Principals, Policies and Keys in a Secure Distributed Programming Language

KDLM is a model of information flow control for distributed programming languages, introducing a distributed notion of selective declassification. KDLM uses type-level notions of both principals and policies to model security policies incorporating access control and information flow control, with run-time representatives based on cryptographic keys. This article addresses the question of how this control is enforced when programs cross address spaces and networks. An object calculus is introduced that extends KDLM. This uses a notion of locally opaque principal and key names in objects to transmit and enforce restrictions across address spaces, with key equality used as a form of runtime type discrimination to reveal “opaque” principal names at the receiver.