HR: 1340h
AN: IN43B-0334    [Abstracts]
TI: Grid Based Techniques for Visualization in the Geosciences
AU: * Bollig, E F
EM: bollig@msi.umn.edu
AF: Dept. of Geology and Geophysics and Minnesota Supercomputing Institute, University of Minnesota -- Twin Cities, Minneapolis, MN 55455 United States
AU: Sowell, B
EM: benny@msi.umn.edu
AF: Dept. of Geology and Geophysics and Minnesota Supercomputing Institute, University of Minnesota -- Twin Cities, Minneapolis, MN 55455 United States
AU: Lu, Z
EM: zhenyulu@cs.fsu.edu
AF: Florida State University, 489 Dirac Science Library, Tallahassee, FL 32306-4120 United States
AU: Erlebacher, G
EM: erlebach@csit.fsu.edu
AF: Florida State University, 489 Dirac Science Library, Tallahassee, FL 32306-4120 United States
AU: Yuen, D A
EM: davey@krissy.geo.umn.edu
AF: Dept. of Geology and Geophysics and Minnesota Supercomputing Institute, University of Minnesota -- Twin Cities, Minneapolis, MN 55455 United States
AB: As experiments and simulations in the geosciences grow larger and more complex, it has become increasingly important to develop methods of processing and sharing data in a distributed computing environment. In recent years, the scientific community has shown growing interest in exploiting the powerful assets of Grid computing to this end, but the complexity of the Grid has prevented many scientists from converting their applications and embracing this possibility. We are investigating methods for development and deployment of data extraction and visualization services across the NaradaBrokering [1] Grid infrastructure. With the help of gSOAP [2], we have developed a series of C/C++ services for wavelet transforms, earthquake clustering, and basic 3D visualization. We will demonstrate the deployment and collaboration of these services across a network of NaradaBrokering nodes, concentrating on the challenges faced in inter-service communication, service/client division, and particularly web service visualization. Renderings in a distributed environment can be handled in three ways: 1) the data extraction service computes and renders everything locally and sends results to the client as a bitmap image, 2) the data extraction service sends results to a separate visualization service for rendering, which in turn sends results to a client as a bitmap image, and 3) the client itself renders images locally. The first two options allow for large visualizations in a distributed and collaborative environment, but limit interactivity of the client. To address this problem we are investigating the advantages of the JOGL OpenGL library [3] to perform renderings on the client side using the client's hardware for increased performance. We will present benchmarking results to ascertain the relative advantage of the three aforementioned techniques as a function of datasize and visualization task. [1] The NaradaBrokering Project, http://www.naradabrokering.org [2] gSOAP: C/C++ Web Services and Clients, http://www.cs.fsu.edu/~engelen/soap.html [3] JOGL, https://jogl.dev.java.net/
DE: 9820 Techniques applicable in three or more fields
SC: Earth and Space Science Informatics [IN]
MN: Fall Meeting 2005