10th IFIP TC-6 TC-11 International Conference on Communications and Multimedia on Security (CMS’06), LNCS 4237, Springer, pp. 54-66, October, 2006. DOI
This paper presents a model for delegation based on partial orders, proposing the subclass relation in OWL as a way to represent the partial orders. Delegation and authorization decisions are made based on the context. In order to interact with the context, we define the Type of a credential as a way to introduce extra information regarding context constraints. When reasoning about delegation and authorization relationships, our model benefits from partial orders, defining them over entities, attributes and the credential type. Using these partial orders, the number of credentials required is reduced. It also classifies the possible criteria for making authorization decisions based on the context, in relation to the necessary information.
3rd international conference on Mobile multimedia communications (MobiMedia ’07), ICST, pp. 43:1–43:6, 2007.
When delegation is implemented using the attribute certificates in a Privilege Management Infrastructure (PMI), this one reaches a considerable level of distributed functionality. However, the approach is not flexible enough for the requirements of ubiquitous environments. Additionally, the PMI can become a too complex solution for devices such as smartphones and PDAs, where resources are limited. In this work, we solve the previous limitations by defining a second class of attributes, called domain attributes, which are managed directly by users and are not right under the scope of the PMI, thus providing a light solution for constrained devices. The two classes of attributes are related by defining a simple ontology. We also introduce in the paper the concept of Attribute Federation which is responsible for supporting domain attributes and the corresponding ontology.
6th International Conference on Trust, Privacy and Security in Digital Business (TrustBus’09), Springer-Verlag, pp. 86-94, September, 2009. DOI
SCADA systems represent a challenging scenario where the management of critical alarms is crucial. Their response to these alarms should be efficient and fast in order to mitigate or contain undesired effects. This work presents a mechanism, the Adaptive Assignment Manager (AAM) that will aid to react to incidences in a more efficient way by dynamically assigning alarms to the most suitable human operator. The mechanism uses various inputs for identifying the operators such as their availability, workload and reputation. In fact, we also define a reputation component that stores the reputation of the human operators and uses feedback from past experiences.
Security and Communication Networks, vol. 9, issue 12, Wiley, pp. 1769-1785, 08/2016. DOI (I.F.: 1.067)
Several generic methods exist for achieving chosen-ciphertext attack (CCA)-secure public-key encryption schemes from weakly secure cryptosystems, such as the Fujisaki–Okamoto and REACT transformations. In the context of proxy re-encryption (PRE), it would be desirable to count on analogous constructions that allow PRE schemes to achieve better security notions. In this paper, we study the adaptation of these transformations to proxy re-encryption and find both negative and positive results. On the one hand, we show why it is not possible to directly integrate these transformations with weakly secure PRE schemes because of general obstacles coming from both the constructions themselves and the security models, and we identify 12 PRE schemes that exhibit these problems. On the other hand, we propose an extension of the Fujisaki–Okamoto transformation for PRE, which achieves a weak form of CCA security in the random oracle model, and we describe the sufficient conditions for applying it