联系我们

论文查询结果	返回搜索

论文编号：	15739
作者编号：	2320234124
上传时间：	2025/12/9 18:05:25
中文题目：	基于DevOps的A公司软件研发质量管理改进研究
英文题目：	Research on the Improvement of Software R & D Quality Management in Company A Based on DevOps
指导老师：	石鉴
中文关键字：	DevOps；软件质量管理；工业互联网；流程改进；合规管理
英文关键字：	DevOps; Software Quality Management; Industrial Internet; Process Improvement; Compliance Management
中文摘要：	在工业互联网与能源数字化深度融合的背景下，工业软件研发面临“高可靠性、强合规、协作低效”三重行业共性挑战：工业系统停机损失巨大，要求软件可用性达99.99%以上；IEC 62443、等保2.0等标准强制研发全流程嵌入安全审计；传统职能架构导致研发、测试、运维信息孤岛，引发交付延迟与质量波动。A公司作为全球能源管理企业在华研发中心，聚焦工业互联网平台与能源数字化系统，其研发流程中“需求到上线周期长、缺陷逃逸率高、工具链割裂”等问题尤为突出。探索基于DevOps的质量管理改进路径，既能帮助A公司突破发展瓶颈，也能为同行业提供可复用的实践范式，兼具实践价值与理论参考意义。 本研究采用“诊断-设计-落地-验证”的闭环研究思路，整合多方法开展分析。首先通过3个月实地调研，收集A公司2022-2024年12个核心项目的定量数据（Jira缺陷记录、SonarQube报告等），并对86名跨职能人员开展半结构化访谈，结合质性与统计分析，识别“组织僵化、流程不标准、工具链整合不足”三大症结；其次结合A公司核心特征，从三维度设计方案：技术上构建“GitHub Actions+Docker+K8s”CI/CD流水线，嵌入代码质量门禁与监控体系；流程上推行“测试左移”与“分级发布”，将合规转化为可执行节点；组织上组建跨职能团队，建立质量KPI考核体系；最后通过“试点-推广”分阶段落地，借助t检验、相关性分析等方法验证成效，并引入对标企业数据排除外部干扰。 研究表明，改进方案有效解决A公司核心问题，实现质量、效率、合规三维度协同提升。核心创新体现在三方面：一是构建“约束适配型DevOps方法论”，以企业约束为轴设计可裁剪模板；二是提炼9项可推广措施，明确复用条件与操作要点；三是界定3类不适用场景（轻量软件、需求极不稳定项目、小型团队），填补DevOps在工业软件领域适用边界的研究空白。本研究为A公司提供落地路径，也丰富“DevOps和工业软件”理论体系，可为同行业质量管理转型提供参考。
英文摘要：	Amid deep integration between the Industrial Internet and energy digitalization, industrial software development faces three major industry-wide challenges: high reliability requirements, strict compliance mandates, and inefficient collaboration. Industrial system downtime incurs substantial losses, demanding software availability of over 99.99%; standards such as IEC 62443 and Cybersecurity Classified Protection 2.0 mandate the integration of security auditing throughout the entire R&D process; and traditional functional organizational structures create information silos among R&D, testing, and operations teams, leading to delayed deliveries and quality fluctuations. As the China-based R&D center of a global energy management enterprise, Company A focuses on Industrial Internet platforms and energy digitalization systems, with prominent issues in its R&D processes, including prolonged lead times from requirement definition to launch, high defect escape rates, and fragmented toolchains. Exploring DevOps-based quality management improvement pathways will not only help Company A break through its development bottlenecks but also provide reusable practical models for the industry, holding both practical and theoretical value. This thesis employs a closed-loop research framework of "diagnosis-design-implementation-verification" and integrates multiple methods for analysis. First, through three months of on-site research, quantitative data (including Jira defect records and SonarQube reports) from 12 core projects of Company A between 2022 and 2024 were collected, and semi-structured interviews were conducted with 86 cross-functional personnel. Combining qualitative and statistical analysis, three core issues were identified: rigid organizational structure, non-standardized processes, and insufficient toolchain integration. Second, tailored improvement plans were designed from three dimensions based on Company A’s core characteristics: technically, an end-to-end CI/CD pipeline (GitHub Actions+Docker+K8s) was built, integrated with code quality gates and monitoring systems; process-wise, "shift-left testing" (embedding test case design in the requirement phase) and "phased release" (canary releases for core functions, blue-green deployments for non-core functions) were implemented, translating compliance requirements into actionable checkpoints; organizationally, cross-functional teams (development+security+operations+SRE+compliance) were established, along with a quality KPI assessment system. Finally, the plan was implemented in phases ("pilot promotion"), with effectiveness verified using methods such as t-tests and correlation analysis, and data from benchmark enterprises were introduced to rule out external interference. The research shows that the improvement plan effectively addresses Company A’s core issues, achieving coordinated improvements in quality, efficiency, and compliance. Three core innovations are highlighted: first, the development of a "constraint-adapted DevOps methodology," with customizable templates designed around enterprise constraints; second, the extraction of 9 scalable measures, with clear reuse conditions and operational guidelines; third, the identification of 3 scenarios where the methodology is inapplicable (lightweight software, projects with highly unstable requirements, small teams), filling the research gap in defining the application boundaries of DevOps in industrial software. This study provides an actionable pathway for Company A, enriches the theoretical system of "DevOps+Industrial Software," and offers references for quality management transformation in the industry.
查看全文：	预览  下载(下载需要进行登录)