China Grid and Related Dependability Research - Grid Computing: Infrastructure, Service, and Applications - page 83

The basic realization of a digital virtual human is to identify and classify
inner viscera, rebuild the edge proi le data of outer body and three-
dimensional proi le data of inner viscera, construct a high-precision digital
human grid, and use these data to achieve the visualization of a human
body. In addition to the massive amount of body slices (a suit of raw data
requires 20 GB storage space), the amount will increase greatly for mesh
construction. Only a computer resource grid can afford this processing.
Before we analyze and dispose these data, some optimized resource
scheduling algorithms must be brought forward, such as the highly precise
algorithm for contour extraction and highly precise parallel algorithm for
reconstructing three-dimensional grids of the digital virtual human.
Remote sensing can be dei ned as the collection of data about an object
from a distance. Geographers use the technique of remote sensing to monitor
or measure phenomena found in the Earth's lithosphere, biosphere, hydro-
sphere, and atmosphere. Remote sensing imagery has many applications
in mapping land use and cover, agriculture, soil mapping, forestry, city
planning, archeological investigations, military observation, and geomor-
phological surveying, among other uses.
Medical imaging processes are dedicated to the storage, retrieval, distri-
bution and presentation of images, which are handled by various modali-
ties, such as ultrasonography, radiography, magnetic resonance imaging,
positron emission tomography, and computed tomography. It consists of
a central server that stores a database containing the images. A doctor and
his patients can get more reliable and precise medical images (such as
X-ray slices, CT, and MR) with more quick, exact, and simple modes, so as
to increase the rate of inchoate morbid detection and diagnoses and living
opportunity.

4.4.3 Computational Fluid Dynamics Grid

The computational l uid dynamics (CFD) grid provides the infrastructure
to access different CFD applications across different physical domains
and security i rewalls. By dei ning a standard CFD workl ow and general
interfaces for exchanging mesh data, the platform facilitates interoperat-
ing between CFD applications. An application development package
(ADP) is being developed and an application development specii cations
(ADS) is being dei ned, shipping with the platform to make migrating
CFD applications feasible. The major function of ADP is to schedule jobs,
recover from faults, and discover services effectively.
The CFD platform includes two levels: a high level for service-oriented
applications (SOA) and a low level for parallel computing. The physical
architecture of the CFD platform is shown in Figure 4.3 .
In the high level, all services are registered in the main hub or subhubs.
The function of the main hub is to provide job dei nition. Service requestor
visits the main hub for the service interface dei nition and i nds the