Semantic Techniques to Support IoT Interoperability
243
6 The WSDL and OWL-S Representations
Once the analysis of the API has terminated and the corresponding ontology has been
created, building the WSDL representation becomes a trivial matter. Indeed, WSDL
has a very precise structure, which can be easily replicated with the information
contained in the API ontology. For our purposes, WSDL is useful in different ways.
First of all, it provides a means to automatically build calls to the sensors’ interfaces,
using a standard representation which can be easily interpreted by external tools.
Second, it is the standard language used to describe the grounding of OWL-S which
is fundamental in our approach for the description of sensors’ compositions. Listing
1 reports an example of WSDL representation, referring to the GoogleFit service.
OWL-S representations at this stage are not very meaningful, as sensors do not have
complex workflows to describe, unless they are part of a sensor network. In this case,
OWL-S can be exploited to describe the net process and data workflows, by using the
semantic annotations provided by the ontology itself. At this stage, however, we have
no network configurations to represent, and OWL-S becomes a mere interconnection
between WSDL parameters and ontological concepts, as shown in Fig. 2 where only
the main ontologies and schemas are imported and connected.
7 Conclusions and Future Work
In this paper, a prototypal tool for the automatic analysis of sensors’ REST API has
been presented. The tool aims not only at automatically analyze the APIs, but also
to derive a semantic representation of them, which can be then used to support the
manual annotation with external domain ontologies. The main steps followed by
the tool have been described: the critical analysis phase, in which a combination of
automatic and manual approaches lead to the production of a semantic API Grap;
the manual annotation phase, where the semantic representation is enriched with
connections to domain ontologies; the WSDL and OWL-S production phase. At the
moment, our research is focused on two aspects: to improve the analysis phase, in
order to reduce the toll on the human operator; to actually build and represent sensor
network using OWL-S and exploit it for orchestration.
Acknowledgements This work has received funding from the University of Campania Luigi Vanvitelli V:ALERE research programme under the SSCeGov (Semantic, Secure and Law Compliant
e-Government Processes) project.
243
6 The WSDL and OWL-S Representations
Once the analysis of the API has terminated and the corresponding ontology has been
created, building the WSDL representation becomes a trivial matter. Indeed, WSDL
has a very precise structure, which can be easily replicated with the information
contained in the API ontology. For our purposes, WSDL is useful in different ways.
First of all, it provides a means to automatically build calls to the sensors’ interfaces,
using a standard representation which can be easily interpreted by external tools.
Second, it is the standard language used to describe the grounding of OWL-S which
is fundamental in our approach for the description of sensors’ compositions. Listing
1 reports an example of WSDL representation, referring to the GoogleFit service.
OWL-S representations at this stage are not very meaningful, as sensors do not have
complex workflows to describe, unless they are part of a sensor network. In this case,
OWL-S can be exploited to describe the net process and data workflows, by using the
semantic annotations provided by the ontology itself. At this stage, however, we have
no network configurations to represent, and OWL-S becomes a mere interconnection
between WSDL parameters and ontological concepts, as shown in Fig. 2 where only
the main ontologies and schemas are imported and connected.
7 Conclusions and Future Work
In this paper, a prototypal tool for the automatic analysis of sensors’ REST API has
been presented. The tool aims not only at automatically analyze the APIs, but also
to derive a semantic representation of them, which can be then used to support the
manual annotation with external domain ontologies. The main steps followed by
the tool have been described: the critical analysis phase, in which a combination of
automatic and manual approaches lead to the production of a semantic API Grap;
the manual annotation phase, where the semantic representation is enriched with
connections to domain ontologies; the WSDL and OWL-S production phase. At the
moment, our research is focused on two aspects: to improve the analysis phase, in
order to reduce the toll on the human operator; to actually build and represent sensor
network using OWL-S and exploit it for orchestration.
Acknowledgements This work has received funding from the University of Campania Luigi Vanvitelli V:ALERE research programme under the SSCeGov (Semantic, Secure and Law Compliant
e-Government Processes) project.