Distributed Hybrid Simulation of the Internet of Things and Smart Territories (1710.04252v2)

Abstract: This paper deals with the use of hybrid simulation to build and compose heterogeneous simulation scenarios that can be proficiently exploited to model and represent the Internet of Things (IoT). Hybrid simulation is a methodology that combines multiple modalities of modeling/simulation. Complex scenarios are decomposed into simpler ones, each one being simulated through a specific simulation strategy. All these simulation building blocks are then synchronized and coordinated. This simulation methodology is an ideal one to represent IoT setups, which are usually very demanding, due to the heterogeneity of possible scenarios arising from the massive deployment of an enormous amount of sensors and devices. We present a use case concerned with the distributed simulation of smart territories, a novel view of decentralized geographical spaces that, thanks to the use of IoT, builds ICT services to manage resources in a way that is sustainable and not harmful to the environment. Three different simulation models are combined together, namely, an adaptive agent-based parallel and distributed simulator, an OMNeT++ based discrete event simulator and a script-language simulator based on MATLAB. Results from a performance analysis confirm the viability of using hybrid simulation to model complex IoT scenarios.