Knowledge Modeling for cross-organizational DevOps Teams
DevOps is a well known means to support the development of software by removing the borders between development and operations teams in software development projects. Software is continuously evolving and is not finished after it is shipped to the customer. Thus, DevOps strives for removing siloed organized teams which are part of the software development and software operations processes, such as development and operations teams.
However, breaking organizational structures the way that development and operations teams are not organized in silos, is very difficult and sometimes not even possible. For example, in a heterogeneous organization in which the development and operation of software is done by cross-organizational providers, the silos of development and operations teams are already defined by the way the organization is organized. Implementing a DevOps environment in such cases is a very difficult task. Each team and each organization follows its own interests, policies, concerns, etc. Thus, it is of tremendous importance that all competences being involved in the software development process communicate with each other in order to collaborate succesfully.
As a means of communication, models have proven to be an important and useful tool. They describe certain aspects using abstraction and build a basis to share knowledge between several individuals. Ontologies are a special form of models, being frameworks for representing shareable and reusable knowledge across a domain. However, the knowledge being expressed by the ontology can only be captured when its modelling format is suitable for describing it, i.e. each knowledge concept can be described in the chosen modelling technique as well as that every required individual can read and understand the model.
In the cooperation between the SWC chair and the “Deutsche Ärzte- und Apothekerbank” a model has been developed which should serve as an ontology to express and share knowledge between cross-organizational teams. Up to now the model lacks expressiveness of the knowledge to be be modelled. In this thesis we want to work out what the reasons for the lack of expressiveness are and find a suitable modelling approach, which is able to capture all expressed knowledge.