Taxonomy of Grid Systems

2.1. Grid Computing

Basically, grid computing (Foster & Kesselman, 2003) is a relatively new distributed system paradigm where computational resources are coupled together to form a large-scale distributed system where all resources are available for sharing. This has the great advantage of providing a resource-rich infrastructure capable of solving data intensive and complex computational problems such as protein folding and weather forecasting in an acceptable time and at a reasonable cost.

Indeed, there are as many definitions to the grid as the growing number of organizations utilizing it. A common theme underlying these definitions is the coordinated resource sharing and problem solving in a virtual organisation (VO). A VO is a dynamic set of participants defined around a set of resource sharing rules and conditions as shown in Figure 1. Some grid definitions add additional criteria requiring the grid resources to be distributed across multiple administrative domains (Foster, 2001) (Joseph & Fellenstein, 2003) or to be geographically distributed (Buyya, 2005). These additional criteria exclude clusters, where shared resources are usually in the same locality and administrative domain, from the grid definition. However, some leading grid authorities, such as Sun Microsystems (Sun Microsystems, 2009), do consider clusters as grid environments, they use the term ‘Cluster Grids’ to refer to them (Sun Cluster Grid, 2002).

Figure 1.

Grid environment