This document consists of notes on the general integration of EOT-PACI activities and the alliance roadmaps. As a matter of principle, EOT-PACI represents the human infrastructure elements of the PACI activity and therefore our interaction with the roadmaps concerns areas of education, access and inclusion of individuals and communities.

Underrepresented Groups: EOT-PACI can directly address capability computing by assisting under-represented researchers (such as at Minority Serving Institutions) to become capability users of PACI resources. This effort will involve identifying a few researchers in FY’00 that are prime candidates for becoming capability users, and making connections between them and the appropriate PACI teams (AT, ET, and PACS). It will also require extending the grid to their facilities, and providing portals (see portals section below) to meet their research needs. (see grid access element below).

Government: EOT-PACI can directly address capability computing by identifying government agency and private sector researchers that are prime candidates to become capability users by extending the grid to them, tailoring portals to their research needs, and by offering training and support to the researchers.

Education/Pipeline: EOT-PACI will address capability computing pipeline issue by preparing the next generation of capability users, including under-represented populations. This will include working with undergraduate students through REU experiences to immerse them in projects involved in capability computing. It will involve working with faculty to develop computational science courses that teach students to effectively utilize PACI technologies and resources. It will also require working with pre-service and in-service teachers to prepare them to make computational science an integral part of the K-12 curriculum.

The applications of capability computing (especially Breakthrough Simulations and High Throughput simulations) represent the apex of what HPC is can achieve.

Computational Science Models & Educational Content: EOT-PACI can work with selected CC applications to help incorporate key ideas, methods, and results in an educational context.

Desktop to XX-flop: EOT-PACI will look for opportunities to encourage students and researchers to think beyond their desktops with models that involve scalable simulations. We will work to link educational groups to high throughput CC at LES and PACS where feasible. This models a key element of CC -- that one brings a wide range of computing capabilities to bear on real problems.

EOT assessment of CC Impact: EOT-PACI could help assess the broader impact of breakthrough CC applications. Such assessment should be an element of the decision of which "hero" simulations should receive priority access to CC resources.

EOT-PACI will identify a few key testbeds for extending the grid. These will occur in K-12 and undergraduate education settings, in government agencies, and in private sector locations. The unique needs of these organizations need to be factored into the development of the grid – such testbeds will enable the ET teams to understand the diverse needs and requirements for a national computational grid. While the majority of testbeds will be national, there is a need to develop an international testbed using STAR TAP among US and European collaborators.

The EOT-PACI audiences will bring unique requirements for collaboration, processor scheduling, quality of service, visualization facilities and tools, security, and data management and handling which will benefit the grid development process. For example, government agencies bring a complexity of scaling up of problems including not tens, but rather thousands of databases. The ability of the grid and portal efforts to support the real-world problems facing government agencies and the private sector are critical to the adoption and usage of the grid technologies on a national scale. The EOT-PACI testbeds will immerse these unique needs into the development process.

The ACCESS grid is an essential element in this effort. Through facilities such as the ACCESS Centers, EOT-PACI will be able to reach a much wider audience to ensure that their needs are taken into consideration in the development of the grid. The ACCESS grid will also enable the EOT-PACI and PACS partners to disseminate the technologies and resources of PACI to more people through demonstrations, training and collaboration capabilities offered by the ACCESS Centers. The ACCESS Centers can be used by EOT-PACI to ensure the training of trainers within diverse communities that will advance and perpetuate the grid technologies.

In those institutions where it is practical, this effort will include creating local computational grids as stepping stones to the national computational resources within PACI. This may often mean utilizing cluster architectures to provide local computing power.

EOT-PACI also seeks to prepare the next generation of people that will continue to evolve and improve the grid, as well as to provide students who can now both engage in the development efforts to improve their skills, and accelerate the development efforts.

EOT-PACI has far-reaching, long-term needs and requirements for the development of the Common Portal Architecture to support a diverse audience – education, government, and business. The impact of fulfilling these needs will be broad adoption of the technology by a diverse set of audiences.

The architecture of distributed object component technologies complements development work underway in NSF funded education research projects that would benefit greatly from collaboration with the portals development effort. Government agencies are in need of easy to use interfaces and meta-data standards for complex data management and mining. NSF is looking for solutions for K-12 digital libraries which the portals and grid teams can uniquely provide to the country. Participation of EOT-PACI in the development of these portals solidifies opportunities to leverage efforts across the NSF directorates.

Problem solving, web-based workbenches are of direct need to all audiences being served by EOT-PACI. The diverse needs of the audiences that EOT-PACI brings into the process will ensure that the resulting workbenches can better serve a broader audience than what might otherwise result from the scientific drivers of the efforts. For example, educators now working with the Biology Workbench and Chemical Engineering Workbench teams are introducing needs and requirements that were not identified by the scientists, but which when added to the workbenches, will make them useful to broad educational applications. Workbenches and portals need to be intelligent about the users – to track where they’ve been, to support them based on their skill level, and to facilitate rather than impede progress. They can only do this by being cognizant of the individual users needs and abilities.

A workbench should allow a student (K-12, undergraduate, or workforce member) to learn to use the workbench assuming little knowledge of the tool, and yet allow the student to solve as complex a problem as a scientist may post to the workbench when they are ready to do so. EOT-PACI, through testbeds and partnerships with diverse audiences, intends to help facilitate such advances in workbench design. The workbenches will in turn allow educators to better prepare the next generation of scientists and engineers to utilize the infrastructure to advance their fields of study.

An area of particular interest and need to EOT-PACI and all of the teams is distributed learning and working. The ability to support remote collaboration and learning is critical not only to the PACI teams, but to all sectors of society. Today’s tools fall far short on the issues of usability, visual quality, quality of service, reliability and functionality. EOT-PACI will work with the grid and portals teams to address these issues and to significantly move the infrastructure ahead. We need major advances in this area, not just minor tweaks to existing tools. It’s an area where PACI can take national leadership.

The UD/DA partners of EOT-PACI will help to ensure that the technology designs are well suited for use by people with disabilities to ensure that the resulting tools are in fact usable and accessible.

EOT-PACI will provide an evaluation of the impact and benefits of PACI technology to the diverse audiences. This will enable PACI to measure the effectiveness of the resulting tools in meeting diverse national applications.

EOT-PACI will bring diverse needs and requirements, diverse testbeds, and energetic participants to advance the development and deployment of the national technology grid. EOT-PACI will also help to prepare the next generation of scientists and engineers to both help advance the infrastructure and well as to make effective use of the infrastructure in their own research and development efforts.