The Design Of Internet Of Things Platform Based On Microservicesand Container Deployment Research

The Internet of things has shown huge market potential and broad development prospects,and the Internet of things platform is a critical part in the industrial structure of the Internet of things.As the backbone of the Internet of things,it not only faces the connection of thousands of physical devices,but also needs to meet the service demand of users for high availability of applications.How to design and implement a low-coupling,easy to expand,scalable,and highly available Internet of things platform based on the rapidly developing virtualization Docker container technology has become one of the most urgent problems to be solved.This thesis firstly studies the design of microservice architecture and some key technologies applied.Based on the design principle of low coupling and easy extension,the system is divided into services through business functions,and the distributed service framework Spring Cloud is adopted to develop micro-service modules.To ensure the high availability of micro service module,using Consul to manage service providers and service consumers.In order to track the Spring Cloud service invocation chains and ensure the stability of the system,Zipkin is used for real-time monitoring.Then the system architecture of the whole Internet of things platform is designed according to the specific business requirements,and the design scheme of each part of the system is provided.In order to improve the extensibility of the program and the convenience of deployment,Kubernetes and Docker container technology are adopted to deploy the service.Based on the original RPC communication framework of Spring Cloud,Thrift was used to replace the use of Feign to improve the performance of the system under high concurrency.In view of the defect that the original scaling index of Kubernetes has only CPU utilization,increase Kubernetes to expand and shrink according to QPS.Finally,the whole system application server system function test and performance test is done.The advantages of Thrift in throughout are demonstrated by comparing the test result of RPC communication framework of Spring Cloud with Thrift.The advantage of container deployment in availability and reliability are shown by comparing the deployed Kubernetes services with the standalone service.