An Adaptive Decision-Driven Architecture for Information-Centric Internet-of-Things Applications

Abstract

This paper introduces a decision-execution supporting environment for the information-centric IoT applications. The main objective of this envisioned architecture is optimizing the IoT decisions' execution. The hallmark of the proposed architecture is the conjugate of the awareness about the congestion situations along with the quality of the decisions' supporting information. Controlling the incoming decisions' upsurge at the presented admission module is accomplished on a fashion that respects their degree of emergency. Wherefore, the Active Weighted Random Early Detection (AWRED) algorithm is integrated into autonomic admission judgment procedures to underpin the congestion control concept. Afterwards, an adaptive information-centric scheduling module is presented in order to devise an optimized information-aware execution schedule for the admitted real time decisions. The proposed scheduling algorithm integrates a priority-driven scheme with a Global best Harmony search (GBH) based schedule amending approach. This integrated scheduling concept empowers the proposed algorithm from tuning the schedules of the admitted real-time decisions such that the completeness opportunities of decisions are elevated without sacrificing the rest architecture objectives. Simulations results demonstrate the robustness of the proposed architecture performance against various congestion degrees, an essential feature in a highly dynamic environment as the IoT. Furthermore, it has been proved that the proposed architecture significantly surpasses state-of-art approaches in terms of both Quality of Information (QoI) and Quality of service (QoS) obligations represented by deadline missing ratio and responsiveness obligations, especially during congested situations.