Design And Implementation Of CDN Edge Computing Platform Based On Docker

With the rapid development of computer technology in recent years,cloud computing has experienced great progress and achieved good results.The use of cloud computing and data center technology greatly reduces the cost of using the IT infrastructure.However,there are also some shortcomings in the resource centralization of the cloud data center.After resource centralization,the end-users are far away from the data center.Long distance network transmission not only takes up a lot of network resources,but also brings the high latency to slow down the response speed of user requests and affect user experience.CDN technology can reduce the response time of user requests by caching network resources at a network edge closer to the user.The existing CDN accelerates static resources and has a very limited acceleration for dynamic content.Edge computing distributes computing power to the edge of the network to achieve dynamic content in the edge nodes.It is an important means to accelerate the dynamic content of CDN system.Up to now,the CDN edge computing platform has not yet had a mature standard.In the field of virtualization,Docker,as a major innovation,not only surpasses the common virtualization in performance,but also introduces the concept of container for the first time.Docker builds the program and the running environment together,realizes the application to " Build once,Run anywhere".If Docker is applied to the CDN edge computing platform,the complex requirements for the generalization of the edge computing platform can be met.In this thesis,a CDN edge computing platform based on Docker is designed and implemented.The platform can be divided into three subsystems:operation management,content routing and edge computing.The edge computing subsystem consists of a number of edge nodes,each of which is a container cluster that is managed by Kubernetes container orchestration tool.The content routing subsystem first routes the user request to the edge node through DNS parsing.Then,through the core module EC Proxy implemented in 'this thesis,the request is forwarded to the specific container in Kubemetes to process.The operation management subsystem is responsible for the creation,update,and unload process of the service.Besides implementing the basic working process of the system,the data synchronization mechanism between edge nodes and source nodes is also analyzed.Through reading and writing separation and master-slave synchronization strategy,data synchronization needs between them can be satisfied.Finally,the CDN edge computing testbed is built in this thesis and the experimental data are obtained through the simulation test.According to the analysis of response time and computing and storage performance,the feasibility and superiority of the CDN edge computing platform based on Docker is verified.