Design And Implementation Of Traffic Scheduling System Based On Ant Colony Algorithm

Under the background of booming digital economy,some large business websites and financial service systems are often faced with high concurrency and high traffic service scenarios.With the emergence of load balancing technology,the system can still run stably in this severe environment.There are many load balancing service firms in the market,including hardware firms and cloud service providers.Professional load balancing hardware products such as F5 have strong performance and steady running,but they are also expensive.Software load balancing,which has a slightly lower performance,can not only help users cope with bad service scenarios,but also help reduce costs.LVS and Nginx are representative projects in software load balancing.They work at different levels of OSI model and can cooperate with each other to provide load balancing services for the system.In order to ensure the good running,enterprises often need to spend a lot of engineers to deploy,manage and maintain these software,which is very costly.The traffic scheduling system is designed to further reduce the consumption of these resources.It through to the centralized management of the load balancing software to achieve the aim of traffic scheduling server in the cluster.And the system in the form of API provides services,can be very convenient access management platform.At the design level,traffic dispatching system is divided into three subsystems,followed by cluster management subsystem,open data subsystem and monitoring alarm subsystem.These three systems work together to provide users with a complete traffic scheduling service.The following are the functions of each subsystem:(1)Cluster management subsystem:the system includes load cluster control module,configuration update module and task center module,which provides the load of cluster management function,monitoring management function and set listening and forwarding rule function,etc.Users will manage their own load clusters through this system.(2)Open data subsystem:the system includes data collection module,data storage module and data open interface,which provides load cluster related data query function and calculation function.Users can through the operation of the subsystems to grasp own load cluster in the relevant data and carry on the related data statistics.(3)Monitoring alarm subsystem:the system includes monitoring module,timing query module,alarm module,and a monitoring panel,when users submit a task to cluster management system,the system will provide monitoring services for the execution of tasks and alarm services for abnormal condition.In order to optimize the cluster scheduling effect,this paper,inspired by ant colony algorithm,integrates the self-developed load strategy in the system.Mapping of the various parameters by using a cluster resource scheduling in the algorithm,the resource scheduling problem of sub-cluster is optimized.This strategy cooperates with the upstream module in Nginx,and improves the quality of service on the basis of the original system.