Feature - People behind the grid: Simon Lin
(Editor's note: Simon C. Lin is currently coordinator of the Asia Federation in EGEE and is responsible for the ASGC-Academia Sinica Grid Computing Center-the only WLCG Tier-1 Center in Asia. He is deputy project manager of the EUAsiaGrid project, coordinated by ASGC and INFN in Italy.)
iSGTW: How did you get involved in particle physics?
I am a theoretical physicist by training. Like many of my generation, I was inspired by the great names in physics of the first half of the 20th century. I was fortunate enough to do my Ph.D. at Edinburgh University under Peter Higgs. Firstly, I worked on applying group theory to supersymmetry but then turned to field theory on surfaces and interfaces later. I still feel theoretical particle physics is excellent training for physicists, independent of what they go on to do.
iSGTW: How did you get involved in computing?
In Edinburgh, I had Dave Wallace as a co-supervisor. Apart from his important work on renormalization group theory, he was also a pioneer in scientific computing. For example, he introduced massively parallel processing in Edinburgh at a very early stage. Later, he helped people like Tony Hey start the e-Science program in the U.K.
Dave's enthusiasm for scientific computing was a big influence on me. So when I went back to Taiwan I pursued a keen interest in computing for physics. As a result, I became director of the computing center at Academia Sinica for 15 years.
iSGTW: How did Academia Sinica become the Asian Tier-1?
By about 2002, many different flavors of grid middleware were appearing in Asia. These were basically regional variations of Globus. However, I was attracted by the more pragmatic approach being taken for LHC computing, where the idea was to take existing middleware from different sources, scrutinize it, fix potential weaknesses and deploy the result. The fact that LHC physicists would rely on this middleware for at least a decade convinced me that it had a good chance of survival.
As a result, Taiwan became one of the first Asian countries to participate in the LCG Grid Deployment Board meetings. At first, I didn't think that the enormous resources for a Tier-1 would be feasible in Taiwan. Also, we didn't have real broadband connections between Europe and Asia at that time. On the other hand, it became clear to me that if LCG were going to be a 24/7 service, being in Asia was an advantage when it came to supporting that service, because of the time zone difference.
While running regional grid operations and user support may not sound like a glamorous task, I saw this as an opportunity for our staff at Academia Sinica to become experts in this new technology and also contribute to it. So, with the help of a grant from the National Science Council, we moved quickly to build up our expertise, and sent a number of young researchers to CERN for training. Thanks to this early start, and with the support of our Japanese colleagues, Academia Sinica became the Asian Tier-1. Indeed, we were the first of all Tier-1s to sign a Memorandum of Understanding with CERN, in 2005.
A view from the East
iSGTW: What are the main challenges of running the only Tier-1 in Asia?
A big technical challenge has been getting bandwidth at a reasonable price. When we started to get involved in LCG, there were no 10-gigabit lines available between Taipei and CERN. Even today, trans-Pacific links are twice the cost of trans-Atlantic ones, and going via the Indian Ocean is eight times more expensive. Fortunately, I had been involved in the '90s in promoting major network upgrades for Taiwan, with the support of the president of Academia Sinica, Nobel laureate Y. T. Lee. Because of this background, I have been able to negotiate good deals for setting up the necessary connectivity for the Tier-1.
Another particular issue in Asia is that we are traditionally playing technological catch-up with the West. So we tend to collaborate with the United States or Europe, rather than with other Asian countries, which we may see as regional competitors. I think that LCG and the Tier structure has provided us with a strong motivation for better regional collaboration.
iSGTW: Looking beyond LCG, what do you see as the opportunities for grid computing?
I think grid software has great room to improve. We have not learned enough from the success of TCP/IP. Some sort of simple, universal protocols need to evolve, to overcome the current challenge of interoperability between different grids.
At Academia Sinica, we are involved in several other grid-based activities, including the WISDOM virtual drug-docking challenges we have collaborated on for malaria and avian flu, which showed how quickly and effectively a grid infrastructure could respond to real research needs. Grid technology for preservation and long-term digital archiving as well as disaster mitigation, are areas where I see a lot of potential, if we can convince different regional institutions to share data.
More profoundly, I've always felt that the development of computers would have a long-term impact on science. Inspired by visionaries like Richard Feynman, scientists have long speculated on what computing might be like if it were based on new paradigms, like quantum computing. In the same way, I think that grid computing is a conceptual change that may have a huge long-term impact.
-Francois Grey for iSGTW