A key issue in RBAC systems is how to efficiently handle the user authorization process. That is, whether or not to grant a user's request to acquire a set of requested permissions or to activate a set of requested roles in a single session. The presence of hybrid hierarchies as well as the cardinality and dynamic separation of duty constraints make the issue more complex. In this paper, we define this issue as the user authorization query problem consisting of a role mapping problem and an activation checking problem. We also propose a set of algorithms to solve the role mapping and the activation checking problems. We show that our model is practical and flexible, and can deal with various cases in presence of the hybrid hierarchy and cardinality/DSoD constraints.

A role hierarchy defines permission acquisition and role-activation semantics through role--role relationships. It can be utilized for efficiently and effectively structuring functional roles of an organization having related access-control needs. The focus of this paper is the analysis of hybrid role hierarchies in the context of the generalized temporal role-based access control (GTRBAC) model that allows specification of a comprehensive set of temporal constraints on role, user-role, and role-permission assignments. We introduce the notion of uniquely activable set (UAS) associated with a role hierarchy that indicates the access capabilities of a user resulting from his membership to a role in the hierarchy. Identifying such a role set is essential, while making an authorization decision about whether or not a user should be allowed to activate a particular combination of roles in a single session. We formally show how UAS can be determined for a hybrid hierarchy. Furthermore, within a hybrid hierarchy, various hierarchical relations may be derived between an arbitrary pair of roles. We present a set of inference rules that can be used to generate all the possible derived relations that can be inferred from a specified set of hierarchical relations and show that it is sound and complete . We also present an analysis of hierarchy transformations with respect to role addition, deletion, and partitioning, and show how various cases of these transformations allow the original permission acquisition and role-activation semantics to be managed. The formal results presented here provide a basis for developing efficient security administration and management tools.

Recently, administration of RBAC systems using a role-based approach has become very appealing because of the benefits that such an approach typically brings. This approach uses RBAC itself to manage RBAC policies so that the administration functions can be decentralized and made more efficient. Existing RBAC administration models, however, fail to deal with RBAC systems with hybrid hierarchy, which has been shown to be necessary to specify fine-grained RBAC policies. In this paper, we propose a Scoped Administration model for RBAC with Hybrid Hierarchy (SARBAC07) by using the notion of an administrative scope that was earlier proposed in the SARBAC model. We show that our model keeps all the advantages of the original model and can deal with more complex situations where hybrid hierarchy is needed.

Multi-domain environments where distributed multiple organizations interoperate with each other are becoming a reality as witnessed by emerging Internet-based enterprise applications. Ensuring secure interoperation in such multi-domain environments has drawn considerable research works in the past, especially in tightly coupled, federated environments. However, methods applied to such environments are not suitable in emerging loosely-coupled environments where the inter-domain interactions are transient and based on specific requirements within a given context (e.g., time, location), which is typical in web service, P2P and Grid-based applications. In this paper, we propose a request-driven secure interoperation framework to facilitate secure interoperation in loosely-coupled environments where the individual domains employ role-based access control policies. In particular, our proposed framework is driven by the service requirements and dynamically integrates relevant policy components between interacting domains.

Phishing is a significant problem involving fraudulent email and web sites that trick unsuspecting users into revealing private information. In this paper, we present the design, implementation, and evaluation of CANTINA, a novel, content-based approach to detecting phishing web sites, based on the TF-IDF information retrieval algorithm. We also discuss the design and evaluation of several heuristics we developed to reduce false positives. Our experiments show that CANTINA is good at detecting phishing sites, correctly labeling approximately 95% of phishing sites.