Dr Oliver Keeble (CERN)
Grid computing as currently understood is normally enabled through the deployment of integrated software distributions which expose specific interfaces to core resources (data, CPU), provide clients and also higher level services. This paper examines the reasons for this reliance on large distributions and discusses whether the benefits are genuinely worth the considerable investment involved in their maintenance. Looking ahead to a context of mature standards, pervasive virtualisation and administrative decentralisation, is it time to embrace alternative models in order to optimally enable a grid infrastructure?
|Presentation type (oral | poster)||oral|