Service-Oriented Architecture in industrial vehicles: Applicability of the Arrowhead framework for distributed and dependable computing

The expected 4th industrial revolution comprising concepts such as Internet of Things, cloud computing, edge computing and machine learning is gradually happening. In its core is the real-time data which is being produced and utilized by cyber-physical systems. Cars and industrial vehicles are examples of such cyber-physical systems producing constantly increasing amount of data. At the same time we are witnessing the strengthening trend of edge computing in which the data is processed by the system which collected it or within the direct proximity. Edge computing can be seen complementing the already established cloud computing paradigm. New solutions must be adopted in both hardware and software level to cope with the increased amount of edge computing. The research objective of the conducted study was to gain knowledge of the service-oriented architecture's applicability as industrial vehicles' software architecture. Particularly concerned are the requirements set by compute-intensive edge computing applications in vehicles, such as machine vision. In addition, it was studied how applicable would service-oriented architecture adhering Arrowhead framework be in the described use context. The research was carried out as an information system Design Science Reseach. It consistent of a comprehensive literature review, design and construction of an IT-artifact, and the evaluation of the artifact. As the IT-artifact, an information system of possible in-vehicle subsystem capable of performing machine vision inference and utilizing Ethernet-based communication was chosen. The subsystem was designed following the Arrowhead framework guidelines and constraints. The evaluation of the artifact focused on illustrative scenario evaluating its fault tolerance. Moreover, the performance of the artifact was measured. The results indicated that the dynamic service orchestration of Arrowhead does not suit for fast response time requiring applications. This covers majority of the vehicle applications. In addition, Arrowhead framework was found containing a few flexibility and adaptability related issues. Service-oriented architecture was seen as a promising development trend for industrial vehicles, and other similar cyber-physical systems. Among other benefits, service-orientation enables distribution of the application components, autonomous interaction between the systems and increases the reusability of the software