Enabling the Collaborative Collection of Uncertainty Sources Regarding Confidentiality

Aus SDQ-Institutsseminar
Vortragende(r) Gabriel Gehrig
Vortragstyp Bachelorarbeit
Betreuer(in) Sebastian Hahner
Termin Fr 17. November 2023
Vortragsmodus in Präsenz
Kurzfassung With digitalization in progress, the amount of sensitive data stored in software systems is increasing. However, the confidentiality of this data can often not be guaranteed, as uncertainties with an impact on confidentiality exist, especially in the early stages of software development. As the consideration of uncertainties regarding confidentiality is still novel, there is a lack of awareness of the topic among software architects. Additionally, the existing knowledge is scattered among researchers and institutions, making it challenging to comprehend and utilize for software architects. Current research on uncertainties regarding confidentiality has focused on analyzing software systems to assess the possibilities of confidentiality violations, as well as the development of methods to classify uncertainties. However, these approaches are limited to the researchers’ observed uncertainties, limiting the generalizability of classification systems, the validity of analysis results, and the development of mitigation strategies. This thesis presents an approach to enable the collection and management of knowledge on uncertainties regarding confidentiality, enabling software architects to comprehend better and identify uncertainties regarding confidentiality. Furthermore, the proposed approach strives to enable collaboration between researchers and practitioners to manage the effort to collect the knowledge and maintain it. To validate this approach, a prototype was developed and evaluated with a user study of 17 participants from software engineering, including 7 students, 5 researchers, and 5 practitioners. Results show that the approach can support software architects in identifying and describing uncertainties regarding confidentiality, even with limited prior knowledge, as they could identify and describe uncertainties correctly in a close-to-real-world scenario in 94.4% of the cases.