Korea plans to be in the top seven nations of supercomputing by 2017, and Dr. Jysoo Lee, as director general of the recently established National Institute of Supercomputing and Networking (NISN) within KISTI (the Korea Institute of Science and Technology Information), is one of the key figures advancing this goal. Dr. Lee also serves as chief professor for the Grid and Supercomputing Program of Korea’s University of Science and Technology. In this Q&A, he evaluates progress on the effort he drove to establish and implement Korea’s first national supercomputing plan.
HPCwire: Talk about HPC in Korea before passage of the National Supercomputing Act. Why was that legislation needed?
Lee: In 1988 Korea purchased its first supercomputer, which wasn’t far behind systems in leading countries of the world. If you compare the current level of HPC adoption in Korea with that of developed countries, however, there is a clear gap that has been widening. For example, there was no system from Korea which made the Top500 list in 2009. The incident served as a wake-up call. We all know that HPC is an important tool for enhancing the competitiveness of a country, and it was clear that Korea was not doing a good job. It’s not so important how many supercomputers are in Korea, per se, but it is important how well Korea uses them for the nation’s benefit.
HPCwire: What process did you follow to gather support for the legislation?
Lee: It is very important to have a balanced ecosystem, which includes applications, infrastructure, and human resource, for maximum effect. For example, it would not be wise to have a large number of supercomputers, but a small number of experts to use them. In that respect, as the leader of this effort I worked with key communities and tried to represent their voices. For example, opinions on infrastructure were heavily gathered from the Korea Supercomputing Centers Alliance, and the Korean Society for Computational Sciences and Engineering played a similar role for applications. I think that people in the national assembly of Korea recognized the voices from these communities.
HPCwire: When was the National Supercomputing Act passed?
Lee: The act was first proposed in regular session of the national assembly of Korea in September 2009. It then went through several committees, sub-committees, and hearings, and was approved in the general session in April 2011. It was enacted in June 2011, and the enforcement ordinance and regulations were established by the end of the year.
HPCwire: Can you provide a summary of the act?
Lee: Its goal is to enhance people’s quality of life and the economic development of the nation through the efficient use of supercomputing. The key action for achieving the goal was set out as establishing and implementing a national plan for the supercomputing ecosystem of Korea. It is the national supercomputing committee who makes decisions on key issues. The chairman of the committee is the Minister of Science, ICT, and Future Planning, and high-ranking officials from nine other ministries such as the Ministry of National Defense and the Ministry of Trade, Industry and Energy are members. Also included are several HPC experts from academia, private and public sectors.
HPCwire: You said the act required the establishment and execution of a national supercomputing plan for Korea. Can you give a summary of the plan?
Lee: It took the whole year of 2012 to establish the first five-year master plan for 2013-2017. Its goal is for Korea to be in the top seven nations of supercomputing by 2017. The plan focuses on three key areas: applications, infrastructure, and technology. These are further organized into ten programs. Major emphasis is given to building a balanced ecosystem that addresses all the key components in the whole workflow, from system to solution. Also, a year-by-year implementation plan was established and this was carried out starting in 2013.
HPCwire: One consequence of the act was the creation of NISN. Why was this important and how did it happen?
Lee: Throughout the whole process of implementing the national plan, it was clear that the national supercomputing committee would need a lot of help, and that this is exactly the role a national supercomputing center should play. In fact, its mission and functions are clearly spelled out in the act. The government designated KISTI, the Korea Institute of Science and Technology Information, to assume the role of the national center and asked us to propose a plan. KISTI responded by establishing NISN, the National Institute of Supercomputing and Networking. NISN is based at the KISTI supercomputing center, which has been serving the country for 25 years or so. NISN officially began operating in 2013, and I am honored to be its founding director-general.
HPCwire: Are there any new projects that have begun recently?
Lee: There are a few projects that started in the last couple of years, but they are relatively small in size. In order to start major projects from the programs in the master plan, additional efforts like detailed planning are necessary, and it usually takes a couple of years for projects like that to start. Currently, detailed plans for three major projects have been completed. First is the National Supercomputing Infrastructure Initiative, which is analogous to XSEDE or PRACE. Second is the National Supercomputing Education and Training Framework. Last is “SuperKorea 2020,” a plan which includes procuring a national leadership computing system, similar to the Tier-0 systems of PRACE, and developing an indigenous Korean supercomputer. Also in progress is the construction of NISN’s new data center, which will be completed by the summer of next year.
HPCwire: Often around the world, legislation is passed but the funding has to be pursued separately. What is the situation in Korea?
Lee: That is also the case in Korea. It certainly is easier to start a project when the project is included in the government’s master plan, but it does not guarantee funding. A budget table was included in the master plan, but merely used as a reference. As I have said before, additional efforts are required to fund large-scale projects like ours.
HPCwire: What are the risks to Korea if the act is not adequately funded?
Lee: The master plan’s goal for Korea to become a top seven supercomputing nation is quite ambitious, and there is not much room for error to achieve it by 2017. It certainly is very difficult, if not impossible, to realize the goal without timely funding. Although the Korean government and our supercomputing communities in Korea are trying hard, progress has been relatively slow. I still believe the goal can be achieved, but it is clear we don’t have too much time on our hands.