Suntae Hwang

Deploying Scientific Applications to the PRAGMA Grid Testbed: Strategies and Lessons

Recent advances in grid infrastructure and middleware development have enabled various types of applications in science and engineering to be deployed on the grid. The characteristics of these applications and the diverse infrastructure and middleware solutions developed, utilized or adapted by PRAGMA member institutes are summarized. The applications include those for climate modeling, computational chemistry, bioinformatics and computational genomics, remote control of instruments, and distributed databases. Many of the applications are deployed to the PRAGMA grid testbed in routine basis experiments. Strategies for deploying applications without modifications, and those taking advantage of new programming models on the grid are explored and valuable lessons learned are reported. Comprehensive end to end solutions from PRAGMA member institutes that provide important grid middleware components and generalized models of integrating applications and instruments on the grid are also described.

An XML Schema-based Semantic Data Integration

Cyber-infrastructures for scientific and engineering applications require integrating heterogeneous legacy data in different formats and from various domains. Such data integration raises challenging issues: (1) Support for multiple independently-managed schemas, (2) Ease of schema evolution, and (3) Simple schema mappings. In order to address these issues, we propose a novel approach to semantic integration of scientific data which uses XML schemas and RDF-based schema mappings. In this approach, XML schema allows scientists to manage data models intuitively and to use commodity XML DBMS tools. A simple RDF-based ontological representation scheme is used for only structural relations among independently-managed XML schemas from different institutes or domains. We present the design and implementation of a prototype system developed for the national cyber-environments for civil engineering research activities in Korea (similar to the NEES project in USA) which is called KOCEDgrid

X-SIGMA: XML Based Simple Data Integration System for Gathering, Managing, and Accessing Scientific Experimental Data in Grid Environments

Effective scientific data management is crucial for e-Science applications. Scientific data management raises challenging requirements: (1) support for not only experimental data but also context data, (2) both schema evolution and integration, (3) and compatibility with legacy data management conventions or environments. In this paper, we present a scientific data management system (called XSIGMA) which is designed to address those issues. X-SIGMA is a Grid-based integrated system for managing, integrating, and accessing scientific experimental data. A prototype system is implemented and has been being used to develop the scientific data management system for the national cyberinfrastructure project called KOCED in Korea.

A workflow management and grid computing approach to molecular simulation-based bio/nano experiments

In this paper, we propose an approach to molecular simulation-based experiments which combines workflow management and grid computing techniques to address both the computing issue due to the challenging computation requirement from molecular simulation and the management issue due to distributed, heterogeneous computing platforms. We present a workflow management system customized for such experiments and explain how this workflow system can be integrated with computational grids.

Modeling Occurrence of Urban Mosquitos Based on Land Use Types and Meteorological Factors in Korea.

Mosquitoes are a public health concern because they are vectors of pathogen, which cause human-related diseases. It is well known that the occurrence of mosquitoes is highly influenced by meteorological conditions (e.g., temperature and precipitation) and land use, but there are insufficient studies quantifying their impacts. Therefore, three analytical methods were applied to determine the relationships between urban mosquito occurrence, land use type, and meteorological factors: cluster analysis based on land use types; principal component analysis (PCA) based on mosquito occurrence; and three prediction models, support vector machine (SVM), classification and regression tree (CART), and random forest (RF). We used mosquito data collected at 12 sites from 2011 to 2012. Mosquito abundance was highest from August to September in both years. The monitoring sites were differentiated into three clusters based on differences in land use type such as culture and sport areas, inland water, artificial grasslands, and traffic areas. These clusters were well reflected in PCA ordinations, indicating that mosquito occurrence was highly influenced by land use types. Lastly, the RF represented the highest predictive power for mosquito occurrence and temperature-related factors were the most influential. Our study will contribute to effective control and management of mosquito occurrences.

Glyco-MGrid: A Collaborative Molecular Simulation Grid for e-Glycomics

In this paper, we present a novel approach to glycomics: collaborative molecular simulation. Glyco-MGrid is a collaborative grid environment for glycobiology which supports simulation, databases and trajectory analyses in an integrated way. It allows the global glycomics community to share simulation results and to continue further research from previous results by other individual scientists or research groups. Currently, over five hundreds of simulation results for glycans and glycoproteins are available from the Glyco-MGrid system (http://glyco.mgrid.or.kr) running on the grid testbed at the Konkuk University Applied Grid Computing Center. In addition, we have been extending the Glyco-MGrid system to build a shared simulation data repository for the Avian flu grid project at PRAGMA (http://avianflugrid.pragma-grid.net).

A grid computing-based monte carlo docking simulations approach for computational chiral discrimination

A validity of Grid computing with Monte Carlo (MC) docking simulations was examined in order to execute prediction and data handling for the computational chiral discrimination. Docking simulations were performed with various computational parameters for the chiral discrimination of a series of 17 enantiomers by β-cyclodextrin (β-CD) or by 6-amino-6-deoxy-β-cyclodextrin (am-β-CD). Rigid-body MC docking simulations gave more accurate predictions than flexible docking simulations. The accuracy was also affected by both the simulation temperature and the kind of force field. The prediction rate of chiral preference was improved by as much as 76.7% when rigid-body MC docking simulations were performed at low temperatures (100 K) with a sugar22 parameter set in the CHARMM force field. Our approach for Grid-based MC docking simulations suggested the conformational rigidity of both the host and guest molecule.

An analysis of idle CPU cycles at university computer labs

Grid computing has a great potential for grand challenge scientific problems such as Molecular Simulation, High Energy Physics and Genome Informatics. Exploiting under-utilized resources is crucial for a cost-effective, large-scale grid computing platform (i.e., computational grid), but there has been little research work on how to predict what resources will be underloaded in the near future. In this paper, we analyze idle CPU cycles of PCs at university computer labs and present techniques for predicting idle cycles to be efficiently scheduled for parallel/distributed computing. Our experiments with eight month monitoring data show that the accuracy of our prediction techniques is over 85%. Especially, the ratio of critical failure, which predicts that what is actually busy be idle, was only 3.2% out of total subject PCs during the experimental period.

Compressing Method of NetCDF Files Containing Clustered Sparse Matrix

Many results from scientific calculations are large-scale sparse matrices. The results of simulating volcanic ash diffusion are also a sparse matrix, and the values are clustered because of the characteristics of ash diffusion. The cost to store or transmit scientific data is usually high because such data are large scale. In this paper, we suggest a new storage format that is more efficient for storing clustered sparse matrix. Coordinate values are compressed more in the proposed format by saving only the first key value of consecutive non-zero elements and its length. The performance of the new format is the best among existing similar formats on ash diffusion simulation data, and the compressed size of the resulting file is comparable to a ZIP file. Because the new format can be applied partially to the data part of Network Common Data Form (NetCDF) files only, its header information is still readable directly from the compressed file, unlike zipped files.

Workflow Based on Pipelining for Performance Improvement of Volcano Disaster Damage Prediction System

A volcano disaster damage prediction system supports decision making for counteracting volcanic disasters by simulating meteorological condition and volcanic eruptions. In this system, a program called Fall3D generates predicted results for the diffusion of ash after a volcanic eruption on the basis of meteorological information. The relevant meteorological information is generated by a weather numerical prediction model known as Weather Research & Forecasting (WRF). In order to reduce the entire processing time without modifying these two simulation programs, pipelining can be used by partly executing Fall3D whenever the hourly (partial) results of WRF are generated. To reduce the processing time, successor programs such as Fall3D require that certain features be suspended until the part of the results that is based on prior calculation is generated by a predecessor. Even though Fall3D does not have a suspend or resume feature, pipelining effect can be produced by using the programs restart feature, which resumes simulation from the previous session. In this study, we suggest a workflow that can control the execution type.