Peer-reviewed

The use of languages based on positive or negative expressiveness is very common for the deployment of security policies (i.e., deployment of permissions and prohibitions on firewalls through single-handed positive or negative condition attributes). Although these languages may allow us to specify any policy, the single use of positive or negative statements alone leads to complex configurations when excluding some specific cases of general rules that should always apply. In this paper we survey such a management and study existing solutions, such as ordering of rules and segmentation of condition attributes, in order to settle this lack of expressiveness.