Dynamic de-duplication of bibliographic data for user services
Los
Alamos National Laboratory, Research Library
DLF
Forum, October 26 2004, Baltimore, MD
LANL Netrics hardware setup
bibliographic
citation
IBM
HS20 blade server
Dual
2.8 GHz Xeon CPU
8 Gb
RAM
IBM
HS20 blade server
Dual
2.8 GHz Xeon CPU
8 Gb
RAM
IBM
HS20 blade server
Dual
2.8 GHz Xeon CPU
8 Gb
RAM
IBM
HS20 blade server
Dual
2.8 GHz Xeon CPU
8 Gb
RAM
IBM
HS20 blade server
Dual
2.8 GHz Xeon CPU
8 Gb
RAM
IBM
HS20 blade server
Dual
2.8 GHz Xeon CPU
8 Gb
RAM
IBM
HS20 blade server
Dual
2.8 GHz Xeon CPU
8 Gb
RAM
IBM
HS20 blade server
Dual
2.8 GHz Xeon CPU
8 Gb
RAM
• Cluster of IBM HS20
blade servers:
• Dual 2.4 Ghz Xeon CPU
• 8 Gb RAM
• Netrics:
• 1 daemon per 2 Gb RAM
• 4 daemons per blade server
• Bibliographic data:
• 36,000,000 biblio keys + identifier
• 6,000,000 per daemon
• 24,000,000 per blade server
• 2 blades for ISI bibliographic keys [12 Gb RAM]
• Citation data:
• 500,000,000 citation keys + identifier
• 136,000,000 unique citation keys +
identifiers
• 6,500,000 per daemon
• 26,000,000 per blade server
• 6 blades for ISI citation keys [42 Gb RAM]
RAM