🤖 AI Summary
This paper addresses the fair termination problem for distributed concurrent programs in resource-constrained active object systems. Methodologically, it introduces a core calculus integrated with a graded semantics and fairness guarantees, pioneering the combination of a graded resource model with fairness-enforcing mechanisms from synchronous session types; type rules statically constrain object behavior and resource consumption to ensure weak fairness—i.e., every enabled action is eventually scheduled—along all global execution paths. Contributions include: (1) the first active object calculus supporting *joint verification* of resource awareness and fair termination; (2) a proof of type safety and *strong fair termination*: all well-typed programs provably terminate fairly under any fair scheduler; and (3) a formal foundation for high-assurance distributed systems that simultaneously ensures resource controllability and liveness guarantees.
📝 Abstract
Active object systems are a model of distributed computation that has been adopted for modelling distributed systems and business process workflows. This field of modelling is, in essence, concurrent and resource-aware, motivating the development of resource-aware formalisations on the active object model. The contributions of this work are the development of a core calculus for resource-aware active objects together with a type system ensuring that well-typed programs are fairly terminating, i.e., they can always eventually terminate. To achieve this, we combine techniques from graded semantics and type systems, which are quite well understood for sequential programs, with those for fair termination, which have been developed for synchronous~sessions.