Hierarchical Control for Self-adaptive IoT Systems A Constraint Programming-Based Adaptation Approach

Abstract : The self-adaptation control of Internet of Things (IoT) systems ought to tackle uncertainties in the dynamic environment (application level), as well as the dynamic computation infrastructure (architecture level). While the control of those two levels is generally separated, they should coordinate to guarantee functionality and quality. This paper proposes a conceptual model for the separation of concerns in controlling the environment and infrastructure events. The approach is applied on a real case: Melle-Longchamp area's smart power transmission network (in France). A hierarchical architecture with a control mechanism formalized with constraint programming (CP) is modeled. The control system assesses the reconfigurations that enhance the quality of service (QoS) while considering the internal and external limitations. The CP considers the desired application level control modes and assesses their feasibility by computing the response time and availability using a Netflow algorithm. The outcomes of this research supported design decisions and provided architectural reconfiguration solutions to the French Power Transmission Company (RTE).