| Many of the grid systems that exist today are data intensive and must be capable of transporting very large amounts of data, both static and in real time. The ability to access and manage data is another important area in computational and data grids. This involves moving data between grid resources, and/or allowing resource access to the data, On Demand. These kinds of data movement activities are very common in job schedulers . Figure 14.1 shows the logical layers in grid architecture. The base services and the data services together provide most of the basic facilities we have previously discussed. These are required for high-level grid applications and services, including job managers, schedulers, resources, and provisioning agents . These high-level services and applications utilize these base services for resource discovery, classification, resource allocation, resource usage planning, monitoring, and scheduling. Figure 14.1. This illustration depicts a logical view of the levels of grid services.  It is recommended that one implement a middleware tool to provide some of these base services. Globus Toolkit (GT3) provides a number of these high-level services. These high-level services are utilizing the core GT3 framework as the fundamental set of building blocks. In GT3, there are information services, resource allocation services, and data services available to the developer. We will discuss each of these component details in the coming pages. |
