18 th WGISS Meeting, September 6-10, 2004, Beijing, People Republic of China Activities on Grid Technology at GISTDA Pakorn Apaphant GISTDA
Mar 27, 2015
18th WGISS Meeting, September 6-10, 2004, Beijing, People Republic of China
Activities on Grid Technology at GISTDA
Pakorn ApaphantGISTDA
18th WGISS Meeting, September 6-10, 2004, Beijing, People Republic of China
Grid Activities in Thailand
• Grid technology is– Being explored by university sector– Collaboration projects
• ThaiGrid• ApGrid, APAN, PRAGMA
– Still under experimental phase in Thailand
18th WGISS Meeting, September 6-10, 2004, Beijing, People Republic of China
ThaiGrid• A partnership project for grid computing
started since December 2000– Kasetsart University– King Mongkut’s Institute of Technology North
Bangkok– Suranaree University of Technology – Asian Institute of Technology– Chulalongkorn University– Walailak University– Chiangmai University– KMUTT– NECTEC (National Electronics and Computing
Technology)
18th WGISS Meeting, September 6-10, 2004, Beijing, People Republic of China
ThaiGrid Testbed
• 6 Universities collaboration
• 100+ processors
www.thaigrid.net
18th WGISS Meeting, September 6-10, 2004, Beijing, People Republic of China
Pilot study: GISTDA Grid Computing
Data Data
Laptop com puter
Laptop com puter Laptop com puter
GISTDA Ground Station
Data Data
Laptop com puter
Laptop com puter Laptop com puter
Users
INTRANET
Data AnalysisGISTDA Gateway-Data Clearing House-Map Server
PM OC
GISTDA HQ
18th WGISS Meeting, September 6-10, 2004, Beijing, People Republic of China
• How to build a ubiquitous data sharing infrastructure among geographically distributed sites
• How to process, manage, store large data set– Transfer large satellite images
between “workstation-server” and “server-server”
• How to build secure environment for organization critical data
Challenges !!
18th WGISS Meeting, September 6-10, 2004, Beijing, People Republic of China
Status
• Encouraged by WGISS/CEOS• A pilot project is funded to explore
the challenges/ problems involve –June 2004
• System will be implemented in couple month
18th WGISS Meeting, September 6-10, 2004, Beijing, People Republic of China
Solution
• Using Grid and cluster computing – Using grid replica service to unify
geographically distributed data.– Use cluster and parallel file systems
technology to provide large and scalable file storage at economical cost
– Deploy grid middleware (Globus 3.2) to build a secure infrastructure
18th WGISS Meeting, September 6-10, 2004, Beijing, People Republic of China
IBM
IBM
Data Grid & Replica Catalogue
File Server
GISTDA Ground StationGISTDA HQ
IBM
IBM
Super File Server
Data Grid technology will be deployed to link 2 sites
“Using grid replica service to unify geographically distributed data”
18th WGISS Meeting, September 6-10, 2004, Beijing, People Republic of China
I/O 4
I/O 2
Frontend 2
I/O n
I/O 3
I/O 1
Frontend 1
Division 2
Division 1 Super File ServerConfiguration
• Cluster front-end is configured to allow workstations to seamlessly get the data from back-end
• Multiple I/O nodes work collectively using parallel file system (PVFS)
“Use cluster and parallel file systems technology to provide large and scalable file storage at economical cost”
18th WGISS Meeting, September 6-10, 2004, Beijing, People Republic of China
NETWORK
MGR
I/O N 1PVFS
mount &
Samba Gateway
I/O N 2
I/O N n
Metadata
Disks
NETWORK
Window Client
Applies Storage Server <= PVFS + Samba
Private Network Public Network
NETWORK
Window Client
Window Client
18th WGISS Meeting, September 6-10, 2004, Beijing, People Republic of China
Expected Benefits
• Reduce the cost of building scalable storage by making use of open source technology
• Reduce access time for large file using powerful cluster file server
• Strong secure infrastructure using grid security
• Building computing power for further complex applications.
18th WGISS Meeting, September 6-10, 2004, Beijing, People Republic of China
Conclusion
• Grid and cluster have been considered as potential technology to build a strong and secure IT infrastructure
• Wider deployment in the future using experiences learned