Recent advances in sciences and business models required the invention of new and innovative types of systems in order for them to be used as a development and deployment platform for applications. Examples of such systems are the Grid and Cloud computing paradigms. Both of them are evolutionary distributed and collaborative systems, which have currently become the de facto platforms for the development and deployment of various types of applications. Despite the different nature of these two types of systems, several requirements and principles remain the same in both of them. Security is an essential principle and it is required to be maintained during any collaboration among participants. Despite the benefits of existing security solutions there are few proposals that addressed the problem of how to maintain security among domains where each implement its own access control (AC) policy. Moreover, the majority of existing solutions are static in nature and not suitable for the examined systems. models and mechanisms using the conceptual categorization, their pros and cons are exposed. Apart from the mapping of the AC area in Grid and Cloud systems, the given comparison renders valuable information for further enhancement of current approaches. Moreover, we define an enhanced Role-Based Access Control (RBAC) model entitled domRBAC for collaborative systems, which is based on the ANSI INCITS 359-2004 AC model. The domRBAC is capable of differentiating the security policies that need to be enforced in each domain and to support collaboration under secure inter-operation. Cardinality constraints along with context information are incorporated to provide the ability of applying simple usage management of resources for the first time in a RBAC model. Furthermore, secure inter-operation is assured among collaborating domains during interdomain role assignments, gradually and automatically. Yet, domR-BAC, as an RBAC approach, intrinsically inherits all of its virtues such as ease of management, and Separation of Duty (SoD) with the latter also being supported in multiple domains. As a proof of concept, we implemented a simulator based on the definitions of our proposed AC model and conducted with experimental studies to demonstrate the feasibility and performance of our approach. Lastly, we provide a formal definition of secure inter-operation properties in temporal logic, which can be verified using model checking techniques. The proposed technique consists of a generic one, and thus, can be used in any RBAC model to verify indirectly the correctness of the secure inter-operation functions that implement the global security policy. As a proof of concept, we provide examples that illustrate the enforcement of the defined secure inter-operation properties, which have to be verified in RBAC policies, and a performance analysis of the proposed technique.