The University of Arkansas (UA), Fayetteville, the University of Arkansas at Little Rock (UALR), and the University of Arkansas for Medical Sciences (UAMS) are partnering with state industrial affiliates currently including representatives from Acxiom and Accelerate Arkansas, and state government representatives currently including the Arkansas Science and Technology Authority and the Arkansas Department of Information Systems, to leverage the Arkansas Research and Education Optical Network (ARE-ON) in a vision and plan to execute that vision for the state of Arkansas in high-performance computing for research, education, and business infrastructure.  We believe that the availability of high-performance computing infrastructure will be essential to the economic development of any state in the 21^st Century.  High-performance and advanced computing capabilities and technology for the understanding and solution of complex problems in science, engineering, and industry are critical to scientific leadership and economic competitiveness in the state of Arkansas.  This is in keeping with the findings of the report from the President's Information Technology Advisory Committee (PITAC) (http://www.nitrd.gov/pitac/reports/20050609_computational/computational.pdf).

 

The debut of the Arkansas Research and Education Optical Network (ARE-ON) is a clear indicator that the state of Arkansas is taking a fresh and energetic approach to high performance computing for educational benefit and economic development.  ARE-ON came online to the University of Arkansas, Fayetteville, in December 2006, and has already been instrumental in UAF participation in a collaborative course with Louisiana State University (LSU) this past spring.  This is one example of one type of educational activity that ARE-ON will support.  ARE-ON represents a statewide initiative that already puts Arkansas ahead of some other states. 

 

Just as ARE-ON is a statewide effort for connectivity, the Arkansas Partnership for Advanced Computing recognizes that there needs to be a complementary statewide effort to support computational infrastructure.  Three additional indicators show that the timing and support are right for such an effort:

 

1) This growing partnership between UA, Fayetteville, UALR, and UAMS and several statewide industrial and government partners provide a solid foundation for a state vision for high performance computing.

2) Legislative support of the Arkansas Science and Technology Authority (ASTA) may provide funding potential that is an opportunity to gain seed funding for an initiative.

3) Funding from the National Science Foundation (NSF) provided for the purchase and deployment of the first supercomputer in Arkansas, Red Diamond, on the UA, Fayetteville, campus, in February, 2005.  Additional funding from the National Science Foundation and EPSCOR in 2006 to UALR and from the NSF in 2007 to UA, Fayetteville, is enabling the significant expansion of high-performance computing infrastructure.  With coordination between the researchers on both campuses it will be possible to double the size of Red Diamond at UA, Fayetteville, and to establish a complementary cluster at UALR.  ARE-ON will provide the link to connect the supercomputing facilities as a nationally-competitive high-performance computing grid that is accessible to researchers across the state.

 

Fundamentally, this initiative for high-performance and advanced computing is about quality, quantity, and the diversity of an emerging workforce.  The workforce includes:

 

 

To achieve these goals requires a statewide commitment to modernization and technology—a move to the 21^st century.  With these indicators and goals, a plan has been made that is economically sound that will move rapidly toward the goals, with modest risk.  The plan will reach the goals with minimum cost, with the maximum likelihood of success, and will mix external expertise with internal experts and leaders in the state.  This is a three-pronged attack:

 

1) With support from UAF, UALR, ASTA, and the National Science Foundation, we have formulated a high performance computing External Advisory Committee (EAC) to look at the requirements and needs of the state.  This external experience base will make recommendations, and provide guidelines and milestones. Dr. Thomas Sterling, Professor, LSU has provided some initial guidance on our current status and has recommended that this is the fastest way to get the high-quality insight necessary to leap-frog our current position. Dr. Dan Reed, director of the Renaissance Computing Institute in North Carolina, Chancellor’s Eminent Professor and member of the PITAC committee, has agreed to be the Chair of the EAC.

The external advisory committee will visit Arkansas over a three day period in October.  They will conduct a series of brief interviews with stakeholders in the state, spending a day each in Little Rock and Fayetteville.  The deliverable of the EAC is a strategic plan that describes the scope and a roadmap for developing high performance computing infrastructure in the state of Arkansas.

2) We have implemented a standing Internal Review Committee composed of experts within the state of Arkansas. This committee consists of approximately two dozen participants from the state of Arkansas, and an additional one, two, or three external participants.  The Internal Review committee will refine the statement of goals that the External committee has developed.  This committee will be an interface to the academic community, K-12, and industry.

3) We will be in partnership with the state legislature, the Governor, and key leaders across the state to develop a sustainable funding model.

 

Industrial partners from Acxiom and Accelerate Arkansas have been participating in this discussion for over a year.  One thing that will help to drive this effort is the identification of one or more “Killer Applications” (ones that grab the attention of funding agencies) that ARE-ON and the computational infrastructure can facilitate, and these may originate from industry, agriculture, or academics.  For example, in Louisiana, “Killer Applications” include:

 

1) modeling of storm surge to avoid damage and save lives during hurricanes and other storms,

2) modeling of depleted oil wells and seismology studies that can help to avoid wild cat digging that wastes millions of dollars and harms the environment,

3) modeling of the preservation and ecological changes to wetlands, and

4) education as a first-class application, to improve the competitiveness of Louisiana as a state.

 

High performance computing must be a synergy of education, industry, and research and is a requirement for ensuring that all Arkansans can fully participate in the digital world.