Leveraging Observability Via Tracing For Software Regression Detection and Root Cause Analysis

通过跟踪利用可观察性进行软件回归检测和根本原因分析

• 批准号: 572127-2022
• 负责人: EzzatiJivan, NaserNN
• 金额: $ 1.46万
• 依托单位: Brock University
• 依托单位国家: 加拿大
• 项目类别: Alliance Grants
• 财政年份: 2022
• 资助国家: 加拿大
• 起止时间: 2022-01-01 至 2023-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=749667
• 关键词: Leveraging; Observability; Via; Tracing; Software

Execution tracing can provide valuable insights into the runtime behaviour of software systems. It can be used to detect, debug, analyze, and resolve a number of software problems, including memory leaks, latencies, network congestion, and security flaws. However, tracing can seriously degrade the system performance as it is difficult to produce a golden set of events possessing an optimal balance of minimal overhead, overlap, and redundancy while being precise enough to cover a wide range of problems. Providing software developers and testers with a small subset of key trace events and metrics can improve the design, testing, and debugging processes. Toward solving this challenge, Ciena, a telecom network equipment and software services supplier company, has partnered with a research group at Brock University to build and promote algorithms and tools for leveraging software observability through a cost-aware adaptive tracing model for software regression detection and analysis. This includes two sub-objectives. The first is to devise algorithms, strategies, and tools for the self-adaptive tracing of software systems. The second sub-objective focuses on root-cause analyses of latency changes detected by regression tests. This will include correlating changes in software behaviour with code blocks using traces collected at various levels of the system.The tools and methods provided by this research could be utilized by a wide range of companies in Canada to enhance the overall performance of their systems while reducing the time and costs of maintaining, monitoring, debugging, and tuning them. This is especially true for smaller companies lacking the resources to carry out a detailed analysis of their software solutions. This project will be primarily open-source and most results, tools, and contributions will be published and applied to open-source systems. Thus, apart from software companies, this work would directly benefit many industrial venues in Canada, including IT and phone companies, transportation, energy, and the research community in software engineering and telecommunications.

执行跟踪可以为软件系统的运行时行为提供有价值的见解。它可用于检测、调试、分析和解决许多软件问题，包括内存泄漏、延迟、网络拥塞和安全缺陷。然而，跟踪会严重降低系统性能，因为很难生成一组具有最小开销、重叠和冗余的最佳平衡的黄金事件，同时又足够精确以涵盖广泛的问题。为软件开发人员和测试人员提供关键跟踪事件和度量的一个小子集可以改进设计、测试和调试过程。为了解决这一挑战，电信网络设备和软件服务供应商Ciena公司与Brock大学的一个研究小组合作，建立并推广算法和工具，通过成本意识自适应跟踪模型来利用软件可观察性，用于软件回归检测和分析。这包括两个子目标。首先是为软件系统的自适应跟踪设计算法、策略和工具。第二个子目标侧重于对回归测试检测到的潜伏期变化进行根本原因分析。这将包括使用在系统的各个级别收集的跟踪将软件行为中的变化与代码块关联起来。本研究提供的工具和方法可以被加拿大的许多公司用于增强其系统的整体性能，同时减少维护、监视、调试和调优系统的时间和成本。对于缺乏资源对其软件解决方案进行详细分析的小公司来说尤其如此。这个项目将主要是开源的，大多数结果、工具和贡献将被发布并应用于开源系统。因此，除了软件公司之外，这项工作还将直接使加拿大的许多工业场所受益，包括IT和电话公司、交通、能源以及软件工程和电信的研究界。