Dynamic Composition of Coordination Abstractions for Pervasive Systems: The Case of LogOp

This paper is about the LogOp coordination model, and has a twofold goal: to provide it with deeper foundation, and to express it in terms a more fitting formal semantics. Accordingly, we first define a new criterion for interpreting, comparing and possibly integrating coordination models based on the dynamic composition of coordination abstractions—feature that makes them suitable for the engineering of complex computational systems like pervasive and knowledge-intensive ones. Then, we present the main features of LogOp, and show that linear logic, the logic of consumption of resources, represents the most suitable choice as the conceptual framework for the formal specification of LogOp semantics. Given the well-known intricacies and subtleties of coordination models for distributed systems, we argue that a well-defined and expressive formal semantics is an essential requirement for a foreseeable scenario where abstractions and mechanisms from diverse coordination models (in particular, those discussed here) are exploited altogether in the engineering of complex systems like pervasive ones.