ThreMA: Ontology-Based Automated Threat Modeling for ICT Infrastructures

Abstract

<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Threat Modelling</i> allows defenders to identify threats to which the target system is exposed. Such a process requires a detailed infrastructure analysis to map threats to assets and to identify possible flaws. Unfortunately, the process is still mostly done manually and without the support of formally sound approaches. Moreover, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Threat Modelling</i> often involves teams with different levels of security knowledge, leading to different possible interpretation in the system under analysis representation. Threat modelling automation comes with two main challenges: (i) the need for a standard representation of models and data used in various stages of the process, establishing a formal vocabulary for all involved parties, and (ii) the requirement for a well-defined inference rule set enabling reasoning process automation for threat identification. The paper presents the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">ThreMA</i> approach to automating threat modelling for ICT infrastructures, aiming at addressing the key automation issues through the use of ontologies. Specifically, a formal vocabulary for modelling an ICT infrastructure, a threat catalog and a set of inference rules needed to support the reasoning process for threat identification are provided. The proposed approach has been validated against actual significant case studies provided by different Stakeholders of the Italian Public Sector.