Page - 171 - in Intelligent Environments 2019 - Workshop Proceedings of the 15th International Conference on Intelligent Environments

However, the concept of interoperability is interpreted in variousways due to the complexity and versatility of the topic or sometimes due to aweak understanding [6]. The lack of interoperability in the context of IoT applications is often complained, but there are very fewstudies that addresses this issue systematically and to the full extent. Existingcomprehensivestudies(e.g.[3],[6]and[7])typicallyaddressinteroperabilityby attempting to unify asmanydifferent aspects of interoperability in one commonmodel as possible. Furthermore, thesemodels try tomake the degree of interoperabilitymea- surable and comparable (i.e. a value indicating towhat extent interoperability is given between two ormore objects). This degree is typically measured by the fulfilment or non-fulfilmentofcriteriadefinedbytherespectivemodel:Themorecriteria fulfilled, the higher thedegreeof interoperability.Somehigher (i.e.moreadvanced)criteria canonly be fulfilled if lower criteria are fulfilled aswell, thus leading to a hierarchyof different criteria and, consequently, a hierarchical scale ofmeasurement.With our approach,we want tomake interoperabilitymeasurablenot onlyonahierarchical scale, but alsowith metricvalues.With thisfirst approach, the focushere isnoton interoperability in its en- tirely (i.e.with all its characteristics, aspects andcriteria), but onlyona specific aspect, namely the call of operations bymessage passingwhich is ubiquitous in IoT environ- ments.More specifically,wemodel this specific aspect ofmessage passing as a causal model.Nevertheless,ouraspectandthus themetricmeasurementmethodcanbeembed- ded in a hierarchically orientedmodel, so that only this aspect ismeasuredmetrically, but the remainingaspects are still evaluatedusingahierarchical scale. This paper is organized as follows: Section 2 gives a brief overviewof definitions of interoperability to provide a first insight into this complex topic. In section 3, we analyze several approaches for classifyingandmeasuring interoperability aspects based on a hierarchical scale and select one framework in order to embed ourmeasurement method. The causalmodel is derived in section 4 and analyzed in section 5. Finally, a conclusion isgiven in section6. 2. Definitionof Interoperability Theproblemof lacking interoperability is neither limited to the areaof IoTnor is dedi- catedto theapplicationfieldof informationsystemsexclusively. IdeandPustejovsky, for instance, defines interoperability as“ameasure of degree towhich diverse systems, or- ganizations,and/or individualsareable towork together toachieveacommongoal”[8]. According to this definition, interoperability is presentwhenever there is the necessity that different participantsmeet and interactwith eachother to fulfill a commongoal, no matter whether these participants are physical machines, IT systems or human actors. Moreover, this definition describes interoperability as ameasurable property indicating howgoodorbad theparticipants can interactwitheachother, or inotherwords, towhat degree they are able to inter-operate. In the context of information systems, IEEE de- fines interoperability as“theability of twoormore systemsor components to exchange information and to use the information that has been exchanged”[9]. A similar defini- tion is given byTsilaswith“the ability of heterogeneous IT networks, applications, or components to exchange and use information, that is to ‘talk to and understand’ each other”[10].Bothdefinitionsareclose to the initialgenericdefinition,but incomparison, the participants are specified as hardware and software components and the interaction S. Kotstein and C. Decker /AnApproach for Measuring IoT Interoperability 171