Design And Implementation Of Load Balancing Scheduling System Based On Server Cluster

With the rapid development of artificial intelligence technology,intelligent video surveillance system has been applied to various fields.The system uses artificial intelligence,computer vision and other technologies to process the video information collected by the camera.The system provides portrait analysis service with face recognition algorithms.The primary functions of the system are portrait comparison,portrait retrieval,real-time video surveillance,etc.However,as the number of videos increases.the performance of a single server cannot meet the needs of many video real-time analysis and processing.Therefore,it is necessary to deploy the system to the clusters and schedule the task to server node to execute.So,this thesis implements a load balancing scheduling system based on server clusters to schedule tasks to multiple servers.Utilize the resources of the server cluster to improve the processing capacity of the system.The load balancing scheduling system based on server cluster is designed and implemented by this thesis.First,this thesis analyzes the functional requirements and non-functional requirements of the system.Then this thesis designs the outline of the system.The system adopts a shared state scheduling architecture,divided the system into software level and functional module architecture.Then this thesis designs and describes the system interface.Subsequently,the system is designed and implemented in detail.The main functions and business processes of each module are designed in detail.Then develop and implement the system in Go programming language,use gRPC communication technology as the main communication method between the system,and use RabbitMQ technology to achieve decoupling of the system.use Consul manages the system services deployed in the clusters.Finally,test the sy-stem,execute the test to verify the main functions,interfaces and performance of the system,and analyze the test results after the test is completed.Finally,from the analysis of test results and system execution effect,the function test and non-function test of the system has passed,and the task scheduling business content of the intelligent video surveillance system can be completed.At the same time,the scheduling system decouples resource allocation,task scheduling and execution from the intelligent video surveillance system,getting an independent operation scheduling system,which can be extended to other upper-level business platforms to provide task scheduling services.