Experiments on a Grid Layer Prototype for Shared Data Programming Model

In the grid context, there is little support for programming paradigms such as shared data or associative programming. We have previously proposed an original idea to attack shared data programming on the grid by making use of both relaxed consistency models and user specified type consistency in an object oriented model. In this paper we present the experimental results of GUN, a lightweight Java prototype implementation of our shared data programming model. In order to assess GUN's performance, some generic experiments are designed. First, GUN's correctness has been proved in a real grid deployment where a limited number of nodes were available. In order to observe GUN's scalability with an increasing number of nodes, experiments have been conducted in a simulated large scale grid environment. The conducted experiments have shown a good scalability of GUN, as well as promising results in terms of response times and usability.