Composable Semantic Models for Actor Theories

  • Authors:

  • Carolyn L. Talcott

  • Affiliations:

  • Stanford University, Computer Science Departments, Stanford, CA 94305. clt@cs.stanford.edu

  • Venue:
  • Higher-Order and Symbolic Computation
  • Year:
  • 1998

Quantified Score

Hi-index 0.00

Visualization

Abstract

We are interested in developing a semantic foundation that supportsspecifying, composing, and reasoning about components of open distributedsystems. The actor model provides the basic elements for open distributedcomputation: encapsulation of state; independent concurrent units ofcomputation; interaction; and dynamic creation and interconnection. Inorder to provide for composability, and for reasoning about properties atmany different levels of abstraction we introduce the notions of actorcomponent, actor component algebra, and actor component algebra morphism.Morphisms from syntactic to semantic algebras give composable semantics.We first illustrate features of the actor model and a variety of componentalgebras and morphisms through a series of simple examples. We then definethree semantic models for actor computation starting with a generalizationto open systems of Clinger‘s event diagram model, and forming twoabstractions: interaction diagrams and interaction paths. An algebra isdefined on each semantic domain with operations for parallel composition,hiding of internal actors, and renaming, and we show that component algebralaws hold. We use these models to provide semantics for descriptions ofactor components based on actor theories and show that the semantics is acomponent algebra homomorphism.