Outline What is the GRID ? Why GRID at CBPF ? What are our needs ? Status of GRID at CBPF
What is the GRID ? Web is a service for sharing information over the Internet, Grid is a service for sharing computer power and data storage capacity over the Internet. The five big ideas Resource Sharing Secure Access Resource Use The Death of Distance Open Standards Direct access to remote software, computers and data Access policy, authentication and authorization You should be able to calculate the optimal allocation of resources High-speed connections between computers applications made to run on one resource will run on all others
LCG – Global GRID for High Energy Physics LCG LHC Computing Grid project Launched in 2002 at CERN. Mission: to integrate thousand of computers worldwide to store and analyze the huge amount of data that will be produced by the LHC. LHC will produce ~ 15 Petabytes of data (15x10 6 Gb) each year. Data sould be available to the thousand of scientists independent of their location. LCG involves today more than 200 sites in over 30 countries worldwide.
LCG is part of EGEE (Enabling Grids for E-sciencE) In April 2004 it was decided to build a permanent Grid infrastructure for scientific applications in Europe. The work has been carried out by a collaboration led by CERN. By the end of 2006, engineers and scientists of EGEE were managing ~ 20000 CPUs over 39 countries and 5 Pb of data storage. Six major scientific fields were included in the EGEE: physics, earth observation, climate prediction, petroleum exploration, astronomy and drug discovery. From Oct 2004 to Oct 2005 two million jobs have been successfully run on this Grid. EELA is a project related to EGEE Initiated in January 2006, coordinated by CIEMAT (Spain). Mission: to bring the e-Infrastructures of Latin American countries to the level of those of Europe. Will benefit of the Alice project and the RedCLARA network. Will focus on Grid infrastructure and related e-Science applications, identifying and promoting a sustainable framework for e-Science (in Latin America). America Latina Interconectada Com Europa Project set up in 2003 to develop the RedClara network 80% founded by the European Commission 19 Latin American and 4 European partners Cooperación Latino Americana de Redes Avanzadas Initiated in 2003 Linked to GÉANT (European advanced network)
Why GRID at CBPF ? CBPF has two groups participating in large experiments at LHC-CERN, LHCb and CMS. Both groups require of huge computational resources in terms of processing power and data storage. The CBPF computational facility has to be a dedicated resource for LHCb and CMS, but possibly open for other LHC experiments. The CBPF computational facility has to meet all the requirements of the CERN Data GRID
What are our needs/wishes ? Production center Distributes RAW data in quasi real time to Tier-1s Will hold a copy of RAW data Responsible for all the production and processing phases associated with real data, including (user) data analysis Primarily MC production centers Eventually in the future: data analysis Total CPU requirements for 2008: 12.97 MSI2k.years (1000 Intel Xeon 3.06 GHz = 6 TF = 1.1 MSI2k) CERN 7% Tier 1s 34% Tier 2s 59% Disk requirements for 2008: ~3.3PB CERN 0.8 PB Tier 1s 2.4 PB Tier 2s 0.1 PB 1.1 MB/s 90-128 MB/s year average CBPF wants to be a Tier 2
Status of GRID at CBPF 11 dual CPU dual core servers with a 160 GB Hard Disk. 1 dual core server with four 320 GB Hard Disks and two GigaBit network interface cards. All CPUs are Intel Xeon 5050 3.06 GHz – 64 bits. Initial GRID setup @ CBPF: SL 3.0.8 (32 bits) installed in all machines 1 Storage Element (SE) Server with 1.2 TB disk 1 Monitor (MON) 1 Computer Element (CE) 9 Worker Nodes (WN) Total cost, including a 24 ports hub-switch, two racks and a 10 KVA UPS ~ 120 KR$ Job request processing Torque server Scheduler Job distribution Data storageJob processing Middleware: CE LCG 2.7 SE gLite 3.0 MON gLite 3.0 WN LCG 2.7 Status: All computers certified. Software installed and configured. GRID node linked to a 1 Gb/s network (RedeRio). Onsite tests done and passed. Waiting for EELA tests and approval to be integrated to EELA Virtual Organization (VO).
Near future: New servers will be bought in the very near future (in the next two or three weeks) (~ 10-12 machines). New servers will be dual CPU, dual core, probably Intel Xeon 5050 or better. All the system will be installed in a definitive location. ~ 38 CPUs in the near future
CBPF is a Registration Authority (RA) Authorized by the Certification Authority (CA) at UFF What a RA do: Deals with users registration Deals with computational resources registration
"The world will only need five computers attributed to Thomas J. Watson, IBM "640 kilobytes is all the memory you will ever need" attributed to Bill Gates, Microsoft The end