A Challenge in Building a A Challenge in Building a National Scale - - PowerPoint PPT Presentation

A Challenge in Building a A Challenge in Building a National Scale Grid National Scale Grid Infrastructure in Thailand Infrastructure in Thailand

Putchong Uthayopas Putchong Uthayopas Thai National Grid Project Thai National Grid Project Software Industry Promotion Agency (SIPA) Software Industry Promotion Agency (SIPA) Ministry of Information and Communication Ministry of Information and Communication

Grid: What and Why

• Grid is a technology that incorporates distributed

computing systems into a massive computing and storage pool with efficient resource sharing.

• Grid can strengthen national competitiveness in

ICT industries as well as promote sciences and engineering research.

• Grid benefits several application domains:

– E-commerce, E-government, E-learning – Digital Media and Animation – Geoinformatics, Bioinformatics – Financial Analysis – Etc.

Thai National Grid Project

• A 4 years project under

Ministry of Information and Communication Technology

• 5.5 Million US$ program
• Start from 13 universities

and 1 government

• rganizations
• Extending rapidly

Objective

• To accelerate research and education in Thailand by

providing a strong infrastructure

• To increase the competitiveness of Thailand’s IT industry

using Grid technology

• To build collaborations for interdisciplinary research
• To create knowledge workers for the next generation IT

and enterprise computing using grid technology

• To increase awareness on Thailand’s IT competitiveness
• To create more business opportunities using grid

technology

Project Organization

Minister MICT National Grid Committee Director SIPA

Board SIPA

Admin Staffs

Member Institutes

Director Thai National Grid Center

Researchers

Board Thai National Grid Center Associate Director Thai National Grid Center

Partner Institutes

Bangkok

Chaingmai University (CMU) Chulalongkorn University (CU) Kasetsart University (KU) King Mongkut’s Institute of Technology North Bangkok (KMITNB) King Mongkut’s University of Technology Thonburi (KMUTT) King Mongkut’s Institute of Technology Ladkrabang (KMITL) Mahidol Unversity (MU) Silpakorn University (SU) Meteorological Department of Thailand (TMD) Walailak University (WU) Prince of Songkla University (PSU) Khonkhan University (KKU) Suranaree University of Technology (SUT)

Nakhon Ratchasima Khonkhan Chaingmai Nakhonsrithammarat Songkla

Thai Research & Education Networks

ThaiRen: UNINET ThaiSARN (NECTEC)

T h ai R esearch and E du cation N etw o rk

T h aiR E N

R E N O C U n iN e t T h a iS a rn

T h a iR E N

In te rn e t2 T E IN 2 A P A N

T h ai R esearch and E du cation N etw o rk

T h aiR E N

R E N O C U n iN e t T h a iS a rn

T h a iR E N

In te rn e t2 T E IN 2 A P A N

1 Gbps

ปตตานี ยะลา นราธิวาส แมฮองสอน เชียงราย ลําปาง พะเยา นาน แพร อุตรดิตถ สุโขทัย ลําพูน ตาก กําแพงเพชร เพชรบูรณ เลย ชัยภูมิ พิจิตร หนองบัว ลําภู อุดรธานี หนองคาย สกลนคร นครพนม กาฬสินธุ มหาสารคาม มุกดาหาร รอยเอ็ด ยโสธร อํานาจ เจริญ บุรีรัมย สุรินทร ศรีสะเกษ นครสวรรค อุทัยธานี กาญจนบุรี สุพรรณ บุรี ลพบุรี สระบุรี ปราจีนบุรี สระแกว ฉะเชิงเทรา จันทบุรี ระยอง ราชบุรี เพชรบุรี ตราด พระนครศรี

ปทุมธานี

สมุทรสงคราม สมุทรสาคร

ประจวบ คีรีขันธ ชุมพร ระนอง สุราษฏรธานี พังงา ภูเก็ต กระบี่ ตรัง สตูล พัทลุง นครนายก เชียงใหม นครราชสีมา อุบลราชธานี สงขลา นครศรีธรรมราช

Inter-University Network (UniNet)

Commission on Higher Education (Center)

24 Public Universities 39 Rajabhat Universities 29 IT Campus 8 Private Universities 13 Mahachulalongkorn Rajawitayalai 9 Other agencies 9 Rajamangala Institue of Technology (44 Campus) Ministry of Education

(for Under Higher Education)

155 Mbps- STM1 34 Mbps - E3 8 Mbps - 4*E1 2 Mbps - E1 64 k – 2Mbps

CMU CMU SUT SUT MUA MUA SU_NMJ SU_NMJ BU BU PSU PSU

Domestic Domestic

1Gbps 1Gbps Internet Internet -

• 622 Mbps

622 Mbps (to Internet 2 155 Mbps) (to Internet 2 155 Mbps) 2.5 Gbps- STM16 1 Gbps

Update on 24 Jan 2006 Update on 24 Jan 2006

THAISARN THAISARN 1Gbps 1Gbps MOEnet MOEnet (K (K-

• 12)

12) 1Gbps 1Gbps (coming soon) (coming soon)

Grid Infrastructure

ThaiRen (Uninet/ThaiSarn)

KMITNB KMUTT KMITL TMD WU KU PSU CMU SUT KKU SU MU CU AIT Teraflop Server at TNGC

2.5Gb

Potential Application Domain

• Life-science Grid
• Cheminformatics Grid
• Geoinformatics Grid
• Digital Archive and E-learning
• Computational Fluid Dynamics Grid
• Simulation and Modeling Grid
• Digital Media and Animation

Impact

• Accelerate research and education in Thailand by

providing a strong infrastructure

– Better ability to store, process, and retrieve massive data set faster – Better and more secured data exchange infrastructure for research, development and academics – Better collaboration for interdisciplinary research

• Enhance IT industry competitiveness
• more competitive of Thailand’s IT industry
• more awareness on Thailand’s IT
• more knowledge workers for enterprise computing using grid

technology

Strategies

• 1. Drive the creation of knowledge workers in Grid

technology

• 2. Build a reliable, secured, and scalable

infrastructure

• 3. Drive the adoption and application of grid

technology in both academic and industry

• 4. Drive research and development in Grid

technology

• 5. Drive the global awareness of Thailand’s

competency in Grid technology

2006 2007 2008 2009

Pilot Grid

R&E Grid

Enterprise Grid

• Pilot Grid and

Cluster infrastructure

• Interoperable

Middleware

• Basic

computational Grid

• Basic Data Grid
• POC applications
• Stable Grid and Cluster

infrastructure

• Interoperable Middleware

and basic services

• Basic computational Grid
• Basic Data Grid
• Grid Portal
• Grid applications
• R&E
• Non production

enterprise app

• Broad academic adoption
• Limited commercial

adoption

• Production Grid and

Cluster infrastructure

• Rich set of services
• Computational Grid
• Enterprise analysis

Grid

• Data and

knowledge Grid

• Collaborative Grid
• Broad commercial

adoption

Time Road Map

High Speed Network

Building the Grid

Grid Middleware

Bioinformatics Digital Media

HEP Medical

Coordination

• Building quick

communication infrastructure for technical contact

– Mailing list, wiki, discussion board, web site

• Building technical training

program

“The secret of war lies in the communication”

Napoleon Bonaparte

Approach

• Building a quick deployment technology

– NPACI Rocks is a key technology being used

• Collaboration SDSC/ TNGC under PRAGMA

– Excellence for first time deployment but need more dynamic deployment

• Building a good infrastructure monitoring and

management system

– Know what your grid are doing. We use SCMS tools (standard for PRAGMA as well) – Develop system monitoring, accounting, testing tools

Approach

• Site certification program

– Level 1: Network firewall compliance – Level 2: Middleware compliance (MPI/G2, Ninf-g) – Level 3: Job scheduling compliance – Level 4: Data Grid compliance

• Standard

– Define clear software stack

Solving network problem

• General problem

– Firewall Compliance – Dealing with smart intrusion detection system – Running Grid on Shared REN/Service network – Network underutilized, unbalance bandwidth

• Develop a network measurement infrastructure

– Collaboration PSU/TNGC/KU/Uninet

• Collaboration with network provider (Uninet)

Network Bandwidth OUT (Mbps)

5 10 15 20 SPI RI T(CU) GRI DCMU(CMU) ENQUEUE(KMI TNB) I NCA(KMUTT) SRI THAN(KKU) RDCGRI D1(NECTEC) RDCGRI D2(NECTEC) PSU-GRI D(PSU) SUTGRI D(SUT)

Network Bandwidth IN (Mbps)

20 40 60 80 SPI RI T(CU) GRI DCMU(CMU) ENQUEUE(KMI TNB) I NCA(KMUTT) SRI THAN(KKU) RDCGRI D1(NECTEC ) RDCGRI D2(NECTEC) PSU-GRI D(PSU) SUTGRI D(SUT)

Thai National Grid Standard Software Stack

Applications Core: Gamess, Autodock Option: Blast Portal: GridSphere Option: GridPort Management Core: SCMS Option: MOGAS Scheduler Core: SQMS , SGE Option: PBS, Condor, CSF Programming Ninf-G MPI-G2 Grid Middleware Globus Toolkit 4.0 Fabric Core: Linux / Rocks on x86_64/32 Option: Windows

C C C C H C C C C H C C C C H C C C C H REN GH

Grid Scheduler

Super Cluster Grid Super Cluster Grid Users see grid as a large single cluster

Thai Grid Operation Center

• Grid Operation Center

Software

– Automatic testing – Service Monitoring – Accounting – System monitoring

• In house development

Building Applications

• Define key competency and strategic

application area

– Need only 1-2 successful application per site

• Prefer application

– Real life application that solve real problem

• Not a short term thesis/project work
• Needed by a group of end users
• Need high computing power and running

frequently

– Interface with industry is preferred

ThaiGrid Drug Discovery Infrastructure

• Partners:

–

• Dr. Jak Saengma, KU

– Somsak Sriprayoonsakul, TNGC – IBM Thailand

• Challenges

– From over 10000 active compounds available from Thai medicinal plants database, find a smallest set of compounds that has a potential to be used as a drug – Very compute intensive. Several month of computing time

• Solution: Use grid to increase computing

power to 10-100 times

– Distributed screening process to hundred of hosts on the Grid

• Benefit

– Speeding up time to results – Allow scientists to spend more time analyzing results

Implicit Narrowband Level Set Method for Fire Spread Model

• Dr. Pallop Huabsomboon

Mahidol University

• Challenges

– Simulation of large scale firespred for forest fire – Very time consuming process

• Solution

– Distribute the computation on the Grid

• Benefit

– Expect 10x faster simulation speed – Ability to model much larger problem within a reasonable time

• Part of TNGP/ Hewlett Packard Join

Research Program

The Parallel Simulation Model for Thin Film Deposition Using the DSMC Method

• Dr. Rardchawadee Silapunt
• Dr. Tiranee Achalakul

KMUTT, Thailand

• Thin film deposition process is

used for many manufacturing industries including very large scale integrated circuits, optics, magnetic, hard and corrosion resistant coatings, micro- mechanics.

• Grid system will be used to

accelerate the computation and simulation process

• Part
• f

TNGP/ Hewlett Packard Join Research Program

Grid based data assimilation using RS data

• Dr. Kiyoshi Honda, AIT
• Dr. Putchong Uthayopas, THNC
• Challenges

– Using GA based data assimilation for agricultural monitoring – Very long computation time from months to years – Large amount of RS data needed to be moved around and process

• Solution:

– Using Grid to harvest more computing power – Hide data assimilation process behind the service using MPI, GridRPC (ninf-g)

• Benefit

– Speeding up the calculation

Medical Grid

Suwachai Sieanoon, Somkiat Chatchuenyot, Dr. Panjai Tantasnawong Computer Science Department, Silpakorn University

• To build and develop a

distributed information management system using DataGrid technology.

• Apply the distributed

information management system to test on exchanging and managing the Patient Records System, which is record complexity and, different format and large size

• f data.

G rid S e rv ic e s (In te rn e t)O p tio n H

• s

p ita l_ E G rid S e rv ic e s

H

• s

p ita l_ F H

• s

p ita l_ D

GRI D Services

Data Agent Service Registr y Meta Data (HL7)

XML I MAGE

Access Grid

• First APEC Virtual

Symposium on Pandemic Preparedness using Access Grid, January 20, 2006

• Presentation from

Medical Organization in many economies

– China, Taiwan, USA, Vietnam, Thailand, Singapore, Philippine

• Access Grid are used as

a key communication infrastructure

Building Grid Community

• Identify Key Player
• Identify their motivation
• Building up communication

structure

– For information dissemination – Web, wiki, mailing list

• Building up working group

– Resources working group – Education Working Group

Building an Awareness

• Educating potential users

– What is really Grid? – What it can do for real? – Reduced confusion, misunderstanding from the beginning

• Stimulating people to see the potential

– Focus on what Grid can be benefit to them. Users always comes first

Bridging the Gap

• Technical Competency Gap

– Different level of competency among partners

• Motivational Gap

– Different level of motivation and commitment

• Policy Gap

– Different level of management support and commitment

Partners

• Class 1: Get to

work

• Class 2: Training
• Class 3: Talk to

them, motivate their management

• Class 4: Move

them to class 2 later

1

Motivated

2 3 4

Not Motivated Competence Not Competence

Computing Center Computer Scientists Domain Experts Thai National Grid Partner Institute

Collaboration is Important

• Current collaboration

– PRAGMA

• GIN test bed http://goc.pragma-grid.net/gin/gin-

monitor.htm

– GRMAP (TNGC/NTU) – HP/IBM/Sun Microsystems/Oracle – APAN Grid Committee

SEAGF (South East Asia Grid Forum)

• Newly established forum of national grid projects in

South East Asia region

• Goal

– Act as a forum to define a common direction and policy for grid computing in the region – Creating research synergy among grid practitioners in the region

• Partners

– National Grid Office, Singapore – Thai National Grid Center, Thailand – MIMOS, Malaysia

Summary

• Thai National Grid Project is in the first

phase of infrastructure building

– Computing infrastructure – Networking infrastructure – Human infrastructure

• Moving toward

– Application development and deployment – Working with community to identify the application

• Community builder is critical to

success

– Building awareness, partnership

Thank you

me