Towards dependable CPS/IoT ecosystem
This thesis defines a concept of CPS/IoT Ecosystem as a hierarchical structure, that governs practices and procedures for modeling, design, development, execution and operation of smart systems. We divide these systems in three loosely dependent scopes of operation: the cloud, the fog, and the swarm. Furthermore, we propose a series of methods and approaches that support the dependable design, execution, and operation of CPS/IoT Ecosystems: the methods for ensuring the deterministic execution of tasks in safety constrained applications, a communication channels virtualization for many-core architectures, and a secure communication architecture for many-core platforms. A CPS/IoT Ecosystem is a highly heterogeneous environment with hardware and software components that are designed and implemented by multiple organizations. To ensure coherence between different components and to reduce complexity we propose a continuous integration and deployment (CI/CD) scheme for CPS/IoT Ecosystem. Furthermore, we demonstrate a runtime verification (RV) mechanism that provides a basis for quality of service (QoS) orchestration and dynamic reconfiguration of CPS/IoT applications. As final step in this thesis we propose methods to achieve energy-sustainable CPS/IoT Ecosystems. In conclusion, this thesis tries to seed methodological guidelines on how to build dependable CPS/IoT Ecosystems for applications with various confidence requirements. We want to understand the upcoming changes and reduce eventual effects of ad-hoc development. To explain physical environments using mathematical models and to learn new emerging behaviors using this massive incursion of new data and new insights.
DOI: https://doi.org/10.34726/hss.2022.103104
Authors:Haris Isakovic
Title of the source: Doctoral dissertation
Publisher: Technische Universität Wien
Relevant pages: 1-155
Year: 2022