GIS and Remote Sensing Unit at FEM-CRI

FEM-GIS cluster

The FEM GIS cluster (Dell Blades system) consists of 12 single-blades with 2 processors each (4 Core Xeon 2.5GHz, 16Gb/32Gb RAM) and 2 double-blades (4x Quad Core Opteron (2.2GHz, 64Gb RAM), all connected via 10 Gb ethernet, so in total 128 nodes with 400 Gb RAM. The file server is a 4-Core Xeon 2.66GHz with 8Gb RAM.

Blades

• PowerEdge M1000e Blade Enclosure
• 12 PE M600 Quad-Core Xeon E5420 2.5GHz/2x6MB 1333FSB, RAM 16GB 667MHz FBD (8x2GB dual rank DIMMs), 2 x 73GB SAS (15,000rpm) 2.5in Hard Drive
• 2 PE M905 4x Quad Core Opteron 8354 (2.2GHz, RAM 64GB, 2 x 73GB SAS (15,000rpm) 2.5in Hard Drive
• 2 PowerConnect M6220 switch 20 Port Redundant (FI)

File server

• 1 PE2950 III Quad-Core Xeon E5430 2.66GHz/2x6MB 1333FSB
• 1 MD3000 external SAS RAID array with 2 single-port controllers
• 14 * 750 GB

UPS

• APC Smart-UPS RT 8000VA - UPS ( esterno ) - 220/230/240 V c.a. V - 8000VA - Ethernet 10/100 - 9 Connettori uscita - 6U

Operating System

• Scientific Linux 5.2: http://www.scientificlinux.org
• Blades: Scientific Linux 5.2 Live CD and DVD - Diskless Client: http://www.livecd.ethz.ch/diskless.html

Backup

• Dell PowerVault 124T, dual magazine tape autoloader, used with Bacula backup software
  

Job scheduler

• Grid Engine (GE 6.1u5): http://gridengine.sunsource.net/

What we do on this cluster

Our projects on infectious diseases require the generation of ecological indicators from vast GIS and remote sensing data bases. For example, we postprocess and reconstruct time series of MODIS satellite data and generate photoperiod maps from high resolution elevation maps which are relevant to distribution modelling of parasites. See our project pages (left menu) to learn more.