Research And Optimization Of Jenkins' Enterprise Distributed Continuous Integration Platform

With the development of computer technology and network technology,distributed system has been widely used.In the area of continuous integration,many companies turn from traditional single-node continuous integration to distributed continuous integration.But,the complexity of the company's projects gradually Increase and the demand of software is becoming more and more complex,which cause distributed continuous integration system node load sharply,even the processing power of the single node is getting stronger,but it can not meet the demand.How to effectively monitor and manage distributed nodes,and the realization of distributed continuous integration system load balancing is a matter of concern.Therefore,this thesis attempts to design a set of distributed continuous integration platform optimization scheme by means of technology to realize the effective monitoring and load balancing and to improve the efficiency of the continuous integration platform,which is based on the usability research of Jenkins(an open source automation service)enterprise continuous integration platform.Firstly,this thesis introduces the jenkins distributed continuous integration platform technology,and then analyzes the commonly used distributed node monitoring and load balancing scheme under the traditional distributed continuous integration platform,which points out the advantages and disadvantages of scheme,and puts forward node monitoring and load balancing optimization scheme based the jenkins distributed continuous integration platform.The communication channel is established based on JNLP and TCP/IP protocol for the node monitoring optimization scheme,and the node state is accurately reflected by the periodic acquisition node data and MA(moving average).For the load balancing optimization scheme,the load capacity of each node is calculated by dynamic weight polling scheduling algorithm,and the task is polled according to it's current actual load capacity.The schemes of node monitoring and load balancing optimization are designed and implemented,which are divided into control layer,the monitoring layer,the data layer and the core layer.The control layer is mainly used to deal with the interaction between the platform UI and the background.The data layer is responsible for storing the data information on the distributed platform nodes.The monitoring layer is responsible for monitoring the status of each node on the distributed platform.The core layer belongs to the platform load balancing optimization system.At the end of this thesis,the node monitoring and load balancing optimization model are tested and the result shows the validity and feasibility of the two optimization systems,which improve the efficiency of enterprise distributed continuous integration platform.