Back to Volume
Paper: The Earth Data System and The National Information Infrastructure Testbed
Volume: 61, Astronomical Data Analysis Software and Systems III
Page: 45
Authors: Christian, C. A.; Murray, S. S.
Abstract: Earth science data is presently stored in many different archives with many different catalog and access methods. These range from large, government archives of satellite imagery with full data management and access services to small, field data sets under the control of an individual scientist. The prototype Earth Data System (EDS) is designed to provide access of such heterogeneous data researchers interested in global weather changes, ocean current movements, and coastal drainage. In particular, the EDS includes a collaborative data browser and analysis capability as an example of a truly distributed computing capability that allows people at geographically separate institutions to search, view, and analyze data together. The EDS is a testbed application being developed by the National Information Infrastructure Testbed organization; a consortium of industry, government and academic institutions. The application is constructed using an implementation of OSF's Distributed Computing Environment (DCE) software, wrapped in a software toolkit reminiscent of the Astrophysics Data System toolkit. It uses a communications network of high-speed cross-country DS3 (45Mbps) lines coupled with ATM switches, and routers with SMDS, FDDI, and Ethernet interfaces. The initial tests in May, 1993, showed outstanding performance with network nodes distributed between California and New Mexico locations linked by a T3 network. The functional system will be demonstrated in November 1993 at several forums including Supercomputing 93. Future developments will incorporate distributed real time device control and data retrieval as well as high speed computing, data retrieval and network management. In this paper we will discuss the EDS reference application as well as several of the technical issues regarding the full deployment of large scale distributed computing infrastructure.
Back to Volume