CC-IN2P3: A High Performance Data Center for Research

Similar documents
The LHC Computing Grid

CERN and Scientific Computing

Grid Computing a new tool for science

CERN openlab II. CERN openlab and. Sverre Jarp CERN openlab CTO 16 September 2008

CSCS CERN videoconference CFD applications

Giovanni Lamanna LAPP - Laboratoire d'annecy-le-vieux de Physique des Particules, Université de Savoie, CNRS/IN2P3, Annecy-le-Vieux, France

Towards a Strategy for Data Sciences at UW

The LHC computing model and its evolution. Dr Bob Jones CERN

Scientific data processing at global scale The LHC Computing Grid. fabio hernandez

Conference The Data Challenges of the LHC. Reda Tafirout, TRIUMF

Online data storage service strategy for the CERN computer Centre G. Cancio, D. Duellmann, M. Lamanna, A. Pace CERN, Geneva, Switzerland

Software and computing evolution: the HL-LHC challenge. Simone Campana, CERN

Travelling securely on the Grid to the origin of the Universe

The CMS Computing Model

Experience of the WLCG data management system from the first two years of the LHC data taking

Physics Computing at CERN. Helge Meinhard CERN, IT Department OpenLab Student Lecture 27 July 2010

Andrea Sciabà CERN, Switzerland

Physics Computing at CERN. Helge Meinhard CERN, IT Department OpenLab Student Lecture 21 July 2011

From raw data to new fundamental particles: The data management lifecycle at the Large Hadron Collider

High Performance Computing from an EU perspective

New strategies of the LHC experiments to meet the computing requirements of the HL-LHC era

Review of the Compact Muon Solenoid (CMS) Collaboration Heavy Ion Computing Proposal

Stephen J. Gowdy (CERN) 12 th September 2012 XLDB Conference FINDING THE HIGGS IN THE HAYSTACK(S)

Insight: that s for NSA Decision making: that s for Google, Facebook. so they find the best way to push out adds and products

CHIPP Phoenix Cluster Inauguration

CERN Lustre Evaluation

High-Energy Physics Data-Storage Challenges

Overview. About CERN 2 / 11

N. Marusov, I. Semenov

Storage and I/O requirements of the LHC experiments

The Grid: Processing the Data from the World s Largest Scientific Machine

Storage Resource Sharing with CASTOR.

IEPSAS-Kosice: experiences in running LCG site

Distributed e-infrastructures for data intensive science

A short introduction to the Worldwide LHC Computing Grid. Maarten Litmaath (CERN)

PLAN-E Workshop Switzerland. Welcome! September 8, 2016

ACCI Recommendations on Long Term Cyberinfrastructure Issues: Building Future Development

High Performance Computing (HPC) Data Center Proposal

Grid Computing at the IIHE

Managing Petabytes of data with irods. Jean-Yves Nief CC-IN2P3 France

Tackling tomorrow s computing challenges today at CERN. Maria Girone CERN openlab CTO

e-research Infrastructures for e-science Axel Berg SARA national HPC & e-science support center RAMIRI, June 15, 2011

Grid Computing: dealing with GB/s dataflows

AGIS: The ATLAS Grid Information System

Reliability Engineering Analysis of ATLAS Data Reprocessing Campaigns

The LHC Computing Grid. Slides mostly by: Dr Ian Bird LCG Project Leader 18 March 2008

CouchDB-based system for data management in a Grid environment Implementation and Experience

Challenges and Evolution of the LHC Production Grid. April 13, 2011 Ian Fisk

LHC Computing Grid today Did it work?

Constant monitoring of multi-site network connectivity at the Tokyo Tier2 center

The Grid. Processing the Data from the World s Largest Scientific Machine II Brazilian LHC Computing Workshop

Using the In-Memory Columnar Store to Perform Real-Time Analysis of CERN Data. Maaike Limper Emil Pilecki Manuel Martín Márquez

Invenio: A Modern Digital Library for Grey Literature

Preparing for High-Luminosity LHC. Bob Jones CERN Bob.Jones <at> cern.ch

Data Transfers Between LHC Grid Sites Dorian Kcira

Batch Services at CERN: Status and Future Evolution

The INFN Tier1. 1. INFN-CNAF, Italy

Virtualizing a Batch. University Grid Center

ATLAS Experiment and GCE

The Future of High Performance Computing

Big Computing and the Mitchell Institute for Fundamental Physics and Astronomy. David Toback

Summary of the LHC Computing Review

e-science for High-Energy Physics in Korea

arxiv: v1 [physics.ins-det] 1 Oct 2009

Brief review of the HEPIX 2011 spring Darmstadt, 2-6 May

Big Data Analytics and the LHC

Operating the Distributed NDGF Tier-1

Storage on the Lunatic Fringe. Thomas M. Ruwart University of Minnesota Digital Technology Center Intelligent Storage Consortium

STFC Cloud Developments. Ian Collier & Alex Dibbo STFC Rutherford Appleton Laboratory UK Research & Innovation HEPiX Spring, Madison, WI 17 May 2018

CC-IN2P3 / NCSA Meeting May 27-28th,2015

Opportunities A Realistic Study of Costs Associated

Distributing storage of LHC data - in the nordic countries

Philippe Laurens, Michigan State University, for USATLAS. Atlas Great Lakes Tier 2 collocated at MSU and the University of Michigan

The grid for LHC Data Analysis

Investor day. November 17, IT business Laurent Vernerey Executive Vice President

Worldwide Production Distributed Data Management at the LHC. Brian Bockelman MSST 2010, 4 May 2010

Top 4 considerations for choosing a converged infrastructure for private clouds

First Experience with LCG. Board of Sponsors 3 rd April 2009

Grid and Cloud Activities in KISTI

Scientific Computing at SLAC. Amber Boehnlein

Distributed Computing Framework. A. Tsaregorodtsev, CPPM-IN2P3-CNRS, Marseille

The coolest place on earth

Austrian Federated WLCG Tier-2

SILECS Super Infrastructure for Large-scale Experimental Computer Science

The JINR Tier1 Site Simulation for Research and Development Purposes

irods usage at CC-IN2P3: a long history

RUSSIAN DATA INTENSIVE GRID (RDIG): CURRENT STATUS AND PERSPECTIVES TOWARD NATIONAL GRID INITIATIVE

Reprocessing DØ data with SAMGrid

CC-IN2P3 activity. irods in production: irods developpements in Lyon: SRB to irods migration. Hardware setup. Usage. Prospects.

Volunteer Computing at CERN

Pan-European Grid einfrastructure for LHC Experiments at CERN - SCL's Activities in EGEE

ATLAS 実験コンピューティングの現状と将来 - エクサバイトへの挑戦 坂本宏 東大 ICEPP

Grid Computing: dealing with GB/s dataflows

Application of Virtualization Technologies & CernVM. Benedikt Hegner CERN

Atlantis: Visualization Tool in Particle Physics

DESY at the LHC. Klaus Mőnig. On behalf of the ATLAS, CMS and the Grid/Tier2 communities

De BiG Grid e-infrastructuur digitaal onderzoek verbonden

The creation of a Tier-1 Data Center for the ALICE experiment in the UNAM. Lukas Nellen ICN-UNAM

CLOUDS OF JINR, UNIVERSITY OF SOFIA AND INRNE JOIN TOGETHER

The European DataGRID Production Testbed

Transcription:

April 15 th, 2011 CC-IN2P3: A High Performance Data Center for Research Toward a partnership with DELL Dominique Boutigny

Agenda Welcome Introduction to CC-IN2P3 Visit of the computer room Lunch Discussion about future partnership Discussion about a possible common event in September 2

CC-IN2P3 within the French scientific framework Main scientific organization in France Dedicated Computing Center National Institute for Nuclear Physics and Particle Physics CC-IN2P3 federates the main computing resources For : High energy physics Nuclear physics Astroparticle physics + some opening to other sciences Manpower: 82 people Budget: 10-12 M with salaries Irfu Atomic Energy Commission 3

A diversity of scientific applications CC-IN2P3 is part of a worldwide network of datacenters for High Energy Physics 4

LHC = Large Hadron Collider The LHC is built at CERN in a 27 km tunnel, 100 m under the Geneva region 2 proton beams circulating in opposite directions Collisions in 4 experimental areas 10 years of construction Cost of the accelerator~3 B More than 900 superconducting magnet Several 1000 physicists and engineers involved

The Large Hadron Collider (LHC): the today s challenge ATLAS 4 experiments with a level of complexity never reached before CMS 10 billions collisions recorded each year Up to 15 PB/year of raw data 6

7

A global grid infrastructure CC-IN2P3 Annecy Nantes CC-IN2P3 FZK TRIUMF RAL CC-IN2P3 T1 (11) Grenoble T3 (many) ASCC CNAF Île de France Lyon Strasbourg Brookhaven NIKHEF T0 NDGF PIC Fermilab Clermont T2 (~60) Marseille 8

9

Grid activities CC-IN2P3 is the main node of a French multidisciplinary grid (France Grille) which is part of the European infrastructure EGI Operation Training Distributed data management Support Interoperability tools and portals based on standards Evolution toward the cloud world Regional initiative 10

CC-IN2P3 current status Today, computing power at CC-IN2P3 is dominated by LHC But several other experiments (~40) relies on CC-IN2P3 for their computing 12 000 cores 10 PB of disk storage 4 robotic silos (up to 40 Pbytes capacity) 11

CC-IN2P3 specificities Very complex infrastructure with a strong specificity on data management, data movement, data mining We start to see limitations related to multicore architecture strong impact on CPU efficiency We need to evaluate new architecture and to propose new solutions At the moment we are leading this effort in the high energy physics community Partnership with DELL 12

Infrastructure evolution The very fast increase of the computing equipments installed at CC-IN2P3 created a huge load on the infrastructure 13

The infrastructure is at its limit A lot of money invested since 6 years But not enough to fulfill our long term scientific commitments 14

Build a new datacenter The design parameters used to define the size of the new datacenter has been the following: Serve ~40 experiments with "standards needs" Fulfill LHC computing commitments and provide first class analysis capability Expect a very significant growth of the Astroparticle community needs: LSST, EUCLID) Add some capacity for network, services, etc. ~ 80 000 today s cores 15

New datacenter Assume Moore law is still valid Extrapolate up to 2019 End up with: indicated power is for computing only power for cooling has to be added 2011 50 racks 600 kw 2015 125 racks 1.5 MW 2019 216 racks 3.2 MW On top of the existing computer room (1 MW) Due to budget constraints the new computer room will start with a limited power (600 kw) Modular design Easily scalable Chilled water and electricity distribution has been designed for the 2019 target value Equipment: transformers, chillers, UPS etc will be added later 16

The new datacenter 17

Computer room in the 2 nd floor (850 m 2 ) All the cooling equipment is on the roof Connection between the 2 computer rooms + elevator Electrical equipment in the basement 18

19

New datacenter The new computer room is available and first computers are being installed An upgrade plan do exist to bring it to full power by 2019 (4.2 MW total capacity for computing equipment) Plenty of room, cooling and power CC-IN2P3 is the only datacenter in the HEP field with such a perspective Huge possibilities for worldwide scientific projects 20

Theoretical physics Lattice QCD Lattice Quantum ChromoDynamics is an area of theoretical physics that needs intensive computation The scientific challenge is huge as many experimental results cannot be interpreted without LQCD inputs Requires massively parallel computation with high speed interconnection Initially the IBM Blue Gene architecture as been developed by former LQCD physicist in Watson labs LQCD need dedicated petaflop computers GPU are very good candidates to build a dedicated LQCD computer at acceptable cost Partnership with DELL 21