Article Preview
Top1. Introduction
The Internet of Things (IoT) is a complex ecosystem of various IoT devices, solutions, services and applications. Interoperability is one of the main problems of the IoT, because it is impossible to manage and connect individually a vast amount of different IoT devices. Interconnecting different things enables applications to understand their environment better and to make intelligent decisions based on this information. Many existing papers dealing with IoT interoperability use Semantic Web languages, tools and technology to achieve the interoperability at different levels of IoT data, devices, services and applications. Semantic Web technologies in IoT domain are also used for other purposes, such as description and search of things and IoT services, composition of services, models and reasoning over IoT resources. Barnaghi et al. (Barnaghi, Wang, Henson, & Taylor, 2012) listed the main scenarios that demonstrate the importance of IoT semantics: semantics for interoperability, IoT data integration, IoT data abstraction and access, resource/service search and discovery, semantic reasoning and interpretation. Some researchers even coined the new term, Semantic Web of Things, to define a new field that combines the IoT and Semantic Web technologies. Also, various ontologies exist that try to describe the IoT ecosystem.
There is a significant number of research papers on the usage of Semantic Web technologies for IoT interoperability, but, to date, there has not been a systematic literature review of research on the application of Semantic Web in the IoT integration, making it difficult to assess the maturity, trends and research gaps of the intersection of these two popular research areas. To demonstrate that a similar review has not been already done, we searched the Google Scholar, IEEE Xplore, ScienceDirect, ACM Digital Library, Wiley Online Library, SpringerLink, and Web of Science Core Collection January 2017 with the following search string Semantic Web AND (Internet of thing OR IoT) AND (Systematic Literature Review OR Systematic Review OR Secondary Study OR Literature Survey OR SLR).
In 2012, Barnaghi et al. (Barnaghi et al., 2012) reviewed the recent developments on applying the semantic technology to IoT and discussed the challenges in information modelling, ontology design and processing semantic data. Applying semantic technologies to IoT promotes interoperability and facilitates effective data access and integration, resource discovery, semantic reasoning and knowledge extraction. Semantic IoT descriptions need to be shared, processed and interpreted. This review includes papers prior to 2012, but our paper will show that most of the semantic IoT papers date from 2014 and 2015, so the results of their study are outdated now. Additionally, they did not use the systematic review method that is more comprehensive and their focus was not on IoT interoperability.
In spite of the synergy between Semantic Web technologies and IoT, the literature still lacks a comprehensive overview and discussion about open issues. In this work, our main aim was to bridge this gap by systematically listing, comparing and classifying the existing research on Semantic Web usage in IoT interoperability research, to identify gaps in the existing works and to find promising research directions. The results of this systematic literature review are beneficial for researchers and practitioners that work either in IoT or Semantic Web area. Considering the importance of semantic technologies usage in IoT and the relative maturity of this field, a consolidation of the existing evidence on Semantic Web for IoT interoperability is timely.
It is important to note while conducting this review new related papers have emerged (like (Goel, Pahal, Jain, & Chaudhury, 2017; B Jia, Li, & Zhou, 2017)) and some new journals. For example, one journal that has gained great impact on the filed in 2016 and 2017 is the new journal “IEEE Internet of Things Journal”. Because of that, the review needs to have a time limit on papers which are included, otherwise new papers would constantly be added and the review would never be finished. For this review time limit is May 6th 2016, when the selected databases where queried. Because of the time limit there are not many articles form 2016, since they are not yet referenced by the databases event though they are published. This fact does not prevents us from answering our research questions, but it reminds why is it important to report the review protocol so everyone can repeat the search if interested in the newest findings in the field.