Observing and controlling the dependability of service provision of complex IoT systems is challenging. In practice, many organizations struggle to derive consumer needs related to quality and to observe and quantify the service provision in the context of the dynamic behavior of a complex distributed system. In this paper, we present an approach to define and evaluate the dependability of complex IoT systems. Our approach is an adaptation of the ISO/IEC 25040, an international standard for the evaluation process for system and software quality, which is part of the systems and software quality requirements and evaluation (SQuaRE) series. Our approach was designed and evaluated with action research in an industrial study at Robert Bosch GmbH. Based on the framework of the SQuaRE series, we integrated different elements of site reliability engineering (SRE) and combined them with distributed tracing as a promising measurement method. Our approach introduces the IoT transaction concept to reduce modeling and observation efforts while increasing operationalization to measure performance against dependability targets. Our adaption was effectively applied, consumer-centricity along different system stakeholders were enhanced, and negative consequences of organizational silos were reduced. This has improved the dependability evaluation of service provision to enable fast feedback cycles for service performance control and improvement.

观察和控制复杂物联网系统服务提供的可靠性具有挑战性。实际上，许多组织都在努力寻求与质量相关的消费者需求，并在复杂的分布式系统的动态行为的背景下观察和量化服务提供。在本文中，我们提出了一种定义和评估复杂物联网系统可靠性的方法。我们的方法是对ISO / IEC 25040（适用于系统和软件质量评估过程的国际标准）的改编，它是系统和软件质量要求与评估（SQuaRE）系列的一部分。我们的方法是在罗伯特·博世（Robert Bosch GmbH）的一项工业研究中通过行动研究进行设计和评估的。基于SQuaRE系列的框架，我们整合了站点可靠性工程（SRE）的不同元素，并将它们与分布式跟踪相结合，成为一种很有前途的测量方法。我们的方法引入了IoT交易概念，以减少建模和观察工作，同时增加可操作性以根据可靠性目标衡量性能。有效地应用了我们的适应方法，增强了以不同系统利益相关者为中心的以消费者为中心的情况，并减少了组织孤岛的负面影响。这改善了服务提供的可靠性评估，以实现用于服务性能控制和改进的快速反馈周期。我们的方法引入了IoT交易概念，以减少建模和观察工作，同时增加可操作性以根据可靠性目标衡量性能。有效地应用了我们的适应方法，增强了以不同系统利益相关者为中心的以消费者为中心的情况，并减少了组织孤岛的负面影响。这改善了服务提供的可靠性评估，以实现用于服务性能控制和改进的快速反馈周期。我们的方法引入了IoT交易概念，以减少建模和观察工作，同时增加可操作性以根据可靠性目标衡量性能。有效地应用了我们的适应方法，增强了以不同系统利益相关者为中心的以消费者为中心的情况，并减少了组织孤岛的负面影响。这改善了服务提供的可靠性评估，以实现用于服务性能控制和改进的快速反馈周期。