In this paper, we present a new security model for distributed active objects. This model emphasizes the aspects of decentralisation and private data of objects. We consider principals as active objects thereby amalgamating subjects and objects into one concept providing a simple uniform security model based on visibility of objects and object local security specification of method accessibility. Decentralized security policies are possible in which every principal has some data that is inaccessible to others. We introduce this new security model algebraically as a semi-lattice contrasting it to the foundations of lattice-based security models similar to Denning's work. As a proof of concept we show how the model can be naturally interpreted for a calculus of active objects.
展开▼
