Design And Implementation Of Load Balancer For High Concurrent Web Server Cluster

In the Internet era,personal computers and handheld devices are becoming more and more popular,which makes the demand for information processing and interaction between devices increasing.In order to meet this growing demand,we should not only consider the security and reliability of the service,but also consider the timely response of the service,so this puts forward higher and higher requirements for the existing servers.It is particularly important to study load balancing technology.As a high performance HTTP and reverse proxy server,Nginx has been favored by many researchers in order to solve the problem of server high load.This thesis first introduces the related technologies of Web server cluster and Nginx,focusing on the analysis of the Nginx system module architecture and the working principle of the process,as well as the advantages and disadvantages of Nginx's weighted polling algorithm,the minimum connection number,the IP_HASH algorithm and the URL_HASH algorithm.In view of their shortcomings,based on the original weighted polling algorithm of Nginx,this thesis proposes that the resource utilization and response time of a large number of cluster server nodes in the past period be taken as parameters,then the proportion of different parameters is analyzed according to the analytic hierarchy process,and finally a comprehensive load weight calculation function is constructed.To construct the comprehensive load weight matrix,the average comprehensive load weight of the server cluster is taken as the threshold value as the first condition to start and modify the effective weight.The server load information is dynamically introduced into the process of Nginx original weighted polling to modify the weights of the servers,thus realizing the dynamic adaptive load balancing strategy and improving the average response time and throughput of the system,enhances system performance.Finally,a comparative experiment is designed to test the performance of Nginx's original weighted polling strategy and the dynamic strategy proposed in this thesis.The experimental results show that the proposed load balancing strategy has obvious advantages in the case of high load,and the performance of the two strategies has little difference in the case of low load.The dynamic load balancing strategy proposed in this thesis has the following characteristics:(1)The backend server CPU,memory,disk I/O of all nodes,network bandwidth load rate and response time performance is collected,through the analysis of user request type,and uses the analytic hierarchy process to get the weight vector of load function,and finally build the comprehensive load weight evaluation function,calculated the performance of the node weight,compared with the empirical assignment,this calculation method is more comprehensive and effective.(2)Dynamically adjust the weight,when the judge to modify weight,weight in order to avoid the frequent modification,the system will first set a threshold value,whether to need to modify the last time the server node processing requests of comprehensive load weight average,as the critical value of start modify the weighting process,then the system will according to gather all the back-end recently for a period of time within the cluster server cluster comprehensive load of each node in the process the request in the weights of building matrix,calculates the variance,it reflects the node load,and the variance of node weights dynamically updated.