Computational Verification of Security Requirements

  • Gideon Dadik Bibu

Student thesis: Doctoral ThesisPhD

Abstract

One of the reasons for persistence of information security challenges in organisations is that security is usually seen as a technical problem. Hence the emphasis on technical solutions in practice. However, security challenges can also arise from people and processes. We therefore approach the problem of security in organisations from a socio-technical perspective and reason that the design of security requirements for organisations has to include procedures that would allow for the design time analysis of the system behaviour with respect to security requirements.In this thesis we present a computational approach to the verification and validation of elicited security requirements. This complements the existing approaches of security requirements elicitation by providing a computational means for reasoning about security requirements at design time. Our methodology is centered on a deontic logic inspired institutional framework which provides a mechanism to monitor the permissions, empowerment, and obligations of actors and generates violations when a security breach occurs.We demonstrate the functionality of our approach by modelling a practical scenario from health care domain to explore how the institutional framework can be used to develop a model of a system of interacting actors using the action language InstAL. Through the application of the semantics of answer set programming (ASP), we demonstrate a way of carrying out verification of security requirements such that it is possible to predict the effect of certain actions and the causes of certain system states. To show that our approach works for a number of security requirements, we also use other scenarios to demonstrate the analysis of confidentiality and integrity requirements.From human factor point of view compliance determines the effectiveness of security requirements. We demonstrate that our approach can be used for management of security requirements compliance. By verifying compliance and predicting non-compliance and its consequences at design time, requirements can be redesigned in such a way that better compliance can be achieved.
Date of Award19 Nov 2014
Original languageEnglish
Awarding Institution
  • University of Bath
SponsorsPetroleum Trust Development Fund
SupervisorJulian Padget (Supervisor), James Davenport (Supervisor) & Marina De Vos (Supervisor)

Keywords

  • Verification
  • Security
  • Requirements

Cite this

'