Correctness of IoT-based systems: From a DSL to a mechanised analysis
Abstract
Systems based on the Internet of Things are being widely used in industrial areas such as smart manufacturing, in smart health monitoring, in entertainment and in home automation application. They impact more and more every day life. Ensuring their correct construction, their well functioning and their reliability is an important issue for some of these systems which can be critical in case of dysfunction. The main requirements on physical architectures and control software are common to most of IoT-based systems. We propose, on the basis of the common architectural properties and the behaviour of IoT-based systems, the rigorous analysis of their intrinsic consistency properties; specific properties may also be considered and analysed. We propose a model-based approach to help in systematically modelling and analysing these systems. Our approach focuses both on the formalisation of the model of the targeted systems for their rigorous analysis purpose, and on the design of a modelling language as a domain specific language dedicated to describe IoT-based systems. We build a generic formal model which captures the common features and the properties required for any IoT-based system. This formal model is then a parametrised model where the parameters are the specific data and properties of a given system. The proposed DSL, named τila , is used to describe an IoT system which will be submitted to formal analysis of correctness. τila is a tiny declarative language, extensible, based on a library of building component blocks available as types. In fact, a description in τila helps to model a system, by collecting the parameters necessary to instantiate the generic formal model. Therefore the analysis requirements are based on the generic model. We experimented our approach using the proposed τila DSL for modelling and the Event-B framework for the formal analysis. The overall proposed approach is accompanied by a model editor generated to assist users in modelling an application and a prototype tool we have developed to assist the users in analysing their systems. Examples are provided. The generic formal model is extensible; it may be profitably adapted to more general hybrid or cyber–physical systems. Moreover, our generic model is independent of the target formal modelling tools; it may be implemented in various other formal analysis environments.