Design And Implementation Of Kubernetes-Oriented Container Network Management System

With the continuous development of technologies such as 5G and the Internet of Things,the amount of data continues to increase,and a large number of computing tasks will be generated in many scenarios.However,cloud computing cannot meet the computing needs in all scenarios,so enterprises have begun to turn their attention to edge-cloud collaborative computing to speed up the execution of computing tasks.Edge-cloud collaboration technology builds a cluster of computing nodes.The cluster divides nodes into two types:edge nodes and cloud nodes according to the computing resources they have.The actual deployment of nodes is often done based on container technology.The management of the container life cycle and the scheduling of computing tasks are currently mainly completed by Kubernetes as the task scheduling system.However,Kubernetes firstly lacks the ability to manage the container network and cannot monitor the network conditions of the entire cluster;secondly,when scheduling tasks,the scheduling strategy lacks the consideration of the container network factor;at the same time,Kubernetes does not have the ability to optimize the container network.When the number of computing tasks is huge,network congestion or even failure may occur.In order to solve the problem of Kubernetes’ insufficiency of container network management capability,this topic designs and implements a Kubernetes-oriented container network management system.The system realizes real-time monitoring of container network data including link data and node data;secondly,it realizes communication with Kubernetes management nodes.On the one hand,it provides network data support for Kubernetes task scheduling,and on the other hand,Kubernetes management nodes generate scheduling After the results,the container network system will receive the batch of task scheduling results and convert them into specific network routing strategies.This topic firstly investigates the research status of container network management systems at home and abroad,and analyzes the advantages and disadvantages of the existing systems,and then analyzes the requirements of the system for the edge-cloud collaboration scenario,designs the overall architecture and business processes,and then analyzes each The functional modules are designed and implemented in detail.In particular,in order to ensure the load balancing of the container network,a load balancing routing algorithm based on traffic prediction is proposed.The algorithm is divided into two steps:the first step is to predict future network traffic based on historical data,and the second step is to use a multi-path routing algorithm based on the predicted traffic to generate a routing strategy with load balancing characteristics.Then,through the system’s flow table delivery module,the generated routing policy is converted into a flow table and delivered to the corresponding switch,thereby realizing load balancing routing.Finally,functional tests and performance tests are done on the developed system.The results show that each function meets the expected effect,the system has stability and high availability,and can well manage the container network cluster.