Sharing a conceptual model of grid resources and services

Abstract

Grid technologies aim at enabling a coordinated resource-sharing and problem-solving capabilities over local and wide area networks and span locations, organizations, machine architectures and software boundaries. The heterogeneity of involved resources and the need for interoperability among different grid middlewares require the sharing of a common information model. Abstractions of different flavors of resources and services and conceptual schemas of domain specific entities require a collaboration effort in order to enable a coherent information services cooperation. With this paper, we present the result of our experience in grid resources and services modelling carried out within the Grid Laboratory Uniform Environment (GLUE) effort, a joint US and EU High Energy Physics projects collaboration towards grid interoperability. The first implementation-neutral agreement on services such as batch computing and storage manager, resources such as the hierarchy cluster, sub-cluster, host and the storage library are presented. Design guidelines and operational results are depicted together with open issues and future evolutions.