Big data applications usually entail the development of complex IT-systems that collect, combine, analyze, communicate and use large amounts of data from different sources. The interactions between system components can be difficult to understand and describe, due to their large number, heterogeneity, and concurrency (e.g. continuous and discrete processes, or different technologies, may interact). In addition, system changes may occur as a result of new requirements or constraints during the lifecycle of any data processing system.

In the system design process, multiple stakeholders need to collaborate and align their individual views of the system, e.g. sensor specialists, hardware designers, IT integrators, operational departments, etc. This alone can be a challenge because different domain experts often have different concerns and use different terminologies and abstraction concepts. Only few of these experts have a system design or modeling background, which can make the elicitation and capture of system knowledge difficult and error-prone.

Our research is concerned with the subject-oriented approach to process modeling (S-BPM) (Fleischmann et al. 2012). We argue that it can address the various design challenges arising from the high complexity of big data systems. It has four foundational characteristics that distinguish it from other approaches (Kannengiesser 2017), Figure 1: 1. notational simplicity, 2. widely-shared, intuitive semantics, 3. encapsulation and 4. seamless integration.