KIT Site Report. Andreas Petzold. www.kit.edu STEINBUCH CENTRE FOR COMPUTING - SCC



Similar documents
KIT Site Report. Andreas Petzold. STEINBUCH CENTRE FOR COMPUTING - SCC

Preview of a Novel Architecture for Large Scale Storage

The cloud storage service bwsync&share at KIT

Data storage services at CC-IN2P3

Why long time storage does not equate to archive

Lessons learned from parallel file system operation

bwgrid Treff HD/MA Sabine Richling, Heinz Kredel Universitätsrechenzentrum Heidelberg Rechenzentrum Universität Mannheim 30.

OSG Hadoop is packaged into rpms for SL4, SL5 by Caltech BeStMan, gridftp backend

Smart Data Innovation Lab (SDIL)

(Possible) HEP Use Case for NDN. Phil DeMar; Wenji Wu NDNComm (UCLA) Sept. 28, 2015

Steinbuch Centre for Computing (SCC) The Information Technology Centre of KIT

Kriterien für ein PetaFlop System

Mass Storage System for Disk and Tape resources at the Tier1.

irods at CC-IN2P3: managing petabytes of data

Maurice Askinazi Ofer Rind Tony Wong. Cornell Nov. 2, 2010 Storage at BNL

Tier0 plans and security and backup policy proposals

Scientific Storage at FNAL. Gerard Bernabeu Altayo Dmitry Litvintsev Gene Oleynik 14/10/2015

Panasas at the RCF. Fall 2005 Robert Petkus RHIC/USATLAS Computing Facility Brookhaven National Laboratory. Robert Petkus Panasas at the RCF

Storage strategy and cloud storage evaluations at CERN Dirk Duellmann, CERN IT

bwgrid Treff HD/MA Sabine Richling, Heinz Kredel Universitätsrechenzentrum Heidelberg Rechenzentrum Universität Mannheim 30.

Managed GRID or why NFSv4.1 is not enough. Tigran Mkrtchyan for dcache Team

Large File System Backup NERSC Global File System Experience

Lustre tools for ldiskfs investigation and lightweight I/O statistics

NERSC File Systems and How to Use Them

Integration of Virtualized Workernodes in Batch Queueing Systems The ViBatch Concept

Flexible Scalable Hardware independent. Solutions for Long Term Archiving

HTCondor at the RAL Tier-1

IT of SPIM Data Storage and Compression. EMBO Course - August 27th! Jeff Oegema, Peter Steinbach, Oscar Gonzalez

A parallel file system made in Germany Tiered Storage and HSM

Forschungszentrum Karlsruhe in der Helmholtz-Gemeinschaft. dcache Introduction

GPFS und HPSS am HLRS

CERN Cloud Storage Evaluation Geoffray Adde, Dirk Duellmann, Maitane Zotes CERN IT

Data management challenges in todays Healthcare and Life Sciences ecosystems

Big Science and Big Data Dirk Duellmann, CERN Apache Big Data Europe 28 Sep 2015, Budapest, Hungary

EMC DATA PROTECTION. Backup ed Archivio su cui fare affidamento

HAMBURG ZEUTHEN. DESY Tier 2 and NAF. Peter Wegner, Birgit Lewendel for DESY-IT/DV. Tier 2: Status and News NAF: Status, Plans and Questions

Globus and the Centralized Research Data Infrastructure at CU Boulder

Computational infrastructure for NGS data analysis. José Carbonell Caballero Pablo Escobar

How To Build A Supermicro Computer With A 32 Core Power Core (Powerpc) And A 32-Core (Powerpc) (Powerpowerpter) (I386) (Amd) (Microcore) (Supermicro) (

Altix Usage and Application Programming. Welcome and Introduction

Performance, Reliability, and Operational Issues for High Performance NAS Storage on Cray Platforms. Cray User Group Meeting June 2007

What is the real cost of Commercial Cloud provisioning? Thursday, 20 June 13 Lukasz Kreczko - DICE 1

IBM System Storage Portfolio Overview

Virtualization, Grid, Cloud: Integration Paths for Scientific Computing

Sun Constellation System: The Open Petascale Computing Architecture

Virtualization of a Cluster Batch System

High Performance Computing within the AHRP

CERN local High Availability solutions and experiences. Thorsten Kleinwort CERN IT/FIO WLCG Tier 2 workshop CERN

Estonian Scientific Computing Infrastructure (ETAIS)

Computing in High- Energy-Physics: How Virtualization meets the Grid

IBM System x SAP HANA

Archival Storage At LANL Past, Present and Future


SURFsara HPC Cloud Workshop

The Hartree Centre helps businesses unlock the potential of HPC

JUROPA Linux Cluster An Overview. 19 May 2014 Ulrich Detert

Big Data and the Earth Observation and Climate Modelling Communities: JASMIN and CEMS

Deploying and managing a Visualization Onera

IT and Storage for Big Data Analytics

Appro Supercomputer Solutions Best Practices Appro 2012 Deployment Successes. Anthony Kenisky, VP of North America Sales

The dcache Storage Element

risk of failure of or access to system applications. Backup software used for backup management is Bakbone Netvault : Backup.

SURFsara HPC Cloud Workshop

U-LITE Network Infrastructure

Open Cirrus: Towards an Open Source Cloud Stack

Report from SARA/NIKHEF T1 and associated T2s

Big + Fast + Safe + Simple = Lowest Technical Risk

Mass Storage at GridKa

Solution for private cloud computing

Computing at the HL-LHC

Forschungszentrum Karlsruhe in der Helmholtz - Gemeinschaft. Holger Marten. Holger. Marten at iwr. fzk. de

Deploying Flash- Accelerated Hadoop with InfiniFlash from SanDisk

Virtualisation Cloud Computing at the RAL Tier 1. Ian Collier STFC RAL Tier 1 HEPiX, Bologna, 18 th April 2013

DSS. High performance storage pools for LHC. Data & Storage Services. Łukasz Janyst. on behalf of the CERN IT-DSS group

Big Data and Cloud Computing for GHRSST

Next Generation Tier 1 Storage

Transcription:

KIT Site Report Andreas Petzold STEINBUCH CENTRE FOR COMPUTING - SCC KIT University of the State of Baden-Württemberg and National Laboratory of the Helmholtz Association www.kit.edu

GridKa Tier 1 - Batch Farm Hardware 620 WNs, mix of Intel & AMD 13k job slots, 150kHS06 (2014) 42kHS06 will be retired, 52kHS06 new Univa Grid Engine 8.1.7 very smooth operation multi-core jobs dynamic slot allocation, no static cluster partition 2 13.10.14 Andreas Petzold KIT Site Report HEPiX 2014 Lincoln, NE

GridKa Tier 1 Disk Storage 14PB on DDN systems S2A9900, SFA10k, SFA12k dcache for ATLAS, CMS, LHCb, Belle2, others Version 2.6.33/34 plan to move to 2.10 by December CMS now using NFSv4.1 automatic file replication off manual data set distribution ATLAS & CMS instances integrated in FAX & AAA w/o xrootd proxy xrootd for ALICE some servers already on version 4 3 13.10.14 Andreas Petzold KIT Site Report HEPiX 2014 Lincoln, NE

GridKa Tier 1 Tape Storage 15.5PB used in 3 libraries (STK, IBM, Grau) tape technology currently LTO3/4/5 with 45 drives, ~22k cartridges 6 new T10kC drives, 400 cartridges in production by end of year tape management currently TSM & ERMM gradual migration to HPSS to start mid 2015 HPSS status intensive testing phase finished dedicated new STK library 1 st user: HLRS (Stuttgart) will archive data to HPSS at KIT starting January 2015 4 13.10.14 Andreas Petzold KIT Site Report HEPiX 2014 Lincoln, NE

GridKa Tier 1 Network currently 80Gb/s WAN connection based on 10Gb/s move to 100Gb/s based setup early in 2015 combine most existing connections into 100G link keep 2x10Gb/s VPN to CERN challenge of remote data access (ALICE&FAX&AAA) GridKa batch farm behind NAT 15Gb/s available bandwidth often saturated now 2015+ 5 13.10.14 Andreas Petzold KIT Site Report HEPiX 2014 Lincoln, NE

Large Scale Data Facility storage and computing for non-hep sciences 6.7PB disk; IBM SONAS + DDN based GPFS cluster w/ HPSS integration Hadoop cluster bwsync&share in production since Jan 1 st 2014 Dropbox-like service for universities and colleges in state of Baden- Württemberg based on Powerfolder steadily increasing number of users; currently ~8000 bwfilestorage central storage for HPC in Baden-Württemberg Storage for Human Brain Project S3 based on DDN WOS systems Climatology data (ENES) replication using EUDAT B2SAFE service 6 13.10.14 Andreas Petzold KIT Site Report HEPiX 2014 Lincoln, NE

Federated Identity Management bwidm federated IdM for Baden-Württemberg all state wide service integrate bwidm bwsync&share, bwfilestorage, bwunicluster,... Shibboleth and LDAP available extend beyond Baden-Württemberg now working on extension to DFN-AAI (Germany wide) integration of Umbrella 7 13.10.14 Andreas Petzold KIT Site Report HEPiX 2014 Lincoln, NE

Smart Data Innovation Lab federally funded project to enable industry and research partners to process their data on commercially available Big Data platforms communities: Industry 4.0, Energy, Smart Cities, Medicine 40 partners from industry and research project leader SAP www.sdil.de 8 13.10.14 Andreas Petzold KIT Site Report HEPiX 2014 Lincoln, NE

Smart Data Innovation Lab platform operated by SCC SAP Hana 4 machines: 2 IBM X3850, coupled with QPI interconnect, 80 (160 w/ HT) cores, 1TB RAM, 30TB HDD IBM Watson Foundations Hadoop, SPSS Modeler, Watson Content Analytics 7 Power8 SL822 machines, 140 cores, 4TB RAM total, 300TB HDD Software AG Terracotta & Apama soon to come will be integrated into federated IdM 9 13.10.14 Andreas Petzold KIT Site Report HEPiX 2014 Lincoln, NE

HPC at KIT SCC operates HPC systems for KIT and state of Baden- Württemberg IC2 for Institutes of KIT (236 on Top500 2012) 485 nodes, 162 TFLOPS, 32.6TB RAM, 470 TB Lustre bwunicluster for general HPC work in BW 520 nodes, 176 TFLOPS, 41.1TB RAM, 470TB Lustre shared with IC2, 230TB Lustre ForHLR1 for HPC research in BW 530 nodes, 216 TFLOPS, 41.1TB RAM, 470TB Lustre shared with IC2, 230TB Lustre ForHLR2 for HPC research in BW 45ºC water cooling, free cooling installation mid 2015 10 13.10.14 Andreas Petzold KIT Site Report HEPiX 2014 Lincoln, NE

New Building for ForHLR II 11 13.10.14 Andreas Petzold KIT Site Report HEPiX 2014 Lincoln, NE

We are hiring! We are looking for computer scientists and physicists interested in IT Security Large Scale Data Management Large Scale Data Analysis Contact andreas.heiss@kit.edu 12 13.10.14 Andreas Petzold KIT Site Report HEPiX 2014 Lincoln, NE