Design And Implementation Of Distributed Scheduling Platform For System Resource Based On Container

With the development of the Internet and Internet of Things,new business scenarios emerge,and cloud computing and edge computing technologies come into being.Cloud computing provides users with a shortcut to access massive computing resources on the Internet.For ordinary users,they don't need to understand the specific implementation of the cloud cluster,and they don't need to have a professional computer background knowledge either,so they can easily access a large number of cloud resources.On the other hand,the shortcomings of cloud computing technology,such as high latency and high energy consumption,limit its use scenarios.Edge computing deploys resources closer to users,which has the advantages of lower network latency and lower broadband demand,and can adapt to the system requirements of applications such as smart home and digital healthcare.The collaboration between cloud computing and edge computing can improve user experience and system resource utilization efficiency.This thesis uses Docker container to replace traditional Virtual Machine,and then combines them with the scenario of the collaboration between cloud computing and edge computing,thus realizes the management of system resources of large-scale distributed cluster.The main research work of this thesis is as follows:1.In this thesis,a container-based distributed orchestration system of resource and application is designed and implemented.The system can realize the management,deployment and distribution of hosts and containers in the cluster.2.For the scenario of the collaboration between cloud computing and edge comput-ing,we design and implement the ant colony algorithm to find the optimal task scheduling solution by imitating the way ants forage,and complete the management system resources such as GPU.The ant colony algorithm can improve the efficiency of task distribution and the efficiency of system resource usage.3.The implementation of the failure recovery mechanism uses a dual-system hot standby technology to ensure the high availability of the entire cluster.The system uses heartbeat messages to achieve real-time monitoring of the status of nodes.The system uses the synchronous backup and delta update features of logs to synchronize critical logs to the backup machine,ensuring the consistency of the hot backup data.4.Function and performance tests.After complete tests on functionality and performance,the system is capable of managing containers and clusters.The system can realize the function of task deployment and container migration.In terms of performance testing,the comparison and analysis of system response time and throughput of different systems shows that this system has features about high performance,fault tolerance and high availability.