SCD is building a web portal that will empower the scientific investigator by providing easy access to SCD's array of services, resources, and information in a highly customizable interactive browser-based interface. Researchers will personalize their view of SCD information by selecting not only what they want from us, but also where that information will appear in their browser's window. This allows researchers to get just the information they need, when they need it.

This is a sample of what researchers will be able to do from the SCD portal:

• Submit and check the status of jobs on the computing resources
• Manage their holdings on the Mass Storage System
• View their accumulated charges
• Interact with other researchers in a collaborative environment for common interests
• Access news, data, and information provided from various sources relevant to their area of research in the atmospheric and related sciences

SCD's strategy involves acquiring and using the machine architecture known as clustered Symmetric Multi Processor (SMP) machines. We will continue to make the best use of available SMP clusters from a variety of vendors to provide the highest computing capability and capacity for NCAR and University scientists.

As part of this effort, SCD will collaborate with the science divisions to develop and release a Request for Proposal (RFP) for a new Advanced Research Computing System (ARCS). Release of the RFP is targeted for late summer 2000.

SCD's Mass Storage System is often called one of the "crown jewels" of NCAR's research-enabling facilities. SCD seeks to provide optimal mass storage services for researchers and enhance the web interface to Mass Storage System data.

As part of this effort, SCD will:

• Continue to provide the best possible mass storage services, while balancing performance and cost
• Provide an enhanced web presence through the modernization of mass storage services interfaces and tools
• Design these interfaces and tools to be used in concert with the larger framework of SCD's computing environment

SCD is dedicated to establishing a premier program in visual data analysis for the earth system sciences. SCD will take a leadership role in enabling wide-area access, manipulation, analysis, and visualization of very large, very complex datasets. Our aim is to develop effective problem-solving environments for knowledge development and sharing. Primary thrust areas include:

• Data access facilities and frameworks for enhanced electronic provision of important community datasets
• A Visual Computing Collaboratory that combines visual supercomputing, large-scale virtual data exploration, and wide-area collaborative visualization environments
• Analysis and visualization applications and frameworks for earth system data
• Collaborative data access, analysis, and visualization efforts with scientific programs
• Collaborative research and development efforts with other labs and universities

Networking capabilities fundamentally support NCAR/UCAR's goals for advancing science, technology, and education. Computer networks are a vital enabling technology that allows SCD to reach and serve a wide community.

Network engineering seeks to develop and implement:

• Networking technologies characterized by a mixture of sufficient bandwidth, reliability, integrity, and scalability
• A network architecture that is cost-effective, maintainable, adaptable, and accommodates future growth
• Strategies that support SCD high-performance networking
• Leadership and participation in high-performance network research and development

The unifying goal of all SCD research is enabling scientific investigators to gain maximum benefit from the available computational, data storage, and data visualization resources. To this end, SCD will pursue this research agenda:

• Research and develop databases and intelligent search engines to catalog and locate information in mass storage systems
• Research new ways to manage data and enable investigators to more effectively visualize and explore scientific datasets through novel techniques
• Research networking techniques that enable ordinary network users to attain full network data rates without the assistance of experts
• Research numerical methods related to simulating fluid dynamics on the sphere, particularly focusing on methods that synergize well with current trends in computer architecture
• Investigate a common modeling infrastructure (framework) for earth systems simulations to: Accelerate scientific progress by reducing duplication of effort; Enable collaboration to permit a sensible division of labor; Improve model reliability, interoperability, flexibility, and performance portability

SCD upholds a long tradition of collecting, quality-controlling, and documenting observational and analyzed datasets that are used to enhance and support the research activities of the atmospheric and related sciences communities. Scientists worldwide have come to rely on SCD to provide "clean," up-to-date, and accurate data as they pursue their own research agendas. To ensure that our community continues to have access to reliable data, SCD will:

• Update current research data archives with scientifically valuable observations and datasets from other regional, national, and international centers
• Improve ease of access to the SCD research dataset holdings
• Augment current data services for our scientific community via enhanced use of web-based technologies to acquire and distribute these data
• Assemble and supply important metadata covering major data holdings in SCD

SCD will actively leverage its role in the community to support activities, programs, and collaborative research projects to enhance education at all levels. As part of this initiative, SCD will:

• Continue our work with the SOARS program, Project LEARN, and science and technology presentations to the general public
• Build on the Virtual Exploratorium Project and seek follow-ons or related projects
• Develop creative options for continuing and enhancing the Colorado Computational Science Fair (CCSF)
• Pursue an NSF/ISE-funded museum exhibit on earth system science in collaboration with several museums
• Enhance university collaborations in computational science research and education by building a visitor program for graduate students, postdocs, and faculty. Explore funding mechanisms.
• Develop education programs for science educators at the secondary and university levels through collaborations with those educators. They would be the target audience for the education programs, and they would actively participate in developing proposals for funding the programs.
• Collaborate with the atmospheric and related geosciences community to develop course content that addresses the need for scientists who are better trained in the efficient use of new computing architectures.

Scientific investigators want to find, interact with, and contribute productively to NCAR’s collection of resources: its people, services, computers, data, and documentation. SCD’s enterprise web services will provide the middleware for an integrated, easy-to-use collaboratory. The complexity of the UCAR/SCD computing environment has increased, and navigating its resources has become nontrivial.

To reduce the researcher’s workload, SCD is creating web-based interfaces that will:

• Interact with all UCAR/SCD resources, providing a common infrastructure to make each researcher’s experience as straightforward and seamless as possible
• Enable our community of researchers to productively apply the vast collection of UCAR and SCD resources to their scientific investigations

As part of SCD's thrust to be outwardly oriented and centered on the needs of researchers, we will support large and small investigative research teams, no matter where they are located. Our services include targeted assistance in the form of consulting services, code conversion, code optimization, support, training, outreach, education, and information services. New services will emphasize providing "just-in-time" information, especially in the areas of code migration to new computing platforms and in making the SCD computing environment easier to use.

SCD will:

• Identify and track major codes used by the university and NCAR communities that have been or should be migrated to symmetric multiprocessors (SMPs)
• Provide regional training to universities using "needs analysis" to optimize the limited time university scientists and graduate students have to spend on learning new computing paradigms
• Develop FAQs (Frequently Asked Questions) from a variety of sources including consulting contacts, training materials, and issues identified by researchers
• Couple these FAQs with a powerful search engine to allow researchers to access only the information they need, just when they need it