Universal model simulation: BG and extended BG as examples

Abstract

This paper focuses on simulations as a means of deriving the relative power of distributed computing models. We describe an abstract simulation algorithm that enables reducing the question of solvability of a generic distributed task in one model to an equivalent question in another model. The technique implies simple equivalents to the fundamental reduction by Borowsky and Gafni, known as BG simulation, as well as to Extended BG, a more recent extension of it to colored tasks. We also sketch how the parameters of our technique can be tuned to derive recent equivalence results for models that use, in addition to basic read-write memory, k-set agreement or k-process consensus objects, or make assumptions on active resilience. © Springer International Publishing 2013.