Resiliency-aware analysis of complex IoT process chains

Abstract

Resiliency-aware Internet of Thing (IoT) can protect sensitive information depending on the ongoing agenda in smart cities in case of failure in critical processes. Guaranteeing resiliency and availability are two critical issues in making safe real-world IoT scenarios. Existing strategies behave differently and most of them fail to pick up all the significant aspects of resiliency-aware IoT in a dynamic system. This paper finds critical activities in IoT process chains in real domains of IoT according to the proposed parameter-based greedy strategy. It analyzes the resiliency and availability of the chains under the failure of these nodes. In this paper, resiliency is measured as the ratio of performance and availability of an activity network before and after disrupting the critical activities. Resiliency is then assessed based on the four selected structural metrics for process chains(Eigenvector Centrality, Closeness Centrality, Betweenness Centrality, and Markov Chain). The results show that the decline rates of network performance, resiliency, and availability are maximum in the case of the failure of activities with a higher value of the Markov Chain (MC) metric, nearly 70%. Moreover, the loss rate of resiliency and availability is evaluated after the failure of high-ranked structural activities and is compared to detect critical activities in a process chain. A replication-based experiment, named the MC-based replication method, is finally proposed to introduce a resilience strategy across multiple situations in IoT settings. Compared to the available replication protocols, the suggested MC-based replication method improves response time and reduces resource redundancy by about 18% and 45%, respectively, for a significant scenario, by selecting efficiently replicated nodes.