the deficiencies in development, the paper points out what problems need to be improved the next stage.