Skip to Main Content
Security requirements analysis for business information systems in today's networked organization is difficult due to the complexity of the systems and the frequent change in the environment. Thus, it requires security knowledge to be explicitly represented, and well understood by system analysts and designer, which in turn being applied in feasible problem contexts. System requirements are often represented in modelling frameworks with different analytical focus, so security requirements knowledge shall reflect such difference and form an integrated treatment. This paper proposes to use modelling concepts from the i* and PF modeling language to capture recurring patterns of security problems. The main concepts used are actors, assets, and relations such as ownership and permissions. The major contribution of the approach is proposing the specific problem frames such as ownership, authorization, attack and protection, by decomposing a large problem into sub-problems (base frames), then evaluate the potential threats (attacking frames) applicable to each sub-problem by evaluate the compatibility of the two, security analysis is integrated into the system design process from the outset. The proposal can be generalized to the design of defensive measures as well as other NFR treatments.