Collective Interfaces for Distributed Components

Abstract

We propose to address collective communications in distributed components through collective interfaces. Collective interfaces handle data distribution, parallelism and synchronization, and they expose collective behaviors in the definition of components. We show, as an illustration, that collective interfaces allow the encoding of SPMD programming in a better structured and less error prone way. We verify the scalability and performance of collective interfaces in an experiment on up to 100 machines.

GCM: a grid extension to Fractal for autonomous distributed components

“…However, we present the particular case of remote method invocations in the remaining of this section because of its richer implications on typing of interfaces and on the component composition. Experiments on the implementation of collective communications for components interacting by asynchronous remote method invocations have been conducted over the ProActive middleware, and proved to be quite efficient and convenient to program distributed applications [7]. However, the notions of multicast and gathercast interfaces are clearly also adapted to other communication semantics, the consequence on type compatibility between interfaces can be inferred from the case presented in this section.…”

Advanced Grid Programming with Components: A Biometric Identification Case Study

“…A further grid specific Fractal extension introduced by GCM is the notion of collective interfaces [9]. This eases parallel programming and allows exposing the collective behaviour of a component on its interface level.…”

On the design of abstract binding connectors for high performance computing component models

“…In general, the cases 1-N, M-1, and M-N are treated separately. Fractal also has been recently extended with collective interfaces of [7].…”