"What are the most important problems in your field? Are you working on one of them? Why not?"
It is better to solve the right problem the wrong way than to solve the wrong problem the right way" - Richard Hamming

sangmin_2009.jpg

Sang-Min Park 

Ph.D. Candidate
Computer Science, University of Virginia

Contacts
Phone: 434-825-4939
Email : sp2kn(no_spam)@virginia.edu
Office: 233 Olsson Hall

Research Group (from 2004)
UVa eScience Group (Advisor: Prof. Marty Humphrey)

Education
Master of Computer Science (2006)   :   Computer Science, University of Virginia
M.E. (2004)    :  Information and Communication Engineering, Ajou University, South Korea    (DMC Lab)
B.S. (2002)   :   Computer Science, Ajou University, South Korea

Job materials -
I expect to finish my PhD in May 2010. Here're links to my CV and research statement.
If you have an opening, please let me know.

News!
Nov 08 - International Science Grid This Week (iSGTW) featured an article introducing my work: Cruise control for eScience



Research Interests

Cloud Computing, HPC and Distributed Computing for E-Science

- Cloud Computing (Virtualized Datacenter)
- Predictable High Performance Computing
- Data-parallel programming such as MapReduce and DryadLINQ
- eScience Workflow Systems
- Scheduling Distributed Computation/Communication
- Authorization and Policy Issues
- Control Theory Applications in Computing System
- Service Oriented Architecture (WS-* / REST)

 

Summary of my PhD Research (History repeats, time-sharing again!)

My PhD research focuses on building predictable high performance computing infrastructure, which is one of the major requirements for emerging dynamic data driven applications. In this new class of applications, complex computations should be finished within explicit deadlines. The examples include LEAD for real-time weather prediction, earthquake analysis, and medical modeling for patient-specific treatments. The current HPC infrastructures, which are largely based on a batch-mode job processing, cannot provide the sufficient level of predictability and adaptability. My research takes a fundamentally different approach to solve unpredictability; rather than relying on the reservation semantic, my research exploits two capabilities -- performance isolation and online resource reconfiguration -- that are made possible with virtualization so as to build a predictable time-sharing system. In my work control theory provides the theoretical underpinning of predictability so that the implemented system can guarantee applications meet deadlines even when severe disturbances affect their progresses.

 

Publications

·         Sang-Min Park and Marty Humphrey. Predictable High Performance Computing using Feedback Control and Admission Control. In submission to a journal (This is an extended version of SC'08 paper; send me an email if you want a copy) (abstract)

·         Sang-Min Park and Marty Humphrey. Self-Tuning Virtual Machines for Predictable eScience. Proceedings of IEEE/ACM International Symposium on Cluster Computing and the Grid (CCGRID'09), May 18-21, 2009, Shanghai, China. slides (acceptance rate: 21.0%)

·         Sang-Min Park and Marty Humphrey. Feedback-Controlled Resource Sharing for Predictable eScience. IEEE/ACM International Conference for High Performance Computing, Networking, Storage and Analysis (SC08), Nov 15-21, 2008, Austin, Texas. slides (Best student paper finalist) (acceptance rate: 21.3%)

·         Sang-Min Park and Marty Humphrey. Data Throttling for Data-Intensive Workflows. 22nd IEEE International Parallel and Distributed Processing Symposium (IPDPS 2008). April 14-18, 2008. Miami, FL, USA. (acceptance rate: 25.6%)

·         M. Humphrey, S.-M. Park, J. Feng, N. Beekwilder, G. Wasson, J. Hogg, B. LaMacchia, and B. Dillaway. Fine-Grained Access Control for GridFTP using SecPAL. 8th IEEE/ACM International Conference on Grid Computing (Grid 2007), Austin, TX, Sept 19-21, 2007.

·         Sang-Min Park, Glenn Wasson, and Marty Humphrey. Authorizing Remote Job Execution based on Job Properties. 2nd IEEE International Conference on e-Science and Grid Computing (e-Science 2006). Dec 4-6, 2006, Amsterdam, Netherlands.

·         M. Humphrey, G. Wasson, Y. Kiryakov, S-M. Park, D. Del Vecchio, N. Beekwilder, and J. Gray. Alternative Software Stacks for OGSA-based Grids. Proceedings of Supercomputing 2005, Seattle, WA, Nov 12-18, 2005.

·         J.V.S. Watson, Sang-Min Park, and M. Humphrey. Toward GT3 and OGSI.NET Interoperability: GRAM Support on OGSI.NET. 2005 International Conference on Computational Science (ICCS 2005). May 22-25, 2005. Emory University, Atlanta, GA, USA.

·         Sang-Min Park, Jai-Hoon Kim, Young-Bae Ko, and Won-Sik Yoon. Dynamic Data Grid Replication Strategy based on Internet Hierarchy. Second International Workshop on Grid and Cooperative Computing(GCC'2003) in Shanghai, China, Dec 2003.

·         Sang-Min Park, Young-Bae Ko, and Jai-Hoon Kim. Disconnected Operation Service in Mobile Grid Computing. First International Conference on Service Oriented Computing(ICSOC'2003) in Trento, Italy, Dec 2003. (acceptance rate:20.9%)

·         Sang-Min Park and Jai-Hoon Kim. Chameleon: A Resource Scheduler in a Data Grid Environment.   2003 IEEE/ACM International Symposium on Cluster Computing and the Grid(CCGRID'2003), Tokyo, Japan, May 2003. (acceptance rate: 34.2%)

·         Sang-Min Park and Jai-Hoon Kim. A Communication Cost Model for Dynamic Selection of Data Replicas in Grid Environment. Korea Information and Communication Socienty Summer Workshop, 2002.07.

Master Thesis

       Sang-Min Park. A Research on Job Scheduling and Data Replication Mechanism in Grid Computing Environments - Ajou University, South Korea, 2004.

 

Experiences

Talks
TA
RA

 

Wisdom

·         You and Your Research by Richard Hamming

This article is a transcript from Dr Hamming’s 1986 talk at Bell Communication Lab. It contains very inspiring
collection of wisdoms that every scientist/engineers would gain significant insights. I made summary of it here:

a.       Work on important problem - "Do first class work"

b.       Emotionally commit to the problem - "Love YOUR problem"

c.      Turn your work into useful one to others

d.       Change your viewpoints; defect will become asset

·         Last Lecture by Randy Pausch
"Brick walls are there for a reason. They let us prove how badly we want things."

·         The Computer Scientists as Toolsmith by Fred Brooks
“The computer scientist is a toolsmith: no more, but no less. It is an honorable calling."
“Arbitrary complexity is our lot, and here more than anywhere else we need the best minds of our discipline
fashioning more powerful attacks on such problems"

·         On Ph.D. Thesis Proposals by H. C. Lauer
This report describes the useful guidance in preparing a Ph.D. thesis proposal in Computer Science.
The process might be well applied to other science/engineering disciplines as well.

 

Research Links

Grid computing
International Conferences (where my paper will go!)

 

Other Life

Light and Salt Church (샬롯츠빌 빛과소금교회): I am serving as a deacon, worship leader, and bible school teacher

 This page is updated in Oct12 2009