Reliability analysis of IoT systems with competitions from cascading probabilistic function dependence

• Published in: Reliability engineering & system safety Vol. 198; pp. 106812 - 15
• Main Authors: Zhao, Guilin, Xing, Liudong
• Format: Journal Article
• Language: English
• Published: Barking Elsevier Ltd 01.06.2020; Elsevier BV
• Subjects: Cascading effect; Combinatorial analysis; Combinatorial model; Competing failure; Dependence; Failure; Failure analysis; Failure times; Internet of Things; Methodology; Probabilistic function dependence; Propagation; Random failure propagation time; Reliability; Reliability analysis; Reliability engineering; Smart buildings; Smart home; System reliability; Systems analysis; Wireless communications; Cascading effect; Competing failure; Random failure propagation time; Combinatorial model; Smart home; Probabilistic function dependence; Reliability
• ISSN: 0951-8320; 1879-0836
• DOI: 10.1016/j.ress.2020.106812

•Failure competition caused by probabilistic function dependence is modeled.•A combinatorial hierarchical reliability analysis methodology is suggested.•Cascading effects and random failure propagation time are considered.•A case study of an example smart home sensor system is provided. This paper models an Internet of Things (IoT) system subject to probabilistic functional dependence (PFD), which takes place when the failure of one component (trigger) causes other system components (dependent components) to become isolated or inaccessible with certain probabilities. Competitions in the time domain between the trigger failure and dependent components’ propagated failures may lead to dramatically different system statuses. The PFD behavior abounds in IoT systems involving relayed wireless communications (e.g., body sensor systems, smart homes). The existing works assume single-level PFD and zero failure propagation time. In practice, however cascading PFD may take place in IoT systems with multi-level configurations. Due to the cascading effect, a component may play a dual role as both a trigger and a dependent component simultaneously, creating correlations among different PFD groups. In addition, the failure originating from a component may take some random time to become effective. In this work, we make contributions by proposing a combinatorial hierarchical methodology for reliability analysis of IoT systems subject to cascading PFD and random failure propagation time. The suggested methodology is applicable to arbitrary types of failure time and propagation time distributions. An example smart home sensor system is analyzed to demonstrate the proposed methodology.