ThaiGrid : From infrastructure to services and beyond. Putchong - - PowerPoint PPT Presentation

thaigrid from infrastructure to services and beyond
SMART_READER_LITE
LIVE PREVIEW

ThaiGrid : From infrastructure to services and beyond. Putchong - - PowerPoint PPT Presentation

ThaiGrid : From infrastructure to services and beyond. Putchong Uthayopas, ThaiGrid, SIPA Putchong_ut@thaigrid.or.th Piyawut Srichaikul, NECTEC Thai National Grid Project A National Project under Software Industry Promotion Agency (Public


slide-1
SLIDE 1

ThaiGrid : From infrastructure to services and beyond.

Putchong Uthayopas, ThaiGrid, SIPA Putchong_ut@thaigrid.or.th Piyawut Srichaikul, NECTEC

slide-2
SLIDE 2

Thai National Grid Project

  • A National Project under Software

Industry Promotion Agency (Public Organization) , Ministry of Information and Communication Technology

  • Started in 2005
  • Start from 14 member organizations and

expanded to 21 organizations in 2007

  • Goal

– Building the next generation E‐infrastructure for Thailand using Grid computing Technology as an enabling technology

slide-3
SLIDE 3

Project Organizational Structure

slide-4
SLIDE 4

21 members 21 members

slide-5
SLIDE 5
slide-6
SLIDE 6

Strategy 1: Creating Grid human resources

  • Encouraging the teaching in grid and parallel computing by creating

the educational platform

  • Community
  • Equipment
  • Courseware
  • Standard
  • Material for 3 training course

– Basic Grid Computing – Grid Administrator – Grid for Developers

  • Material for 3 university level undergraduate and graduate program

are being developed

– Introduction to grid computing – Intermediate Grid computing – Grid computing for Enterprise

slide-7
SLIDE 7

Important Statistics

  • Train more than 800 people during 2006‐

2007

– More than 600 people passed the training workshop at various level – Grid and parallel computing has been taught in 5 major university producing more than 200 grid‐enable CSE

  • We are pushing for more than 1000 people

this year in

slide-8
SLIDE 8

Strategy 2: Building the Grid Infrastructure

  • Thai Grid infrastructure is now functioning well
  • TERA Cluster

– Installed at Thai National Grid Center – Start to provide the service on July 11, 2007 – 200 nodes = 800 cores of Intel Xeon 3.0GHz – Rmax 2.5 Tflops (Rpeak 4.6Tflops)

  • Satellite Cluster

– 14 clusters installed at each member institute – Each cluster has 5 nodes = 10 CPUs of Intel Xeon 2.8GHz

  • Institute Cluster

– Existing cluster owned by some institutes – 3 institutes

  • Total number of CPU core now ~1,000 cores
slide-9
SLIDE 9

WU PSU CMU SUT KKU BU KMITNB KMUTT KMITL TMD KU SU MU CU AIT Tera

More than 1000 CPUs (August 2007)!

21 clusters from 16 sites around Thailand 21 clusters from 16 sites around Thailand

slide-10
SLIDE 10

Grid Middleware

  • Still sticking with Pre‐ws base services

– GT4 is installed in every member institutes, thus enable only Pre‐ws for now

  • Some institutes is moving to GT4

– 4 Institutes pass WS‐Gram test

  • Newly created services will base only on Web‐

services base technology or WSGram

slide-11
SLIDE 11

ThaiGrid Infrastructure usage

  • ThaiGrid provides more

than 111 years of computing time for member

– 7 years on the grid – 104 years on tera

  • 31 projects from 8 areas

are being support on Teraflop machine

  • More small projects on

each machines

slide-12
SLIDE 12

Strategy 3: Driving the Grid Adoption

slide-13
SLIDE 13

Building a Domain Specific Grid for E‐science

slide-14
SLIDE 14

Services on the Grid

  • CSE Online
  • Drug design Portal using Gridway
  • TAVERNA
  • Animation Grid using Web Services

infrastructure

  • Other commercial software provided in the

collaboration with users

– Fluent – Materials Studio

slide-15
SLIDE 15

Example of upcoming grid portal and e‐services

  • Virtual Research Center

http://vrc.thaigrid.or.th/vrcWeb

– Realtime e‐learning anfd collaboration support

  • http://biogrid.thaigrid.or.th/

– Bioinformatic data and service – Partners: PSU, Mahidol, KU, Chula, Biotec, and University of Manchester

  • Drug Discovery Portal

– Chula, KU – Job submission, vitual screening, database

  • Animation Portal

– Partners: Siam university, Thai Grid

slide-16
SLIDE 16

BioGrid

  • Building Grid infrastructure for Bio‐

informatics researcher in Thailand

  • Building network of bio‐informatics

researchers

  • Using TAVERNA + Web services to

drive job workflow

  • Partners

– Prince of Songkla University – University of Manchester – Mahidol University – Kasetsart University – Thai National Grid Center

  • Result

– A web‐based portal for Life‐science Grid base on TAVERNA is being developed – A network of Thai‐International researches is created – Activities

slide-17
SLIDE 17

Drug Design Grid

  • Building Grid‐base Drug Design infrastructure
  • Platform in Drug design grid

– CSE‐Online to drive drug design job – Portal site + Meta scheduler to distribute the job

  • Partners

– Chulalongkorn University – University of Utah – Thai National Grid Center

slide-18
SLIDE 18

Collaboration Grid

  • Using the Grid to drive “Virtual Research

Group”, making the collaboration between institutes more close

  • Partners

– KMUTT, CU, KMITNB, KU – Thai National Grid Center

  • ConferenceXP is currently in action

– Portal site to manage meeting – ConferenceXP is used for communication

slide-19
SLIDE 19

Computational Fluid Dynamics

  • Creating an environment for driving CFD simulation
  • Partners

– Kasetsart University – Thai National Grid Center

  • CFD for research

– Using existing software, such as FLUENT, in cluster level – Collaborating with Microsoft to utilize WindowsHPC cluster

  • CFD for industry

– Utilizing the use of virtual cluster for secured cluster environment – A proof‐of‐concept platform is being developed

slide-20
SLIDE 20

Selected Case

Model with Winglet, Diffuser and Rearview mirror

Velocity Vector

slide-21
SLIDE 21

Impact of blood flow heat transfer during cancer treatment with Hyperthermia method

  • Influence of seed number and configuration on

generated thermal field. Temperature in K.

slide-22
SLIDE 22

NanoGrid

  • Provide platform to help processing

nanotechnology simulation

  • Partners

– Computational Nano Science Consortium (Thailand) – Thai National Grid Center – Accelrys Software Inc.

  • Nanotechnology usage in Thailand

– Using Materials Studio, software from Accelrys co. ltd. in cluster level – Providing back‐end computing infrastructure to help nanotechnology research in Thailand – TNGC participate in porting Materials Studio 4.2 for SGE

slide-23
SLIDE 23

CNC members in Thailand

N

PSU PSU KU KU HW & SW NANOTEC NECTEC ThaiGrid MU MU CU CU CMU CMU KKU KKU SUT SUT AIT AIT UBU UBU

slide-24
SLIDE 24

Nitrous oxide (N2O) is a toxic pollutant gas in the atmosphere and a major cause of such greenhouse effect and global warming. Here we investigate, computationally at nanoscale using Dmol3, reaction mechanisms of the decomposition of toxic N2O to a non‐toxic N2 gas over pristine‐ and Titanium decorated‐carbon nanotubes (Ti/CNTs). By comparing activation barriers of such reaction on both catalysts, we found Ti/CNTs a strong candidate as catalyst in the removal of nitrous oxides (N2O).

N2O Reduction N2O Reduction

slide-25
SLIDE 25

Results from the Adsorption Locator calculations show such probability to find hydrogen molecules (H2) in nanoporous materials. For the metal‐organic frameworks‐5 (MOF‐5), H2 specifically prefers to locate near the metal oxide

  • clusters. No hydrogen molecule was found

near the phenyl ring or at the center of the pore. Unlike the MOF‐5, the zeolite imidazolate framework‐8 (ZIF‐8) shows different behavior

  • n the adsorption of hydrogen molecule,

which is likely to diffuse and locate in the pore space near the imidazole unit and infrequently settle in the zinc metal area.

MOF‐5 ZIF‐8

The behavior of H2 molecule in different nanoporous materials

slide-26
SLIDE 26

2008/4/2 Free template from www.brainybetty.com 26

Blood Brain Barrier

The model predicts blood‐brain barrier (BBB) after oral administration.

Egan and Lauri., Advanced Drug Delivery Reviews , 2002, 54, 273–289

In 2002, Egan and Lauri reported BBB model that used to predicts blood brain penetration after oral administration.

slide-27
SLIDE 27

Animation Grid

  • Provide environment for 3D

animation rendering in Thailand

  • Partners

– Siam University – Thai National Grid Center – Leading animation companies in Thailand

  • Currently an open‐source project
  • n Sourceforge called

“animagrid”

– Utilize web services and existing cluster and grid scheduler to distribute jobs – Base on commercial‐grade renderer software, such as Autodesk Maya or Blender

slide-28
SLIDE 28

Strategy 4: Driving Grid Research

  • 2006‐2007

– 2 National Conference on Grid computing (attendants 300‐500) – 3 International workshops – 67 Publication papers from members

  • 29 papers in 2006
  • 38 papers in 2007
  • Research Area

– Scientific computing in CFD, life science, physics, chemistry, math, Geo science – Grid: middleware, application, algorithm, scheduling, tools

slide-29
SLIDE 29

Strategy 5: Creating Grid Technology Awareness

  • Seminar, talk, workshop

– 9 events in 2006 – 35 event in 2007 ( about 3 events monthly)

  • 4 National Grid Competition (2000‐3000

people)

– Leading Edge (with Microsoft Asia Pacific)

  • University level : Engineering

– National Supercomputing Contest (SCC) (HP)

  • 200‐300 high schools

– Multicore‐Championship (Intel)

  • International Participation

– APAN, PRAGMA, SC2006, SC2007, GridAsia

  • News

– National newspaper, magazine – International

slide-30
SLIDE 30

Pilot Grid and Cluster infrastructure Interoperable Middleware Basic computational Grid Basic Data Grid POC applications 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 Computational Grid Data Grid Grid Portal Grid applications R&E and Non production enterprise applications Broad academic adoption Limited commercial adoption Stable Grid and Cluster infrastructure Interoperable Middleware and basic services Computational Grid Data Grid Grid Portal Grid applications R&E and Non production enterprise applications 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 Production Grid and Cluster infrastructure Rich set of services Computational Grid Enterprise analysis Grid Data and knowledge Grid Collaborative Grid Broad commercial Adoption

slide-31
SLIDE 31

Community driven model

slide-32
SLIDE 32

Objective 2009‐2011

slide-33
SLIDE 33

For more information please visit: http://www.thaigrid.or.th