05.06.2013 Views

PNNL-13501 - Pacific Northwest National Laboratory

PNNL-13501 - Pacific Northwest National Laboratory

PNNL-13501 - Pacific Northwest National Laboratory

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Study Control Number: PN99008/1336<br />

Collaborative Problem-Solving Environments<br />

Deborah K. Gracio, Karen L. Schuchardt, Brett T. Didier<br />

Advanced computational techniques are critical to improving our understanding of scientific processes in a<br />

multidisciplinary research environment. The development of collaborative problem-solving environments, which apply<br />

current research and state-of-the-art technology to create a powerful and convenient computing environment, is one key to<br />

transcending current methods of solving complex problems. General-purpose components that enable the construction of<br />

domain-specific collaborative problem-solving environments are being developed on this project.<br />

Project Description<br />

The purpose of this project is to develop general-purpose<br />

infrastructure components that are required to develop<br />

and deploy collaborative problem-solving environments.<br />

Our requirements were derived from work in the areas of<br />

climate change, chemistry, manufacturing engineering,<br />

and subsurface reactive transport. We developed<br />

prototypes of several components this year: resource<br />

information management and discovery, portal services,<br />

and distributed data and meta-data management. In<br />

addition, we developed a conceptual prototype of a<br />

climate change problem-solving environment with an<br />

emphasis on flexible workflow. This project produced<br />

technology that is now being applied to programmatic<br />

work and advanced our ability to develop domain-specific<br />

collaborative problem-solving environments.<br />

Introduction<br />

A collaborative problem-solving environment is a<br />

computational system that provides a complete and<br />

convenient environment for solving problems in a specific<br />

domain. The target problems are increasingly distributed<br />

in nature. Collaborators, often physically remote and<br />

working for different organizations with different<br />

infrastructures, require access to distributed resources<br />

such as supercomputers, instruments, and research data.<br />

A collaborative problem-solving environment must<br />

support this complex work environment in addition to<br />

supplying the domain-specific tools and capabilities.<br />

Constructing such a system is cost prohibitive unless<br />

components that address the basic infrastructure issues are<br />

readily available. The specific aim of this project is to<br />

apply state-of-the-art technologies to create new solutions<br />

for selected infrastructure components, thus enhancing<br />

our ability to develop and deploy collaborative problemsolving<br />

environments.<br />

106 FY 2000 <strong>Laboratory</strong> Directed Research and Development Annual Report<br />

Another important characteristic of problem-solving<br />

environments is that they must be readily adaptable to<br />

new problems and new solution strategies. The users<br />

require the capability to define and redefine their process<br />

or workflow. We continued our research in this area by<br />

developing a conceptual prototype of a flexible workflow<br />

environment for atmospheric research.<br />

Results and Accomplishments<br />

Two years ago, a collaborative problem-solving<br />

environment was only an idea. Today, it has been<br />

embraced as one of the key capabilities of the <strong>Laboratory</strong><br />

in Information Sciences. The research and technologies<br />

being developed by the collaborative problem-solving<br />

environments team are sought out by many of the<br />

research entities in the <strong>Laboratory</strong> to provide solutions to<br />

aid scientists, decision makers, and policy planners in<br />

their work.<br />

Portal Services<br />

We designed and developed a web-based component for<br />

submission and monitoring of computational jobs. The<br />

design was implemented as a portal service that can be<br />

accessed by applications written in Java, C++, or any<br />

browser supported language. We tested this service by<br />

successfully running computational jobs under a variety<br />

of computing environments including UNIX and Linux<br />

workstations, Linux clusters running the Portable Batch<br />

System, an IBM SP running Maui Scheduler and Load<br />

Leveler, and an Silicon Graphics, Inc. Origin 2000<br />

running the Network Queuing System. The service has<br />

the following features:<br />

• accepts XML job requests over a secure web-based<br />

connection (https)

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!