Systematic development of UMLsec design models based on security requirements

  • Authors:

  • Denis Hatebur;Maritta Heisel;Jan Jürjens;Holger Schmidt

  • Affiliations:

  • Software Engineering, Department of Computer Science and Applied Cognitive Science, Faculty of Engineering, University Duisburg-Essen and Institut für technische Systeme GmbH, Germany;Software Engineering, Department of Computer Science and Applied Cognitive Science, Faculty of Engineering, University Duisburg-Essen, Germany;Software Engineering, Department of Computer Science, TU Dortmund and Fraunhofer Institut für Software- und Systemtechnik, Germany;Software Engineering, Department of Computer Science, TU Dortmund, Germany

  • Venue:
  • FASE'11/ETAPS'11 Proceedings of the 14th international conference on Fundamental approaches to software engineering: part of the joint European conferences on theory and practice of software
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

Developing security-critical systems in a way that makes sure that the developed systems actually enforce the desired security requirements is difficult, as can be seen by many security vulnerabilities arising in practice on a regular basis. Part of the difficulty is the transition from the security requirements analysis to the design, which is highly nontrivial and error-prone, leaving the risk of introducing vulnerabilities. Unfortunately, existing approaches bridging this gap largely only provide informal guidelines for the transition from security requirements to secure design. We present a method to systematically develop structural and behavioral UMLsec design models based on security requirements. Each step of our method is supported by model generation rules expressed as pre- and postconditions using the formal specification language OCL. Moreover, we present a concept for a CASE tool based on the model generation rules. Thus, applying our method to generate UMLsec design models supported by this tool and based on previously captured and analyzed security requirements becomes systematic, less error-prone, and a more routine engineering activity. We illustrate our method by the example of a patient monitoring system.