Semantic interoperability in the Internet of Things: An overview from the INTER-IoT perspective

Abstract

The Internet of Things (IoT) idea, explored across the globe, brings about an important issue: how to achieve interoperability among multiple existing (and constantly created) IoT platforms. In this context, in January 2016, the European Commission has funded seven projects that are to deal with various aspects of interoperability in the Internet of Things. Among them, the INTER-IoT project is aiming at the design and implementation of, and experimentation with, an open cross-layer framework and associated methodology to provide voluntary interoperability among heterogeneous IoT platforms. While the project considers interoperability across all layers of the software stack, we are particularly interested in answering the question: how ontologies and semantic data processing can be harnessed to facilitate interoperability across the IoT landscape. Henceforth, we have engaged in a “fact finding mission” to establish what is currently at our disposal when semantic interoperability is concerned. Since the INTER-IoT project is initially driven by two use cases originating from (i) (e/m)Health and (ii) transportation and logistics, these two application domains were used to provide context for our search. The paper summarizes our findings and provides foundation for developing methods and tools for supporting semantic interoperability in the INTER-IoT project (and beyond).

Introduction

The Internet of Things (IoT), conceptualized as an omnipresent network, consisting of physical or virtual objects/resources, equipped with sensing, computing, communication and actuating capabilities, can be seen as the most recent incarnation of, so-called, ubiquitous computing (Weiser, 1991, Friedewald and Raabe, 2011). With billions of sensors and actuators (things) already deployed, and combined into a number of domain-specific platforms, the vision of the hyper-connected world is closer than ever before.

Dealing with the vast amount of data produced by the things, their varying capabilities, and an exploding number of services, which they can offer (or require, to be “useful”), are among the biggest conceptual and technological challenges of our time. This challenge is further magnified by the typical ills of early-stage technology. With not much exaggeration it can be stated that “every IoT domain and every IoT vendor produces its own IoT platform.” As a matter of fact, different “vendor groups” can be found in different domains, while not a single vendor can be seen as having an “upper hand” in being positioned across all IoT domains. Even the, EU “sponsored,” FIWARE platform (https://www.fiware.org/) has only a limited uptake. Furthermore, as typically happens in early stages of technology adoption, no real (accepted by most players) standards can be found and none can be expected to materialize in the near future.

It is possible to deal with the interoperability challenge on multiple levels of the software stack. However, it is our belief that the key to solving the problem is on the “highest” level. Specifically, common description and data representation frameworks, which will characterize the things, their capabilities and data they produce, in machine-readable and machine-interpretable forms, are needed. Since the IoT can naturally be perceived as a “successor” of “the Web,” it should not come as a surprise that approaches, which are believed to have a chance to be successful in the case of the latter, should be considered for the former. Henceforth, it is reasonable (today—June 2016) to believe that semantic technologies, based on application of ontologies (Staab and Studer, 2009) have the best chance to facilitate interoperability among the things, as well as across the IoT platforms. Thus, ontologies should be used for semantic annotation, managing access, and resource discovery in the IoT. As a result, common interpretation of data and information, based on a shared ontology (or, more likely, multiple shared ontologies), is the best pathway to achieve semantic interoperability, which allows to exchange information such that the meaning of it will be automatically interpreted by the receiver in order to produce useful results.

We make these claims knowing very well that the original vision of the Semantic Web is still to be realized. For instance, as one can see from the, three-volume report, Internet of Things Success Stories (Cousin, 2014), published by the Internet of Things European Research Cluster and Smart Action, actual semantic methods are still used almost exclusively within the research community. However, recent developments in the “world of information processing” (e.g. success of the Linked Data Bizer et al., 2009) make us believe that widespread practical application of semantic technologies is just a matter of time. Note also that semantic technologies are often utilized within multi-agent systems (MAS). At the same time, MAS may provide the right set of mechanisms for implementing IoT interoperability (see, for instance, Fortino et al., 2012a, Fortino et al., 2013c, Fortino et al., 2014c, Fortino et al., 2012b, Underbrink et al., 2008). Therefore, the “push” to introduce semantic technologies into the IoT domain will come not only from within. It will also be facilitated by “outside technologies” that will be tried in the context of IoT interoperability.

This being the case, we have decided to take the bottom-up approach. To be able to apply semantic technologies, one has to have ontologies available. Therefore, we have delved into the state-of-the-art of ontologies in three areas: (1) general ontologies applicable to virtually any IoT platform, (2) ontologies in (e/m)Health and (3) ontologies in transportation and logistics. These ontologies, described in 3 Ontologies in the Internet of Things, 4 (e/m)Health ontologies, 5 Transportation/logistics ontologies, are considered in the context of use case scenarios introduced in Section 2. Finally, in Section 7, we outline a possible approach to use ontologies to achieve semantic interoperability among heterogeneous IoT platforms.