Research And Design Of Task Scheduling System Based On Shared Computing Resources

In the era of data explosion,with the rapid development of Internet information technology and the gradual expansion of enterprise scale,most enterprises need to have their own task scheduling systems to handle various complex data business work.Most of them are timed tasks,such as the daily work report of the company,the statistics of employee attendance and so on.In the past,task requirements were relatively simple,and scheduling could be completed by deploying a single server node.However,as the application scenarios of timed tasks become more and more complex,enterprises have higher and higher requirements for task scheduling systems.Single-server scheduling faces large-scale tasks.The processing efficiency of task requirements is slow,the single point of failure is prone to occur,and the problem of resource contention is serious.Therefore,a multi-server distributed task scheduling system is required.The question that follows is how the cluster distributed task scheduling system should implement the online scheduling of tasks and the trigger execution of tasks and the error correction of tasks and the security management of the system.In order to solve the above problems,this thesis researches and designs a task scheduling system based on shared computing resources,deploys tasks through executor clusters,solves the problem of resource contention,at the same time,the two functions of task scheduling and task execution are separated to improve the coupling between system modules.The system uses Java programming language for development,the task scheduling framework uses quartz,the back-end development uses spring boot technology for integrated development,and the My SQL database is used as the data service registration center of the entire system,and the front-end interface is also configured to make it more convenient for users for system management.This system mainly includes three parts,namely dispatch center,executor cluster and database service registration center.The scheduling center is mainly responsible for the online scheduling of tasks.The main modules include task management and executor management and log management and other management modules,which can add,delete,modify and check tasks,add,delete,modify and check executors,log management,and online execution of tasks and other functions.The executor cluster is mainly responsible for executing the business logic of specific tasks,which meets the above requirements of decoupling the two functions of task scheduling and task execution.The database service registry is mainly responsible for processing system data,including user data,executor registration information,system operation logs and so on.And through the free marker framework engine combined with j Query,a front-end operation control interface is designed.Compared with the traditional task scheduling system,it also adds task failure alarms and running report monitoring functions.Finally,the system test was carried out.According to the results,it was concluded that the system basically achieved the required functions,and also achieved the goals in throughput and response speed.Compared with the existing task scheduling framework,the system has It has obvious advantages in deployment and task execution management functions.The task scheduling system designed in this thesis is developed with a high-performance development framework and adds some functions.Compared with the traditional task scheduling framework,it has higher performance and stability.At the same time,it decouples the scheduling and execution functions to improve the system’s performance Low coupling,spring boot supports packaged deployment to make the system have better usability and scalability,so as to meet the data business scheduling requirements in actual scenarios.