A fuzzy model for IT security investments

Abstract

This paper presents a fuzzy set based decision support model for taking uncertainty into account when making security investment decisions for distributed systems. The proposed model is complementary to probabilistic approaches and useful in situations where probabilistic information is either unavailable or not appropriate to reliably predict future conditions. We first present the specification of a formal security language that allows to specify under which conditions a distributed system is protected against security violations. We show that each term of the security language can be transformed into an equivalent propositional logic term. Then we use propositional logic terms to define a fuzzy set based decision model. This optimization model incorporates uncertainty with regard to the impact of investments on the achieved security levels of components of the distributed system. The model also accounts for budget and security constraints, in order to be applicable in practice.