CIRCE: Architectural Patterns for Circular and Trustworthy By-Design IoT Orchestrations

Abstract

The adoption of Internet of Things (IoT) devices, applications and services gradually transform our everyday lives. In parallel, the transition from linear to circular economic (CE) models provide an even more fertile ground for novel types of services, and the update and enrichment of legacy ones. To fully realize the potential of the interplay between IoT and CE, the design-time definition of IoT orchestrations with proven circularity properties, and the run-time management of these orchestrations based on said properties, is of paramount importance. Nevertheless, the circularity requirements and associated properties are not only difficult to achieve at the IoT orchestration design and deployment initialization phases, but also hard to prove and maintain at run-time. Motivated by this, this paper presents the CIRCE framework for circular and trustworthy by-design IoT orchestrations. The CIRCE approach leverages concepts from pattern-driven engineering, whereby patterns are used to encode proven dependencies between the Location, Condition, and Availability (LCA) properties of individual smart objects and corresponding properties of orchestrations (compositions) involving them. These are augmented by patterns encoding trustworthiness-related properties, namely Connectivity, Security, Privacy, Dependability, and Interoperability (CSPDI). Thereby, these patterns are used to generate IoT orchestrations with proven LCA and CSPDI properties, as needed, at design time. At runtime, these properties are monitored in real-time, leveraging reasoning engines deployed across system layers, triggering adaptations to return the deployed orchestration to the desired LCA and CSPDI states, when required. Details are provided on the above novel combination of IoT, CE and pattern-based engineering, along with a proposed architecture and implementation approach. Furthermore, an assessment of a proof-of-concept implementation is provided, validating the feasibility of the proposed approach.