Research And Design Of Affinity Load Balancing Algorithm For JCF Middleware

With the development of economy,a new journey of building a strong civil aviation country has been opened.In recent years,the number of passengers travelling by air has increased dramatically.This has caused tremendous pressure on the JCF(Java Core Framework)system that handles core ticketing business.At present,fixed scale factor algorithm and minimum uncompleted transaction number algorithm has been deployed on the JCF middleware platform.However,it does not satisfy the requirements of the service affinity,it is easy to cause unreasonable request allocation,and the request cannot receive effective feedback,more effective algorithms are needed to solve the service affinity problem.First,regarding the middleware and load balancing algorithm,the research status at home and abroad is investigated,and the structure and characteristics of middleware and the differences and application scenarios of various load balancing algorithms are analyzed in detail.And studied two algorithms that have been deployed on the JCF middleware platform.Secondly,in order to solve the problem of the service affinity,and an affinity dynamic load balancing algorithm is proposed.The algorithm firstly extracts the user identification information,then uses a hash map to allocate user requests,maps the same user continuous requests to the same target server,and designs a dynamic adjustment strategy for the change of cluster size.The algorithm can effectively solve the problem of service affinity and load balance,and ensure the stability and scalability of the cluster system.Finally,an affinity load balancing optimization algorithm based on the master-slave backup mode in the group is proposed.The algorithm divides server nodes into two groups to form a master-slave backup mode within the group,and backs up each other's status and contents regularly.When the master node fails,user service requests can be smoothly switched to the slave nodes for processing.The change situation has designed the load adjustment strategy of the grouping mode,help to improve the service affinity and system stability.