Font Size: a A A

Research On Load Balancing Of Web Server Cluster Based On Genetic Algorithm

Posted on:2018-04-18Degree:MasterType:Thesis
Country:ChinaCandidate:X WeiFull Text:PDF
GTID:2348330512471492Subject:Master of Engineering
Abstract/Summary:PDF Full Text Request
Load scheduling mechanism selects the appropriate nodes in the network to allocate the load by minimizing the response time and maximizing the utilization ratio of the nodes.However,the construction of load balancing is a Non-deterministic Polynomial complete problem.In the field of load balancing,service cluster has made many research achievements.Although the traditional scheduling algorithm is simple,but a realistic environment combined of a number of complex factors intersect and constraint combination.These point-to-point algorithm will produce a large number of errors tend to peak in the search process,thus affecting the determination of optimal distribution combination.The genetic algorithm adapts to many fields,which is a high degree of parallelism global search algorithm.In the research of load balancing based on genetic algorithm,fitness function is the standard to evaluate the request allocation combination.In this paper,the mean-variance model is introduced to improve the fitness function of load balancing algorithm.The investment selection model mean-variance is used to set the weight of node utilization in service cluster.Through lagrange multiplier to get the optimal weight,which takes the minimum task completion time as a condition.This method improves the accuracy and effectiveness of task scheduling in service cluster.For the load balancing of service cluster,the main contributions of this paper include three aspects:(1)In the coding design of genetic algorithm,this paper proposes the design of spatial candidate solution coding used three-dimensional decimal.In three-dimensional decimal coding,each solution is encoded as a decimal array represented by three attributes.The string in each array has a fixed size quantization value,quantization load is quantified according to the proportion of tasks in the Kennedy space center server log file.Each node in the search space has a string representation,and the string it represents is unique.So the three-dimensional decimal coding design improves the accuracy of genetic scheduling.(2)By studying and analyzing the load balancing method based on genetic algorithm,as the fitness function is used to judge the accuracy of distribution combination,this paper introduces mean-variance model to construct fitness function.Minimizing the response time by portfolio selection model mean-variance,the weight of each server resource utilization can be obtained,thus we can get the optimal allocation combination.By changing the simple superposition mode of node utilization in traditional fitness function,the execution time and node utilization are improved,and the efficiency of load balancing strategy is improved.(3)In this paper,the Mean-Variance model is used to construct the fitness function,which makes the execution time and the node utilization efficiency of the load balancing algorithm be validated.However,some unbalanced service runs still occur in the scheduling algorithm.To solve this problem,an adaptive threshold algorithm is proposed to obtain the average load of a single server according to the load of the current system load and the new scheduling task.At the same time,the system sets the thresholds of heavy and light.If the server's load value exceeds the heavy threshold value or below the light threshold,it will not function properly and will be identified as an unacceptable allocation task server.The adaptive threshold policy of the acceptable allocation server adds flexibility to the load balancing mechanism and adds an important safeguard to the load balancing system.In this paper,a series of experiments are designed and implemented,and compared with other models under different service environment and parameter design.The simulation results show that the optimized load balancing algorithm proposed in this paper makes the service cluster get better balance in terms of node utilization and response time.
Keywords/Search Tags:Load balancing scheduling, genetic algorithm, Mean-Variance model, adaptive threshold algorithm
PDF Full Text Request
Related items