Testability in an IoT system : a full life cycle perspective

Abstract

The Internet-of-Things (IoT) has brought about a major paradigm shift to traditional system analysis and design. System testing, and testability, also changed as more system elements can now be measured than in traditional systems. From a traditional perspective, testability is mostly considered functional testing and verification of devices or components in controlled environments. The massive interconnectivity of systems and devices has thus changed the current testing philosophy, which must be investigated. This study provides findings from a literature review on the definition of testability in IoT systems, taking into account that a system comprises many elements that can be individually and collectively measured, compared, and actioned. The highest impact on testability, brought about by IoT, is found in the system operations and maintenance phase, including installation, commissioning, scheduled and unscheduled maintenance, and support. In order to consolidate the massive amounts of data, an agent is proposed in a high-level orchestration layer of the system hierarchy. Both literature and a case study confirmed the value of such an agent and its purpose in ensuring system availability. The impact of IoT on testability was evidenced from a cost analysis (reduced logistics and labour) and reduced resource dependency in an IoT-enabled agricultural irrigation system. IoT had a structural effect on the system resources by changing the resource organisation from a decentralised to a centralised-hybrid structure. In addition to digital twins, a digital agent ensures testability and reflection on the real-world case study confirmed its importance. A hierarchical framework that will ensure the testability of a system in the IoT context resulted from this study.