Scientists but also industrial researchers nowadays have unprecedented computing and instrumental capability for studying and simulating natural phenomena at greater accuracy, and this leads to the production of huge amount of data. To store and share these data collections is now emerging the use of the Grid as a very large data repository. Advantages are the enormous availability of storage resources and the share of data, a disadvantage is the geographically distributions of data, and the remoteness of users and data. In particular data transmission represents a bottleneck in data visualization across the Grid, because of the amount of data and the complexity of 3D models, and the actual and promised increases in phone and network bandwidth will not suffice to solve this problem. On the other side there is an increasing research activity to design efficient and effective data simplification and compression algorithms in order to make the transmission over the network of large 3D models a feasible task. The use of local parallel processing to make these algorithms even more efficient and satisfy real-time equirements of interactive distributed applications over the Grid is discussed. A specific class of algorithms based on the corner-table data structure is considered, and problems and possibilities related to the parallelisation of these algorithms on cluster of workstation are addressed.
Parallel Computing for 3D data visualization and transmission in Grid based applications.
D. D'AGOSTINO;GIANUZZI, VITTORIA;
2002-01-01
Abstract
Scientists but also industrial researchers nowadays have unprecedented computing and instrumental capability for studying and simulating natural phenomena at greater accuracy, and this leads to the production of huge amount of data. To store and share these data collections is now emerging the use of the Grid as a very large data repository. Advantages are the enormous availability of storage resources and the share of data, a disadvantage is the geographically distributions of data, and the remoteness of users and data. In particular data transmission represents a bottleneck in data visualization across the Grid, because of the amount of data and the complexity of 3D models, and the actual and promised increases in phone and network bandwidth will not suffice to solve this problem. On the other side there is an increasing research activity to design efficient and effective data simplification and compression algorithms in order to make the transmission over the network of large 3D models a feasible task. The use of local parallel processing to make these algorithms even more efficient and satisfy real-time equirements of interactive distributed applications over the Grid is discussed. A specific class of algorithms based on the corner-table data structure is considered, and problems and possibilities related to the parallelisation of these algorithms on cluster of workstation are addressed.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.