Augmented reality permits to embed virtual objects in the real environment to enhance the perception of users. In this paper, we describe an approach for embedding conceptual models using augmented reality in IT maintenance scenarios. It is based on the ArchiMate modeling language that has been extended with the goal of bridging the gap to models for physical environments. This allows, for example, to guide users in IT maintenance tasks by displaying necessary information originating from the models in the real world. The approach has been implemented on a novel metamodeling platform, which natively supports augmented reality scenarios.
been represented in the past primarily using text, graphs, charts, as well as 2D and 3D models.
This includes also the representation using augmented or virtual reality devices for displaying
conceptual enterprise architecture models in 3D space, e.g., to support decision making [
However, what is missing so far is a link between the high-level view of an enterprise architecture for supporting business operations via IT, i.e., the conceptual models, and the low-level view of physical IT hardware components. Integrating these two perspectives would enable maintenance tasks at the physical layer to be supported by knowledge from higher layers. This could be useful, for example, to trace problems in IT applications back to the physical environment, such as an incorrectly connected cable. Such a link could be beneficial not only for professional IT operations, but also for consumer use cases, for example, when having to connect devices in smart home scenarios.
In the following, we propose a first approach to bridge enterprise architecture models on the technology layer using ArchiMate notation with models of physical IT hardware components, thereby adding details to the enterprise architecture information to fuel it into real-time augmented reality supported work scenarios. This is achieved through an intermediary layer that acts as a facilitator for transitioning from the high-level view used in enterprise architectures to the level of physical components. 2. Model-based AR for it maintenance tasks As shown in Figure 1, our approach is based on three layers. At the top layer, traditional enterprise architecture models in ArchiMate notation are used. These may be connected to other ArchiMate models, e.g., for specifying IT services on the application layer, which in turn support particular business processes on the business layer. On the intermediate model layer, a ifrst transition towards physical models is described.
The information from the enterprise architecture needs to be mapped to a conceptual repreEnterprise Architecture Model Layer (Technology) Intermediate Model Layer (Wiring)
Physical Model Layer (Hardware) sentation that focuses on abstract physical devices and their connections. Thus, for example, the concept of a communication network on the EA level needs to be translated into network components and cables for connecting them. Finally, at the bottom layer, a model of the actual physical devices is shown. Here, the concrete devices with the detailed positioning of their ports are shown. Depending on the goals of the intended AR application, further details on coordinates or markers may be added on this layer to permit the positioning of virtual objects in the AR environment.
Figure 2 shows an excerpt of the integrated metamodel for our approach. It features three diferent types of diagrams corresponding to the three model layers, cf. Figure 1, classes, and relationclasses, as well as references between the layers. Due to limitations of space only a subset of concepts is shown here. As an example of references between layers, traces between the Device class on the enterprise architecture model layer, on the intermediate model layer and the concrete manifestations on the physical model layer are depicted. This highlights how high-level concepts from the EA perspective need to be mapped to concrete physical devices on the bottom layer.
In the example, the device needs to be specialized into either a PC, Switch, Modem, or Wall
concept on the physical layer. Thereby, the Wall concept has been introduced as an additional
concept only in the intermediate model layer, as it is necessary for the physical operation, but
has not been part of the models at the EA level.
3. Implementation and preliminary evaluation
The approach has been implemented using a new metamodeling platform based on [
C
R
R C
C C Class R Relationclass
C
C
C C C
R C
C
R R C
C R
C R C Enterprise Architecture
ModelLayer (Technology) Intermediate Model
Layer (Wiring)
PhysicalModel
Layer (Hardware)
reality environments using the W3C WebXR proposal [
Based on the implementation of the metamodels and a separate rudimentary prototypical application1, it was possible to evaluate the approach in four usage scenarios. These included scenarios such as fixing a broken Ethernet cable, the interruption of a network connection to a network attached storage (NAS) due to a malfunctioning Ethernet port, an unplugged coaxial cable that caused a missing internet connection, and a broken LAN port on a desktop PC. All scenarios were first modeled in ArchiMate notation to represent the general architecture components and their conceptual linkages. Subsequently, the intermediate and physical models were added. In addition, a BPMN diagram was elaborated to specify possible resolutions of malfunctions that could be applied to all scenarios. With the help of the prototypical AR application, the models could be displayed on a WebXR compatible tablet and embedded directly in the user’s view based on recognized markers, which acted as surrogates for a full object detection.
In this paper, we could briefly describe our vision towards model-based augmented reality
applications using information from enterprise architecture models. In particular, we highlighted
how an intermediate layer may help to transition to actual physical models that depict how the
information from high-level enterprise architectures maps to the real world. In the current state
of the modeling platform, the processing of multiple models is not yet supported. Future work
will include the integration of the processing and visualization of the created models directly in
1Available on: https://zenodo.org/record/7889218
the modeling platform, as well as the addition of functionalities for reasoning about the current
state of the environment for displaying the necessary model information [
Financial support for this paper is gratefully acknowledged by the Smart Living Lab funded by the University of Fribourg, EPFL, and HEIA-FR.