Design And Implementation Of Load Balancing System Based On Server Cluster

With the development of informatization in the medical industry,the research direction of medical system server architecture design is highly valued.Most medical-related information systems do not take into account the uneven load caused by the impact of high concurrent business traffic and the inability of clusters to meet high availability issues when conducting demand analysis and construction design.As a solution to this problem,load balancing technology is a very popular and well-researched technology.It has huge advantages compared with traditional single-point server architecture technology.Firstly,in order to achieve load balancing of server clusters under high concurrency,this thesis compares the existing data balancing schemes with respect to the data skew problem in the consistent hashing method,and designs a scalable and efficient load balancing method based on improved consistent hashing.This method constructs a virtual node maximum linear value calculation model and a virtual node redundancy value calculation model,and relates the number of virtual nodes to the node performance value and the number of cluster nodes,which can prevent uneven scheduling caused by hash tilt.Secondly,aiming at the master-slave failure and session sharing problems in master-slave model,this thesis compares the existing master-slave construction scheme and session sharing scheme,then designs a scalable,backupable HA model based on the election queue and Redis.This model builds an election queue based on the election state,and at the same time decouples the business logic Session management,which can prevent the service interruption caused by the loss of cache data after the master-slave failure.Finally,this thesis constructs the system's hardware and software environment,and performs prototype implementation and system testing of the functional modules.According to the test results,the system meets the business needs of the medical system in terms of load balancing and high availability.The system has been officially launched in the hospital for one year and has been affirmed by medical staff in all departments.