Grid Computing - Guide to Cloud Computing for Business and Technology Managers - page 250

connecting databases from two separate VOs (representing two data streams
from physics sensors/detectors) in order to output their correlation. Such
a service could not be provided by one organization or the other since the
output relies on the combination of both.
The term computational grid comes from an analogy with the electric power
utility grid. A computational grid is focused on setting aside resources spe-
cifically for computing power and uses networks of computers as a single,
unified computing resource. It is possible to cluster or couple a wide variety
of resources including supercomputers, storage systems, data sources, and
special classes of devices distributed geographically and use them as a single
unified resource. Such pooling requires significant hardware infrastructure
to achieve the necessary interconnections and software infrastructure to
monitor and control the resulting ensemble. The majority of the computa-
tional grids are centered on major scientific experiments and collaborative
environments. Computational grid applications exhibit several functional
computational requirements. These include the ability to manage a variety
of computing resources, select computing resources capable of running a
user's job, predict loads on grid resources, and decide about resource avail-
ability, dynamic resource configuration, and provisioning. Other useful
mechanisms for the management of resources include failure detection,
failover, and security mechanisms.
A data grid is responsible for housing and providing access to data across
multiple organizations and makes them available for sharing and collabora-
tion purposes. These data sources can be databases, file systems, and storage
devices. The requirement for managing large data sets is a core underpin-
ning of any grid computing environment. Data grids can also be used to
create a single, virtual view of a collection of data sources for large-scale
collaboration. This process is called data federation . In data grids, the focus is
on the management of data that are being held in a variety of data storage
facilities in geographically dispersed locations. For example, medical data
grids are designed to make large data sets, such as patient records contain-
ing clinical information and associated digital x-rays, medication history,
doctor reports, symptoms history, genetic information, and so on, available
to many processing sites. By coupling the availability of these massive data
sets with the large processing capability of grid computing, scientists can
create applications to analyze the aggregated information. Searching the
information for patterns or signatures enables scientists to potentially reach
new insights regarding the environmental or genetic causes of diseases.
Data grid systems must be capable of providing data virtualization services
to provide the ability to discover data, transparency for data access, integra-
tion, and processing as well as the ability to support flexible data access and
data filtering mechanisms. In addition, the provision of security and privacy
mechanisms for all respective data in a grid system is an essential require-
ment for data grids.