Types and Effects for Resource Usage Analysis

An extension of the $\lambda$-calculus is proposed, to study resource usage analysis and verification. Resources can be dynamically created, and passed / returned by functions; their usages have side effects, represented by events. Usage policies are properties over histories of events, and have a possibly nested, local scope. A type and effect system over-approximates the set of histories a program can generate at run-time. A crucial point solved here concerns correctly associating fresh resources with their usages within approximations. A second issue is that these approximations may contain an unbounded number of fresh resources. Despite of that, we have devised a technique to model-check validity of approximations. A program with a valid approximation is resource-safe: no run-time monitor is needed to safely drive its executions