Simulating IoT Systems from High-Level Abstraction Models for Quality of Service Assessment

In the context of IoT systems, the use of services is a key element in managing system complexity. Concepts such as service-oriented computing/architecture or quality of service (QoS) are present in many IoT systems and are the aim of several studies. However, the analysis and assessment of the behaviour of these concepts requires the deployment of the IoT system, implying high investments in hardware and software. Thus, in order to decrease these costs, the system can be simulated. In this regard, IoT simulations have been tackled focusing on low level aspects such as networks, motes, etc. rather than on high-level concepts, such as services or computing layers. In this proposal, a model-driven development approach named SimulateIoT is proposed to model, generate code and deploy IoT systems simulations from a high abstraction level (from models). Besides of modeling the IoT environment call generation, the IoT system could be simulated. From these simulations it is possible to assess QoS-related aspects such as the delay or jitter between two nodes, the variation of delay or jitter over time, the use of bandwidth, the packet loss, the variation of these parameters as the system changes (e.g. increase of sensors), check whether Service level agreements (SLA) are met, etc. In order to show the proposal, a case study, focused on an Internet of Vehicles (IoV) system is presented.