A Web-based Scientific Data Server for Accessing and Distributing Earth Science Data
	Liping Di, R. Suresh, K. Doan and Doug Ilg Hughes STX Corporation 7701 Greenbelt Road, Suite 400 Greenbelt , MD 20770 Liping.Di@gsfc.nasa.gov Phone: 301-441-4104 Fax: 301-441-4335	Ken McDonald NASA/GSFC, code 505 Greenbelt, MD 20771

In recent years, one of most exciting developments in information technologies is the proliferation of Internet connectivity and the popularity of WWW services. Web technologies have provided a convenient way to access multimedia information through the Internet. Almost all users with Internet access use Web browsers to access information. Although currently most information accessed through the Web are descriptive, the possibility of distributing scientific data through the Internet using the Web will immediately offer scientific datasets to millions of potential users with little effort. Therefore, the data will be more widely used, and cost savings in data distribution will be achieved. Currently, many prototyping efforts have been done to show the applicability of the Web technologies for the scientific data distribution. The particular one which promotes NASA data standards and provides accesses to NASA datasets is the prototype Data and Information Access Link (DIAL) system developed by Hughes STX Corporation and National Center for Supercomputing Applications (NCSA) with NASA DIAL is the software designed specifically for type of what could be generally termed a "scientific data repository", which is designed to provide scientific data over networks. There will soon be many thousands of such repositories, varying in scale from large global archives to science teams or even individual scientists. These repositories must all interoperate as seamlessly as possible, while operating in heterogeneous computing environments. Groups of interoperating servers will form one or more federations of repositories. The key software technologies to create the DIAL system are: 1) catalog interoperability, and 2) efficient access to complex scientific data objects over networks.

The development of DIAL system is a on-going process. The fully-developed DIAL system will consist of: 1) a catalog interoperability layer to support major protocols, such as Z39.50, CEOS CIP, EOSDIS V0, for interoperation among DIAL servers and between DIAL and other data systems in support of NASA concept of the federation of small data producers; 2) a scaleable scientific data server to serve complex scientific data and data objects in multiple formats and to allow data users to interactively manipulate the selected data so that they can obtain the data in their favorite form in terms of spatial and temporal coverage and resolution, parameters, and format. The data server will address the dynamic and object serving; 3) an open database interface layer to interface with Open Database Connectivity (ODBC) capable metadata catalog databases for powerful data search and finding; 4) a user interface layer to support user-friendly interaction between client and server; and 5) a search engine with a file-based catalog database to provide basic metadata database capability in case the data producer does not have an expensive commercial database system. In addition, a suite of software tools will be available to help data producers to ingest the data into the system and to help data users to download and analyze the data.

Currently a prototype of the DIAL system has been developed. The prototype system has all above mentioned components in some stages of the development. The prototype DIAL is a compact yet powerful Web client-server based data browsing and distribution system, providing low-cost on-line data access to users through the Internet. DIAL enhances data access and interoperability through a simple web interface. It permits data producers to set up a low end workstation as a data server for distributing their data and metadata quickly and easily. Users can access the data server through any web browser, search and query the data archival based on geographic location, time, and other relevant parameters, locate data of interest, view the metadata, browse the data, view, subsample, subset, and download the data to local disk. Data may be download in one of three formats (HDF, ASCII, and binary), or displayed on the screen as a GIF image, HTML text or tables. Binary executables of the prototype software are free to any interested people, and are available on all major UNIX platforms, and will soon be available on the Window/NT platform. The address for the demo site of this software is http://hops.stx.com:8080/dialhome.html.