Towards interoperable, cognitive and autonomic IoT systems: An agent-based approach

Within the Information and Communication Technology hype cycle the Internet of Things (IoT) represents a prominent subject, being rich in potentiality as well as in development issues. In this paper, we face the IoT developing process first “in the small” by designing the Smart Objects (SOs, representing the fundamental IoT building blocks) as agents, and after “in the large” by treating the IoT systems as Multi Agent Systems (MASs). Indeed, the agent abstraction is a suitable paradigm to instill smartness and autonomy within a single SO and consequently to realize distributed, self-steering and heterogeneous IoT systems. In such directions, the Agent-based COoperating SO (ACOSO) middleware represents a viable solution for the programming, development and management of agent-based SO systems while its performance verification on different IoT networks of different scale has been made through the Omnet++ simulator.