标题： 从业者视角下的软件架构新兴趋势：五年回顾 显示英文标题

标题： Emerging Trends in Software Architecture from the Practitioners Perspective: A Five Year Review

摘要： 软件架构在软件系统的设计、开发和维护中起着核心作用。 随着云计算、微服务和容器的兴起，架构实践变得更加多样化。 理解这些变化至关重要。 本研究分析了五年内八个领先行业会议中的软件架构趋势。 我们通过分析顶级从业者会议的演讲来研究软件架构的演变，重点关注推动技术采用的动机和背景。 我们分析了来自八个主要行业会议的5,677个演讲，使用大型语言模型和专家验证来提取技术、其目的和使用场景。 我们还探讨了技术如何相互关联并融入DevOps和部署管道中。 在450种技术中，Kubernetes、云原生、无服务器和容器在频率和中心性方面占主导地位。 从业者主要展示与部署、通信、人工智能和可观测性相关的技术。 我们识别出五个技术社区，涵盖自动化、协调、云人工智能、监控和云边缘。 大多数技术跨越多个DevOps阶段并支持混合部署。 我们的研究表明，像Kubernetes和无服务器这样的少数核心技术主导了当代软件架构实践。 这些技术主要应用于后期DevOps阶段，对早期阶段如规划和编码的关注有限。 我们还展示了从业者如何根据目的和背景来描述技术，反映了行业优先级的变化。 最后，我们观察到只有研究才能提供对架构设计、质量和演化的更全面的视角。 显示英文摘要

摘要： Software architecture plays a central role in the design, development, and maintenance of software systems. With the rise of cloud computing, microservices, and containers, architectural practices have diversified. Understanding these shifts is vital. This study analyzes software architecture trends across eight leading industry conferences over five years. We investigate the evolution of software architecture by analyzing talks from top practitioner conferences, focusing on the motivations and contexts driving technology adoption. We analyzed 5,677 talks from eight major industry conferences, using large language models and expert validation to extract technologies, their purposes, and usage contexts. We also explored how technologies interrelate and fit within DevOps and deployment pipelines. Among 450 technologies, Kubernetes, Cloud Native, Serverless, and Containers dominate by frequency and centrality. Practitioners present technology mainly related to deployment, communication, AI, and observability. We identify five technology communities covering automation, coordination, cloud AI, monitoring, and cloud-edge. Most technologies span multiple DevOps stages and support hybrid deployment. Our study reveals that a few core technologies, like Kubernetes and Serverless, dominate the contemporary software architecture practice. These are mainly applied in later DevOps stages, with limited focus on early phases like planning and coding. We also show how practitioners frame technologies by purpose and context, reflecting evolving industry priorities. Finally, we observe how only research can provide a more holistic lens on architectural design, quality, and evolution.