面向软件缺陷预测的云原生研发效能框架研究  

Software Defect Prediction Oriented Cloud Native Based R&D Efficiency Framework Research

作　　者：刘晶 韩振东[2] 王浏明 Liu Jing;Han Zhendong;Wang Liuming(China Unicom Research Institute,Beijing 100176,China;China United Network Communications Co.,Ltd.,Beijing 100033,China;China Unicom Institute for Software Research,Beijing 100176,China)

机构地区：[1]中国联通研究院,北京100176 [2]中国联合网络通信集团有限公司,北京100033 [3]中国联通软件研究院,北京100176

出　　处：《信息通信技术》2021年第4期26-31,共6页Information and communications Technologies

摘　　要：云计算的飞速发展和上云企业软件规模、业务多样性的日益增长对资源利用率、快速迭代、持续交付和软件质量提出了越来越高的要求,同时针对研发过程缺乏客观的度量与评价体系。为解决上述问题,文章提出一种基于云原生的研发效能框架,包括基础设施层、平台能力层、研发效能平台层和应用层,对研发过程进行全面统一地管理与度量,并通过数据挖掘、机器学习方法构建分析预测模型,为研发过程的团队和项目提供画像分析、决策咨询、缺陷预测等服务。然后对研发效能框架应用层的软件缺陷预测方案进行了研究,基于研发效能框架采集的数据样本,提取软件缺陷预测的度量元指标,构建预测模型。为可能出现的缺陷代码有效分配有限的测试资源,进一步优化研发过程。The rapid development of cloud computing and the growing scale and business diversity of cloud-based enterprises have put forward higher requirements on resource utilization,rapid iteration,continuous delivery,and software quality,while there is a lack of objective metrics and evaluation systems for the R&D process.In order to solve the above problems,this paper proposes a cloud native based R&D efficiency framework,including infrastructure layer,platform capability layer,R&D efficiency platform layer and service layer.To manage and measure the R&D process comprehensively and uniformly,this paper constructs analysis and prediction models through data mining and machine learning methods to provide portrait analysis,decision consulting and defects prediction services for R&D teams and projects.And then,this paper elaborates on the software defect prediction method of the application layer of the R&D effectiveness framework.Based on the data samples collected by the framework,the measurement meta indicators of software defect prediction are extracted and the prediction model is constructed.The limited testing resources are effectively allocated for the possible defective codes to further optimize the R&D process.

关 键 词：云原生 研发效能 度量元 缺陷预测 

分 类 号：TP393.09[自动化与计算机技术—计算机应用技术]