Interactions with Old and New Partners

Since the last annual report, ADL has continued to work with significant number of partners from the commercial, government, and university sectors. Old partners with whom a significant level of continuing or new activity occurred during the past year include: the California Digital Library (CDL); the National Mapping and Imagery Agency (NIMA); Digital Equipment Corporation (DEC); Environmental Systems Research Institute (ESRI); GEOREF; Hughes; Informix; Microsoft NASA; Oracle; San Diego Supercomputer Center (SDSC); the Sierra Nevada Ecosystem Project (SNEP); Utah State University and the Mojave Database Cooperative; the United States Geological Survey (USGS); the United States Navy NAVO facility at Stennis and the United States Navy NRaD facility at San Diego.

Partner Interactions:

• United States Geological Survey continues to work with ADL in cartographic issues, the building of the domestic ADL gazetteer and as an active ADL board advisor. The Colorado subteam of the User Interface Evaluation team completed a series of focus groups to assist U.S. Geological Survey evaluate a prototype interface design for their planned National Atlas of the United States. The product will be a Web-based atlas of maps, data and imagery available from federal data producers, targeted for the lay public and in schools. USGS is interested in further interface evaluation efforts, including transaction log monitoring by the CU Team once a working Web prototype is established. Discussions for USGS funding for that effort are underway.

• In 1997 ADL subcontracted with Microsoft to help develop a 1 TB image database. This database, called "Terra Server", was demonstrated publicly on 20 May 1997 in New York by Microsoft senior researcher (and ADL board member) Jim Gray. Approximately one-fifth (200 GB) of the Terra Server's contents are image data processed by ADL at UCSB. Further information about the Terra Server may be found in http://www.research.microsoft.com/ research/ barc/ Gray/ Two_Commodity_Scaleable_Servers.doc As a result of the experience gained in the Terra Server project, ADL made a commitment to wide-area high-resolution image data as fundamental online collection. Microsoft also continues to contribute software and documentation to ADL.

• In 1997 ADL received an Educational Grant from Informix Corp. Under this grant, ADL has received free software, documentation, training, consulting, and technical support valued at approximately $200K retail, as well as an ongoing 50% discount for further training, and assistance obtaining free or discounted products and services from Informix OEMs.The Informix Dynamic Server / Universal Data Option is ADL's primary database management system.

• Digital Equipment Corporation continues to be an active partner in ADL. They provide support in the form of hardware engineering, discounted systems and active input through the ADL advisory board.

• Earth Data Analysis Center, University of New Mexico and ADL has completed the first phase of adding EDA metadata to ADL. ADL will provide the Web access service for EDA while EDA maintains the actual data at their site. They are scanning historic coverage of aerial photography and mapping of New Mexico for their research clients.

• We have conducted research in the Navy NRaD CONVEX shared memory machine to extend our work in the SWEB/SWEB++ projects. For the shared memory machine, CPUs can be partitioned as subcomplexes and normally different users access different subcomplexes based on the work load needs. We have developed a strategy to monitor the performance of subcomplexes and perform load balancing between subcomplexes. If all CPUs are clustered as one subcomplex, then OS scheduling will dominate the optimization. Our analysis shows that our scheduling outperforms the CONVEX OS process auto-scheduler substantially if we partition the system into several subcomplexes. Load balancing issues have been previously studied extensively, but we think many of previous work are not fully implemented, and experiment results are conducted mainly using simulation. Our analysis based on real-world experiments motivates us to develop new techniques that work in practice. Our experiments also show that resource monitoring is important to gain performance. We also conducted adaptive client-server scheduling research on the CONVEX machine for image browsing. The project is funded by NRaD.

• Hughes Information Technology Systems. Xun Cheng designed a relational database model for the gazetteer based on the proposed gazetteer content standard. A thesaurus of feature types was created to categorize gazetteer entries.

It is clear that these partnerships are not only beneficial to ADL, but are in may senses essential to the success of the project, both in terms of supporting important research activities and in moving towards an operational digital library.