Home : Site Map : Contact Us : Help : Links : Security

Resources in the UK Tier-1/A Centre

Batch CPU Cluster

4 racks with 156 dual P3 1.4GHz CPU systems, each with 1 Gb memory, 40 Gb internal disk and dual 100 Mbit/s network connection

3 racks with 80 dual P4/Xeon 2.66GHz CPU systems, each with 2 Gb memory, 30 Gb internal disk and dual 1 Gbit/s network connection

7 racks with 224 dual P4/Xeon 2.8GHz/533MHz CPU systems, each with 2 Gb memory, 120 Gb internal disk and dual 1 Gbit/s network connection

1 rack with 32 dual P4/Xeon 2.8GHz/533MHz CPU systems, each with 4 Gb memory, 120 Gb internal disk and dual 1 Gbit/s network connection

Disk Cluster

The disk cluster consists of 201 Tbyte of disk space. Of this:

• 146 Tbyte is available for use by HEP experiments
• 48 Tbyte is awaiting assignment
• 7 Tbyte is in reserve (for spaces etc.)

The disk space is served from about 60 dual-processor servers running Linux, each with 2 SCSI RAID controllers using IDE disks. Each server has a 1 Gbit/s interface to the network.

The Atlas DataStore

(Not related to the ATLAS LHC experiment!)

The Atlas Datastore uses an STK Powderhorn tape robot system which serves 200 Gbyte tapes. There are 1400 tapes at present, giving a current capacity of 270 Tbyte. When the robot is full with 6000 tapes, total capacity is 1.2 Pbyte.

Networking

Internal networking for the farm is a mixture of 100 Mbit/s and 1 Gbit/s up to switches which connect at 1 Gbit/s to the RAL backbone network.

The UK academic network, SuperJANET4 (SJ4), has a 20 Gbit/s backbone. RAL has a 2.5 Gbit/s connection into SJ4, and SJ4 has a 2.5 Gbit/s interconnect to GEANT.

Contact Andrew Sansum for further information

---

Last modified Mon 21 April 2008 . View page history

---

Switch to HTTPS . Website Help . Print View . Built with GridSite 1.4.3

For more about GridPP please contact Neasan O'Neill