Presentation is loading. Please wait.

Presentation is loading. Please wait.

CERN TERENA Lisbon The Grid Project Fabrizio Gagliardi CERN Information Technology Division May, 2000

Similar presentations


Presentation on theme: "CERN TERENA Lisbon The Grid Project Fabrizio Gagliardi CERN Information Technology Division May, 2000"— Presentation transcript:

1 CERN TERENA Lisbon The Grid Project Fabrizio Gagliardi CERN Information Technology Division May, 2000 F.Gagliardi@cern.ch

2 CERN TERENA Lisbon F. Gagliardi - CERN/IT-May-2000 2 Summary  High Energy Physics and CERN computing problem  An excellent computing model: the GRID  The Data Grid Initiative  (http://www.cern.ch/grid/)

3 CERN TERENA Lisbon F. Gagliardi - CERN/IT-May-2000 3 CERN organization CERN organization Largest Particle Physics lab in the world European International Center for Particle Physics Research Budget: 1020 M CHF 2700 staff 7000 physicist users

4 CERN TERENA Lisbon F. Gagliardi - CERN/IT-May-2000 4 The LHC Detectors CMS ATLAS LHCb 3.5 PetaBytes / year ~10 8 events/year

5 CERN TERENA Lisbon F. Gagliardi - CERN/IT-May-2000 5 The HEP Problem - Part I The scale...

6 CERN TERENA Lisbon F. Gagliardi - CERN/IT-May-2000 6 ~10K SI95 1200 processors Non-LHC technology-price curve (40% annual price improvement) LHC Capacity that can purchased for the value of the equipment present in 2000

7 CERN TERENA Lisbon F. Gagliardi - CERN/IT-May-2000 7 Non-LHC technology-price curve (40% annual price improvement) LHC

8 CERN TERENA Lisbon F. Gagliardi - CERN/IT-May-2000 8 Long Term Tape Storage Estimates Current Experiments COMPASS LHC 0 2'000 4'000 6'000 8'000 10'000 12'000 14'000 1995 1996 199719981999200020012002200320042005 2006 Year TeraBytes

9 CERN TERENA Lisbon F. Gagliardi - CERN/IT-May-2000 9 HPC or HTC High Throughput Computing  mass of modest, independent problems  computing in parallel – not parallel computing  throughput rather than single-program performance  resilience rather than total system reliability Have learned to exploit inexpensive mass market components But we need to marry these with inexpensive highly scalable management tools Much in common with other sciences (see EU-US Annapolis Workshop at www.cacr.caltech.edu/euus): Astronomy, Earth Observation, Bioinformatics, and commercial/industrial: data mining, Internet computing, e-commerce facilities, …… Contrast with supercomputing

10 network servers tape servers disk servers application servers Generic component model of a computing farm

11 CERN TERENA Lisbon F. Gagliardi - CERN/IT-May-2000 11 The HEP Problem - Part II Geography, Sociology, Funding and Politics...

12 CERN TERENA Lisbon F. Gagliardi - CERN/IT-May-2000 12 CMS:1800 physicists 150 institutes 32 countries World Wide Collaboration  distributed computing & storage capacity

13 CERN TERENA Lisbon F. Gagliardi - CERN/IT-May-2000 13 Regional Centres - a Multi-Tier Model Department    Desktop CERN – Tier 0 MONARC report: http://home.cern.ch/~barone/monarc/RCArchitecture.html Tier 1 FNAL RAL IN2P3 622 Mbps 2.5 Gbps 622 Mbps 155 mbps Tier2 Lab a Uni b Lab c Uni n

14 CERN TERENA Lisbon F. Gagliardi - CERN/IT-May-2000 14 Are Grids a solution?  Change of orientation of US Meta-computing activity  From inter-connected super-computers ….. towards a more general concept of a computational Grid (The Grid – Ian Foster, Carl Kesselman)  Has initiated a flurry of activity in HEP  US – Particle Physics Data Grid (PPDG)  GriPhyN – data grid proposal submitted to NSF  Grid technology evaluation project in INFN  UK proposal for funding for a prototype grid  NASA Information Processing Grid

15 CERN TERENA Lisbon F. Gagliardi - CERN/IT-May-2000 15 The Grid “Dependable, consistent, pervasive access to [high-end] resources” Dependable: provides performance and functionality guarantees Consistent: uniform interfaces to a wide variety of resources Pervasive: ability to “plug in” from anywhere

16 CERN TERENA Lisbon F. Gagliardi - CERN/IT-May-2000 16 R&D required Local fabric  Management of giant computing fabrics  auto-installation, configuration management, resilience, self- healing  Mass storage management  multi-PetaByte data storage, “real-time” data recording requirement, active tape layer – 1,000s of users Wide-area - building on an existing framework & RN (e.g.Globus, Geant)  workload management  no central status  local access policies  data management  caching, replication, synchronisation  object database model  application monitoring

17 CERN TERENA Lisbon F. Gagliardi - CERN/IT-May-2000 17 HEP Data Grid Initiative  European level coordination of national initiatives & projects  Principal goals:  Middleware for fabric & Grid management  Large scale testbed - major fraction of one LHC experiment  Production quality HEP demonstrations  “mock data”, simulation analysis, current experiments  Other science demonstrations  Three year phased developments & demos  Complementary to other GRID projects  EuroGrid: Uniform access to parallel supercomputing resources  Synergy to be developed (GRID Forum, Industry and Research Forum)

18 CERN TERENA Lisbon F. Gagliardi - CERN/IT-May-2000 18 Participants  Main partners: CERN, INFN(I), CNRS(F), PPARC(UK), NIKEF(NL), ESA-Earth Observation  Other sciences: KNMI(NL), Biology, Medicine  Industrial participation: CS SI/F, DataMat/I, IBM/UK  Associated partners: Czech Republic, Finland, Germany, Hungary, Spain, Sweden (mostly computer scientists)  Formal collaboration with USA  Industry and Research Project Forum with representatives from:  Denmark, Greece, Israel, Japan, Norway, Poland, Portugal, Russia

19 CERN TERENA Lisbon F. Gagliardi - CERN/IT-May-2000 19 Status  Prototype work already started at CERN and in most of collaborating institutes  Proposal to RN2 submitted  Network requirements discussed with Dante/Geant

20 CERN TERENA Lisbon F. Gagliardi - CERN/IT-May-2000 20 WAN Requirements  High bandwidth from CERN to Tier 1 centres (5-6)  VPN, Quality of Service  Guaranteed performance during limited test periods and at the end of the project for production quality services  Target requirements (2003) 2.5 Gb/s + 622 Mb/s + 155 Mb/s  Could saturate for limited amount of test time 2.5 Gb/s (100 MB/s out from a 100 PC farm, we plan for 1000’s PC farm)  Reliability is an important factor:  from WEB client-server model to  GRID peer distributed computing model

21 CERN TERENA Lisbon F. Gagliardi - CERN/IT-May-2000 21 Conclusions  This project, motivated by HEP and other high data and computing demanding sciences, will contribute to develop and implement a new world-wide distributed computing model: The GRID  An ideal computing model for the next generation Internet  An excellent test case for the next generation of high- performance research networks


Download ppt "CERN TERENA Lisbon The Grid Project Fabrizio Gagliardi CERN Information Technology Division May, 2000"

Similar presentations


Ads by Google