Mapping Description Communication Pattern Resources and a between Grid Mechanism Scheduling

No application can execute efficiently on the grid that is not aware of the fact that it runs in an heterogeneous network environment with heterogeneous nodes. We report on an ongoing work whose goal is to develop a distributed software framework and an API for grid computing which shall empower applications to perform scheduling decisions on their own and utilizing the information about the grid environment in order to adapt their algorithmic structure to the particular situation. Since our solution hides low-level grid-related execution details from the application by providing an abstract execution model, it is able to eliminate some algorithmic challenges of nowadays grid programming.