RINA (Recursive InterNetwork Architecture) is a promising network architecture based on the concept of distributed Inter Process Communications (IPC). In RINA there is a single type of programmable layer (the DIF "Distributed IPC Facility"), that repeats as many times as needed by the network designer. Each layer is composed of only two programmable protocols: one performs data transfer and data transfer control functions, while the other carries out layer management functions. Multiple policies can be plugged to both protocols at each layer to adapt them to operational requirements. In this paper we explore the advantages of the RINA architecture on the access control area with respect to the IP protocol suite, specially regarding layer management functions. By designing and implementing a capability-based access control policy we show that RINA allows finer granularity in access control decisions than current approaches. Moreover, the policy can be applied to any layer in the network, maximizing the re-use of specifications and implementations