A security model for functional active objects with an implementation in Erlang

In this paper we describe a language based security model for distributed computing based on functional active objects intended to build a basis for Distributed Information Flow Control (DIFC) and contrast it to earlier models in particular the prominent model by Myers and Liskov. We carefully motivate the assumptions concerning secure communication in distributed object-oriented scenarios, show that they naturally lead to futures as the security abstraction of object-oriented message passing, and motivate our assumptions by an example. Finally, the stepping stones to a formal foundation of the model are summarized: the notion of visibility, the security classes, and a definition of information-flow security for distributed active objects. This paper further describes how the concept can be realized using the novel tool box Erlang Active Objects for functional active objects. We introduce Erlang and Erlang Active Objects, and show how our model of decentralized multilateral security can be implemented by a simple algorithm for a dynamic security check for Erlang Active Objects.