Architectural Optimization for Confidentiality Under Structural Uncertainty

Abstract:

More and more connected systems gather and exchange data. This allows building smarter, more efficient and overall better systems. However, the exchange of data also leads to questions regarding the confidentiality of these systems. Design notions such as Security by Design or Privacy by Design help to build secure and confidential systems by considering confidentiality already at the design-time. During the design-time, different analyses can support the architect. However, essential properties that impact confidentiality, such as the deployment, might be unknown during the design-time, leading to structural uncertainty about the architecture and its confidentiality. Structural uncertainty in the software architecture represents unknown properties about the structure of the software architecture. This can be, for instance, the deployment or the actual implementation of a component. For handling this uncertainty, we combine a design space exploration and optimization approach with a dataflow-based confidentiality analysis. This helps to estimate the confidentiality of an architecture under structural uncertainty. We evaluated our approach on four application examples. The results indicate a high accuracy regarding the found confidentiality violations.

BibTeX:

@inproceedings{walter_architectural_2022,
    title = {{Architectural Optimization for Confidentiality Under Structural Uncertainty}},
    author = {Walter, Maximilian and Hahner, Sebastian and Seifermann, Stephan and Bures, Tomas and Hnetynka, Petr and Pacovský, Jan and Heinrich, Robert},
    year = {2022},
    booktitle = {{Software Architecture (ECSA 2021 postproceedings)}},
    editor = {Scandurra, Patrizia and Galster, Matthias and Mirandola, Raffaela and Weyns, Danny},
    publisher = {Springer},
    series = {{LNCS}},
    doi = {10.1007/978-3-031-15116-3_14},
    isbn = {978-3-031-15116-3},
    pages = {309--332},
    volume = {13365},
}

Proceedings paper