Secure Implementation of Asynchronous Method Calls and Futures

Programming languages suitable for distributed computation contain constructs that should map well to the structure of the underlying system executing the programs, while being easily usable by the programmers and amenable to computer-aided verification. For object-oriented languages, asynchronous method calls returning futures that will be filled only after the called method has finished its execution have been proposed as a reasonably simple and analyzable programming construct. In this paper, we show how to map from a language with asynchronous method calls and futures to a language with explicit communication primitives and cryptographic operations. Our target language is reasonably similar to common process calculi, and translating it further to e.g. the applied pi calculus requires only known techniques. The translation is valid even for programs executing in open environments, where method calls and futures can be transmitted between the program and the environment.