Production Grids

One of the main purposes of Grid Technologies is to enable the secure sharing of resources across organization boundaries. This sharing, not subject to a centralized control, has lead to the new collaborative concept of Virtual Organizations (VOs). A VO is identified as the set of resource providers and consumers who have agreed to a common set of rules specifying who and what is allowed to share and under which conditions. The VO concept has grown rapidly in the past few years in the scientific community where people are working together to achieve the solution of large-scale problems which range various fields: from high energy physic to earth observation to environmental management to biomedicine. Many main research institutions dealing with such computing or/and data intensive challenges looked towards Grid Computing as a possible solution to match their needs and have received and invested many funds to build a secure, reliable, widespread Grid Computing infrastructure. These infrastructures are called Production Grids and offer reliable Grid services over a set of heterogeneous computing and storage resources in a dynamic environment. The production quality Grid services are sustained by an organization that rely upon many individuals (system administrators, middleware developers,. . .) and that is capable of supporting the use by large scientific communities. Many Production Grids have the same basic middleware that provide a full set of Grid foundation services (necessary for interoperability with other Grid systems) as well as own-developed middleware providing a range of specialist higher level services. All the offered services serve as an abstraction layer for users to access resources transparently with a transparent support for multiple VOs. In this work we'll give a detailed description of the main existing Production Grids: EGEE, TeraGrid, Open Science Grid, DEISA, NorduGrid, National Grid Service (NGS), PRAGMA and Grid Australia, what kind of services they provide, how do they differ and why.