Interoperability of Grid component models: GCM and CCA case study

Abstract

This paper presents a case study in the generic design of Grid component models. It defines a framework allowing two component systems, one running in a CCA environment, and another running in a Fractal environment, to interact as if they were elements of the same system. This work demonstrates the openness of both Fractal and CCA component models. It also gives a very generic and exhaustive overview of the interaction strategies that can be adopted to allow full integration of these two models, like strategies for reusing in Fractal single components from the CCA world and connecting a Fractal system to an already running CCA assembly. Finally, it presents the implementation and results of investigation of interoperability between two given component frameworks: MOCCA and ProActive. In generall, this paper presents the key concepts useful to make any two component models interoperate.

GCM: a grid extension to Fractal for autonomous distributed components

“…Interoperability between GCM and other standards or component models has been demonstrated, first through effective interactions between CCA and GCM components [30], and second by the possibility to expose component interfaces as web services [19].…”

The design of a CCA framework with distribution, parallelism, and recursive composition

“…It improves the ability to deal with the complexity of large scale applications. Recursive composition is not supported by CCA frameworks [21]. Without recursive composition, CCA frameworks are less suitable to work with fine-grained component ensembles compared to Fractal/GCM and HPE.…”

“…The support for non-functional concerns is another important feature that distinguishes Fractal/GCM from CCA [21]. GCM designers argue that "the absence of standard interfaces dedicated to non-functional concerns makes it hard to achieve certain features, for instance, dynamic reconfiguration based on observed performance or failures" [6].…”

Design Pattern Usage in Grid Computing

“…Here, we found 9 occurrences of 6 different behavioral design patterns and 11 occurrences of 4 different structural design patterns respectively. Flexible design of the dispatcher architecture using a plug-in structure [13]; Interoperability of Grid components [24] Chain of Responsibility Behavioral Allows concurrent usage of multiple implementations of a particular plug-in. [12]; Plugin discovery [13] Command…”