Introduction

The GPS Environmental and Earth Science Information Systems (GENESIS) program, a member of NASA's Earth Science Information Partners (ESIP) Federation, supports archiving of the GPS occultation data and provides web-accessible database, process, and visualization services. For the past two years at USC's Information Laboratory, we have been designing and developing several tools and components to support the underlying database requirements of GENESIS.

Demos:

SOAR Application
We're implementing Sun's Java Web Start. Please try that version of the application. Does nothing happen? Try manually getting Java Web Start before running our application.

Soar Canned Queries
Try our new, experimental, "canned query" webpage. It permits one to seach our databases without executing our front-end application.
Here is yet another, experimental canned query webpage. It's been implemented as a web service with IBM WebSphere Application Server Enterprise Edition and WebSphere Studio Application Developer. This one gives you the result in an XML-like format.

Equipment:

Currently, we have a working prototype consisting of the following pieces:

1. Server

We use an object-relational database management system (OR-DBMS) — in particular, Informix Universal Server v9.2 that runs atop a SUN 250 Enterprise Server with Solaris 2.6. The hardware has two processors, 512 MB of RAM, and approximately 250 GB of disk space. We designed and implemented an OR-based conceptual schema for the metadata of the three product levels of occultation data: level 1a, 1b, and 2, as well as the acutal data from level 2. All of the levels are now populated with real data.

2. Client

A Java-based application. The graphical user interface is implemented in Java and can support queries on the metadata of all three different levels, as well as on level 2 data. In addition, the interface is map-based in that a user can select an area of interest and ask for all the available level 2 occultation data within that area. The client submits SQL3 queries to the server through JDBC and RMI connections.

3. Middleware

The middleware is written in Java and is based on infolab's RIMS toolkit.

4. Remote sources

A websource containing radiosonde data is wrapped and its information can be queried and integrated with the results of the GENESIS-localized queries. We utilize the Ariadne mediator architecture for this purpose.

5. Server utilities

We use the Informix Geodetic datablade, which provides spatial support for our database. Therefore, the latitude and longitude of occulation data points are stored in the database as geospatial coordinates. This eases spatial queries on the data. In addition, the spatial attributes are indexed using a spatial index structure (R-tree) for efficient execution of spatial queries.

6. Client utilities

An XML uploader was developed in Java so that occultation files can be remotely and automatically uploaded from a server at JPL into our database server at USC.

7. Data mining utilities

We believe that these techniques are immediately applicable to level 2 and 3 of occultation data and our system will soon be ready to incorporate these new techniques.

Pictures:

Architecture:


People:

Publications:

Contact us