High-speed data ingest and analysis for photon science at DESY with IBM Spectrum Scale.
|
|
- Gyles Brown
- 5 years ago
- Views:
Transcription
1 High-speed data ingest and analysis for photon science at DESY with IBM Spectrum Scale. Martin Gasthuber, Stefan Dietrich, Manuela Kuhn, Uwe Ensslin, Janusz Malka / DESY
2 DESY Where, Why and What > founded December 1959 > Research Topics > 2 Sites Accelerator Research/Build/Operate Particle Physics (HEP) > i.e. Higgs@LHC Photon Science > i.e. X-ray crystallography, looking at viruses to van Gogh and fuel cells Astro Particle Physics Hamburg Hamburg / main site Zeuthen (close to Berlin) > ~2300 employes, >3000 guest scientist anually Zeuthen Martin Gasthuber SC16 15/11/2016 Page 2
3 PETRA III > storage ring accelerator 2.3km circumference > X-ray radiation source > since 2009: 14 beamlines in operation > starting February 2014: shutdown for new extension 10 additional beamlines Sample raw file Beamline P11 Bio-Imaging and diffraction Martin Gasthuber SC16 15/11/2016 Page 3
4 PETRA III Extension > extension for PETRA III > 2 new experiment halls > 10 new beamlines > bigger and faster detectors... > in operational since April 2015 Martin Gasthuber SC16 15/11/2016 Page 4
5 Shooting crystals... Martin Gasthuber SC16 15/11/2016 Page 5
6 Shooting crystals... Martin Gasthuber SC16 15/11/2016 Page 6
7 New Challenges > New detectors achieve higher data rates: Pilatus 300k: 1,2 MB 200 Hz Pilatus 6M: 25 MB 25 Hz 7 MB 100 Hz PCO Edge: 8 MB 100Hz PerkinElmer: 16 MB Byte 15 Hz Lambda: 60 Gb/s, 2000 Hz file rate Eiger: 30 Gb/s, 2000 Hz file rate > Old storage system hit limits > New storage system has to be installed during PETRA III shutdown! Martin Gasthuber SC16 15/11/2016 Page 7
8 Requirements for New Storage System > High performance for single clients > 1GB/s > Handle data peaks (Burst Buffer) Data-acquisition has bursty nature First measurement, change sample, second measurement and so on Duration: minutes, hours, days > Protection between beamlines Competition, data must not be readable from every PC Data must not be readable by next scientific group using the beamline Data-acquisition at one beamline should not interfere with other beamlines Martin Gasthuber SC16 15/11/2016 Page 8
9 Limitations > datacenter is ~1 km away > low space in experiment hall and at beamline local storage is no option > 10 Gigabit Ethernet available (only) > mix of operating systems and ages: Windows Multiple Linux distributions and ages Sometimes unsupported versions > shared accounts for data-acquisition per beamline > very heterogeneous environment: tech, social, requirements > time, personal and money is limited as usual ;-) Martin Gasthuber SC16 15/11/2016 Page 9
10 We are not alone... Martin Gasthuber SC16 15/11/2016 Page 10
11 but time is not on our side. Martin Gasthuber SC16 15/11/2016 Page 11
12 DESY & IBM Collaboration > Collaboration with IBM within scope of SPEED project > Timeline for SPEED project: June > March 2015 > For DESY: Access to experts from development, research and support > 6 people from DESY, 5+ from IBM > Solution based on IBM Spectrum Scale (GPFS) and IBM Elastic Storage Server (ESS) GPFS ESS supports GPFS Native RAID > Initial invest: 1x GSS24 (232x 3TB NLSAS) > Loan (Beta HW): 1x ESS GL4 (232x 4TB NLSAS) 1x ESS GS1 (24x 400GB SSD) > So far excellent performance and stability (more than expected!) ESS GS1 Martin Gasthuber SC16 15/11/2016 Page 12 ESS GL4/GSS24
13 Logical Dataflow From the cradle to the grave Keep copy of data for several days/weeks Keep copy of data for several months Martin Gasthuber SC16 15/11/2016 Page 13
14 Logical Building Blocks Key Components 4x GPFS clusters - Looks like one cluster from the user point of view - Provides separation for admin purposes 2x GPFS file system - Provides isolation for data taking (beamline file system) and long-term analysis (core file system) 2x IBM ESS GS1 - Metadata core file system - Metadata beamline file system - Tier0 beamline file system - Having two ESS improves high availability for data taking 2x IBM ESS GL4 - Tier1 beamline file system - Data core filesystem - Convert GSS24 to ESS GL4 Additional nodes in access cluster and beamline cluster Analysis cluster already exists All nodes run GPFS All nodes are connected via InfiniBand Martin Gasthuber SC16 15/11/2016 Page 14
15 physical Layout Detector prop. Link Proxy Nodes Control Server Detector prop. Link Control Server up to 24 Detectors Beamline Nodes 10GE Access Layer Total: 160 Gb/s (320 max.) P3 DC ~1km RTT: ~0,24 ms Proxy Nodes Proxy Nodes Proxy Nodes Initial 7 Proxy Nodes Infiniband (56 Gbit FDR) 2x Switches 2x GPFS Fabrics Analysis Cluster GS1 NSD ESS GL6 NSD ESS GL4 NSD DESY Network GS1 NSD P3 Hall Datacenter Martin Gasthuber SC16 15/11/2016 Page 15
16 Access Protocols > Proxy nodes export GPFS filesystem via NFS + SMB + ZeroMQ > Beamline NFSv3 (Kernel) SMB, based on Samba 4.2 ZeroMQ > ZeroMQ: Messaging library > Core Available for multiple languages Multiple message patterns available (PUSH/PULL, REQ/REPLY) One-way tunnel from detector to GPFS vacuum cleaner for data ;-) NFSv3 (NFSv4.1/pNFS with Ganesha awaited) SMB, based on Samba 4.2 Native GPFS Martin Gasthuber SC16 15/11/2016 Page 16
17 Beamline Filesystem > Wild-West area for beamline > Only host based authentication, no ACLs > Access through NFSv3, SMB or ZeroMQ > Optimized for performance 1 MiB filesystem blocksize Pre-optimized NFSv3: ~60 MB/s NFSv3: ~600 MB/s SMB: ~ MB/s 'carve out' low capacity & large # of spindles > Tiered Storage Detector prop. Link Control Server Beamline Nodes Access through NFS/SMB/0MQ GS1 NSD Proxy Nodes Migration between GPFS Pools via PolicyRun GL6/GL4 NSD MD Prx. Nodes Native GPFS Beamline Filesystem Tier 0: SSD burst buffer (< 10 TB) Migration after short period of time Tier 1: ~80 TB capacity Martin Gasthuber SC16 15/11/2016 Page 17
18 Core Filesystem > Ordered World > Full user authentication > NFSv4 ACLs (only) > Access through NFSv3, SMB or native GPFS (dominant) > GPFS Policy Runs copy data Beamline Core Filesystem Analysis Nodes Proxy Nodes Beamline Filesystem GS1 NSD MD Migration between GPFS Pools GL6/GL4 NSD GPFS Policy Run creates copy Single UID/GID ACL inheritance gets active Raw data set to immutable > 8 MiB filesystem blocksize > Fileset per beamtime GL6/GL4 NSD Core Filesystem GS1 NSD MD > XATTR used for data flow steering (policy run) Martin Gasthuber SC16 15/11/2016 Page 18
19 Activities since April 2015 Martin Gasthuber SC16 15/11/2016 Page 19
20 European XFEL - a leading new research facility The European XFEL (X-Ray Free-Electron Laser) is a research facility under construction which will use high intensity X-ray light to help scientists better understand the nature of matter. > Location: Schenefeld and Hamburg, Germany > User facility with 280 staff (+ 230 from DESY) > 2017 start of user operation Schenefeld site at the start of user operation Martin Gasthuber SC16 15/11/2016 Page 20
21 EuXFEL - participants > Organized as a non-profit corporation in 2009 with the mission of design, construction, operation, and development of the freeelectron laser > Supported by 11 partner countries > Germany (federal government, city-state of Hamburg, and state of Schleswig-Holstein) covers 58% of the costs; Russia contributes 27%; each of the other international shareholders 1 3% > Total budget for construction (including commissioning) 1.22 billion at 2005 prices 600 M contributed in cash, over 550 M as in-kind contributions (mainly manufacture of parts for the facility) Martin Gasthuber SC16 15/11/2016 Page 21
22 Facility overview Facility overview Schenefeld Osdorfer Born DESY-Bahrenfeld Experiment hall Laboratories Offices Electron beam to photon beamlines Undulator systems begin Electron source Linear accelerator begins Martin Gasthuber SC16 15/11/2016 Page 22
23 from electron to coherent x-ray Martin Gasthuber SC16 15/11/2016 Page 23
24 DAQ Challenges Readout rate driven by bunch structure 10 Hz train of pulses 4.5 MHz pulses in train ( pulses) Data volume driven by detector type Detector type Sampling Data/pulse Data/train Data/sec 1 channel digitizer 5 GS/s ~2 kb ~6 MB ~60 MB 1 Mpxl 2D camera 4.5 MHz ~2 MB ~1 GB ~10 GB 4 Mpxl 2D camera 4.5 MHz ~8 MB ~3 GB ~30 GB* - volume depends on detector type and pulses per train - 1-N trains per file -> 1GB file or larger * Limited by AGIPD detector internal pipeline depth (352 img/sec), hence factor 3 compare to LPD 1MPx Martin Gasthuber SC16 15/11/2016 Page 24
25 How to cope with that? Standardize detector to DAQ interfaces Multiple 10GE network links to receive data from detector Standard data transfer protocols Standard data formats (HDF5) Include software based computing capability into the DAQ chain Data receiving, aggregation, reduction, formatting Enable bad quality data rejection Provide real time overview of collected data e.g. compute statistics, visualize data Provide highly optimized infrastructure and resources for data recording close to the experiment station Dedicated network for DAQ Distributed storage systems with controlled/restricted access HPC systems for demanding storage GPFS on ESS systems Martin Gasthuber SC16 15/11/2016 Page 25
26 DAQ data flow and processing Martin Gasthuber SC16 15/11/2016 Page 26
27 infrastructure locations SASE3 SASE1 SASE2 4 computer rooms in the experiment hall (red, a.k.a. balcony rooms) Dedicated rack rooms for the instruments (orange) Martin Gasthuber SC16 15/11/2016 Page 27
28 data flow more abstract Detector Train Builder PC-Layer (16-64 Nodes) Online Storage Online FS Offline FS Scratch FS InfiniBand MetroX (FDR/FDR-10) Offline Storage Home FS Per SASE, initially 2, finally 3 10GE UDP + Jumbo Frames Online Cluster ~ 4 km Distance ~22µs latency RDMA enabled Scratch FS shared for all SASE Offline Cluster Train Builder Reshuffles picture modules into whole picture Pictures shuffled in trains Sends single trains per channel PC-Layer Data analysis for monitoring Data Reduction, e.g. FPGA based compression Veto File creation in memory and online filesystem every node creates a 1GB HDF5 file every 1.6s Online Cluster nodes Online data analysis and re-calibration Transfer Online Offline Storage Evaluation: multiple or stretched cluster Evaluation: GPFS AFM or custom scripts Offline Storage Shared across experimental stations (SASE) Data arrives after delay, stored on GPFS Copy data to dcache (tape copy, export) ACLs Raw data access only from dcache Offline cluster stores calibrated data on GPFS User analysis from calibrated data Martin Gasthuber SC16 15/11/2016 Page 29
29 even higher altitude looking at rates two beamlines Martin Gasthuber SC16 15/11/2016 Page 30
30 Quality of Service control IOPS spent profile Martin Gasthuber SC16 15/11/2016 Page 33
31 challenges to continue on bandwidth optimization ingest from detector (the 30GBs detector beamline) highest priority offline storage to dcache feed calibration process and copy to tape preserving GPFS NFSv4 ACLs control user access largely non-predictable (QOS) prove fault tolerance site failure, link failure (Ethernet/InfiniBand) all flash for online storage looks economically feasible (0.5 PB per unit) performance figures under investigation should help a lot to get chaotic user access silently merged save inter-tier copy overhead homogeneous event driven data migration (instead/complementing of policy run) cluster wide inotify (L)ight (W)eigth (E)vent Martin Gasthuber SC16 15/11/2016 Page 34
32 EOF Martin Gasthuber SC16 15/11/2016 Page 35
Data Challenges in Photon Science. Manuela Kuhn GridKa School 2016 Karlsruhe, 29th August 2016
Data Challenges in Photon Science Manuela Kuhn GridKa School 2016 Karlsruhe, 29th August 2016 Photon Science > Exploration of tiny samples of nanomaterials > Synchrotrons and free electron lasers generate
More informationAn ESS implementation in a Tier 1 HPC Centre
An ESS implementation in a Tier 1 HPC Centre Maximising Performance - the NeSI Experience José Higino (NeSI Platforms and NIWA, HPC Systems Engineer) Outline What is NeSI? The National Platforms Framework
More informationAn Introduction to GPFS
IBM High Performance Computing July 2006 An Introduction to GPFS gpfsintro072506.doc Page 2 Contents Overview 2 What is GPFS? 3 The file system 3 Application interfaces 4 Performance and scalability 4
More informationStorage for HPC, HPDA and Machine Learning (ML)
for HPC, HPDA and Machine Learning (ML) Frank Kraemer, IBM Systems Architect mailto:kraemerf@de.ibm.com IBM Data Management for Autonomous Driving (AD) significantly increase development efficiency by
More informationLustre architecture for Riccardo Veraldi for the LCLS IT Team
Lustre architecture for LCLS@SLAC Riccardo Veraldi for the LCLS IT Team 2 LCLS Experimental Floor 3 LCLS Parameters 4 LCLS Physics LCLS has already had a significant impact on many areas of science, including:
More informationNext-Generation NVMe-Native Parallel Filesystem for Accelerating HPC Workloads
Next-Generation NVMe-Native Parallel Filesystem for Accelerating HPC Workloads Liran Zvibel CEO, Co-founder WekaIO @liranzvibel 1 WekaIO Matrix: Full-featured and Flexible Public or Private S3 Compatible
More informationA COMMON SOFTWARE FRAMEWORK FOR FEL DATA ACQUISITION AND EXPERIMENT MANAGEMENT AT FERMI
A COMMON SOFTWARE FRAMEWORK FOR FEL DATA ACQUISITION AND EXPERIMENT MANAGEMENT AT FERMI R. Borghes, V. Chenda, A. Curri, G. Kourousias, M. Lonza, G. Passos, M. Prica, R. Pugliese 1 FERMI overview FERMI
More informationExperiment Control Upgrades at DESY
Experiment Control Upgrades at DESY Teresa Núñez DESY Photon Science PiLC Logic Controller ADQ412 Digitizer Diffractometer in Sardana GPFS storage system Tango Meeting ONERA, 21-06-16 PiLC Logic Controller
More informationScientific Data Policy of European X-Ray Free-Electron Laser Facility GmbH
6 June 2017 Scientific Data Policy of European X-Ray Free-Electron Laser Facility GmbH as approved by the Council on 30 June 2017 1 Preface... 2 2 Definitions... 2 3 General principles... 5 4 Raw data
More informationOnline Data Analysis at European XFEL
Online Data Analysis at European XFEL Hans Fangohr Control and Analysis Software Group Senior Data Analysis Scientist DESY, 25 January 2018 2 Outline Introduction & European XFEL status Overview online
More informationDESY site report. HEPiX Spring 2016 at DESY. Yves Kemp, Peter van der Reest. Zeuthen,
DESY site report HEPiX Spring 2016 at DESY Yves Kemp, Peter van der Reest Zeuthen, 2016-04-18 Accelerators news > XFEL: 1.3.2016: All segments of first light-generating system installed in European XFEL
More informationGPFS Experiences from the Argonne Leadership Computing Facility (ALCF) William (Bill) E. Allcock ALCF Director of Operations
GPFS Experiences from the Argonne Leadership Computing Facility (ALCF) William (Bill) E. Allcock ALCF Director of Operations Argonne National Laboratory Argonne National Laboratory is located on 1,500
More informationControl, data acquisition, management and analysis
Control, data acquisition, management and analysis Thomas M. Baumann Scientific Instrument SQS Hans Fangohr Control & Analysis Software Krzysztof Wrona IT & Data Management SQS Early User Workshop Schenefeld,
More informationConference The Data Challenges of the LHC. Reda Tafirout, TRIUMF
Conference 2017 The Data Challenges of the LHC Reda Tafirout, TRIUMF Outline LHC Science goals, tools and data Worldwide LHC Computing Grid Collaboration & Scale Key challenges Networking ATLAS experiment
More informationNexus Builder Developing a Graphical User Interface to create NeXus files
Nexus Builder Developing a Graphical User Interface to create NeXus files Lilit Grigoryan, Yerevan State University, Armenia September 9, 2014 Abstract This report describes a project which main purpose
More informationData Movement & Tiering with DMF 7
Data Movement & Tiering with DMF 7 Kirill Malkin Director of Engineering April 2019 Why Move or Tier Data? We wish we could keep everything in DRAM, but It s volatile It s expensive Data in Memory 2 Why
More informationEinführung von VoIP am DESY. Kars Ohrenberg IT
Einführung von VoIP am DESY Kars Ohrenberg IT Overview DESY The IP Network @ DESY Telephony @ DESY VoIP Installation, Design and Experience Summary 13th, 2008 2 Deutsches Elektronen- Synchrotron DESY National
More informationThe creation of a Tier-1 Data Center for the ALICE experiment in the UNAM. Lukas Nellen ICN-UNAM
The creation of a Tier-1 Data Center for the ALICE experiment in the UNAM Lukas Nellen ICN-UNAM lukas@nucleares.unam.mx 3rd BigData BigNetworks Conference Puerto Vallarta April 23, 2015 Who Am I? ALICE
More informationAccelerating Spectrum Scale with a Intelligent IO Manager
Accelerating Spectrum Scale with a Intelligent IO Manager Ray Coetzee Pre-Sales Architect Seagate Systems Group, HPC 2017 Seagate, Inc. All Rights Reserved. 1 ClusterStor: Lustre, Spectrum Scale and Object
More informationThe Leading Parallel Cluster File System
The Leading Parallel Cluster File System www.thinkparq.com www.beegfs.io ABOUT BEEGFS What is BeeGFS BeeGFS (formerly FhGFS) is the leading parallel cluster file system, developed with a strong focus on
More informationThe CMS Computing Model
The CMS Computing Model Dorian Kcira California Institute of Technology SuperComputing 2009 November 14-20 2009, Portland, OR CERN s Large Hadron Collider 5000+ Physicists/Engineers 300+ Institutes 70+
More informationStorage and I/O requirements of the LHC experiments
Storage and I/O requirements of the LHC experiments Sverre Jarp CERN openlab, IT Dept where the Web was born 22 June 2006 OpenFabrics Workshop, Paris 1 Briefly about CERN 22 June 2006 OpenFabrics Workshop,
More informationComputing and Networking at Diamond Light Source. Mark Heron Head of Control Systems
Computing and Networking at Diamond Light Source Mark Heron Head of Control Systems Harwell Science and Innovation Campus ISIS (Spallation Neutron Source) Central Laser Facility LHC Tier 1 computing Research
More informationEmerging Technologies for HPC Storage
Emerging Technologies for HPC Storage Dr. Wolfgang Mertz CTO EMEA Unstructured Data Solutions June 2018 The very definition of HPC is expanding Blazing Fast Speed Accessibility and flexibility 2 Traditional
More informationIBM Spectrum Scale IO performance
IBM Spectrum Scale 5.0.0 IO performance Silverton Consulting, Inc. StorInt Briefing 2 Introduction High-performance computing (HPC) and scientific computing are in a constant state of transition. Artificial
More informationAnalyzing the High Performance Parallel I/O on LRZ HPC systems. Sandra Méndez. HPC Group, LRZ. June 23, 2016
Analyzing the High Performance Parallel I/O on LRZ HPC systems Sandra Méndez. HPC Group, LRZ. June 23, 2016 Outline SuperMUC supercomputer User Projects Monitoring Tool I/O Software Stack I/O Analysis
More informationDAQ system at SACLA and future plan for SPring-8-II
DAQ system at SACLA and future plan for SPring-8-II Takaki Hatsui T. Kameshima, Nakajima T. Abe, T. Sugimoto Y. Joti, M.Yamaga RIKEN SPring-8 Center IFDEPS 1 Evolution of Computing infrastructure from
More informationEvent-Synchronized Data Acquisition System of 5 Giga-bps Data Rate for User Experiment at the XFEL Facility, SACLA
Event-Synchronized Data Acquisition System of 5 Giga-bps Data Rate for User Experiment at the XFEL Facility, SACLA Mitsuhiro YAMAGA JASRI Oct.11, 2011 @ICALEPCS2011 Contents: Introduction Data Acquisition
More informationComputer Science Section. Computational and Information Systems Laboratory National Center for Atmospheric Research
Computer Science Section Computational and Information Systems Laboratory National Center for Atmospheric Research My work in the context of TDD/CSS/ReSET Polynya new research computing environment Polynya
More informationData Acquisition. Amedeo Perazzo. SLAC, June 9 th 2009 FAC Review. Photon Controls and Data Systems (PCDS) Group. Amedeo Perazzo
Data Acquisition Photon Controls and Data Systems (PCDS) Group SLAC, June 9 th 2009 FAC Review 1 Data System Architecture Detector specific Photon Control Data Systems (PCDS) L1: Acquisition Beam Line
More informationFEL diagnostics and control system
FEL diagnostics and control system Thomas M. Baumann WP-85, Scientific Instrument SQS Instrument Scientist Satellite meeting Soft X-ray instruments SQS and SCS Hamburg, 24.01.2017 2 Outline FEL diagnostics
More informationSERGEY STEPANOV. Argonne National Laboratory, Advanced Photon Source, Lemont, IL, USA. October 2017, ICALEPCS-2017, Barcelona, Spain
BEAMLINE AND EXPERIMENT AUTOMATIONS FOR THE GENERAL MEDICAL SCIENCES AND CANCER INSTITUTES STRUCTURAL BIOLOGY FACILITY AT THE ADVANCED PHOTON SOURCE (GM/CA@APS) SERGEY STEPANOV Argonne National Laboratory,
More informationIBM Spectrum NAS, IBM Spectrum Scale and IBM Cloud Object Storage
IBM Spectrum NAS, IBM Spectrum Scale and IBM Cloud Object Storage Silverton Consulting, Inc. StorInt Briefing 2017 SILVERTON CONSULTING, INC. ALL RIGHTS RESERVED Page 2 Introduction Unstructured data has
More informationData Management. Parallel Filesystems. Dr David Henty HPC Training and Support
Data Management Dr David Henty HPC Training and Support d.henty@epcc.ed.ac.uk +44 131 650 5960 Overview Lecture will cover Why is IO difficult Why is parallel IO even worse Lustre GPFS Performance on ARCHER
More informationNew Storage Technologies First Impressions: SanDisk IF150 & Intel Omni-Path. Brian Marshall GPFS UG - SC16 November 13, 2016
New Storage Technologies First Impressions: SanDisk IF150 & Intel Omni-Path Brian Marshall GPFS UG - SC16 November 13, 2016 Presenter Background Brian Marshall Computational Scientist at Virginia Tech
More informationSSRS-4 and the CREMLIN follow up project
SSRS-4 and the CREMLIN follow up project Towards elaborating a plan for the future collaboration Martin Sandhop SSRS-4 and the CREMLIN follow up project www.cremlin.eu CREMLIN WP5 Workshop: "Towards a
More informationStore Process Analyze Collaborate Archive Cloud The HPC Storage Leader Invent Discover Compete
Store Process Analyze Collaborate Archive Cloud The HPC Storage Leader Invent Discover Compete 1 DDN Who We Are 2 We Design, Deploy and Optimize Storage Systems Which Solve HPC, Big Data and Cloud Business
More informationHardware Aspects, Modularity and Integration of an Event Mode Data Acquisition and Instrument Control for the European Spallation Source (ESS)
Hardware Aspects, Modularity and Integration of an Event Mode Data Acquisition and Instrument Control for the European Spallation Source (ESS) T Gahl 1,5, M Hagen 1, R Hall-Wilton 1,2, S Kolya 1, M Koennecke
More informationParallel Storage Systems for Large-Scale Machines
Parallel Storage Systems for Large-Scale Machines Doctoral Showcase Christos FILIPPIDIS (cfjs@outlook.com) Department of Informatics and Telecommunications, National and Kapodistrian University of Athens
More informationBlock Storage Service: Status and Performance
Block Storage Service: Status and Performance Dan van der Ster, IT-DSS, 6 June 2014 Summary This memo summarizes the current status of the Ceph block storage service as it is used for OpenStack Cinder
More informationIBM Spectrum Scale vs EMC Isilon for IBM Spectrum Protect Workloads
89 Fifth Avenue, 7th Floor New York, NY 10003 www.theedison.com @EdisonGroupInc 212.367.7400 IBM Spectrum Scale vs EMC Isilon for IBM Spectrum Protect Workloads A Competitive Test and Evaluation Report
More informationXcellis Technical Overview: A deep dive into the latest hardware designed for StorNext 5
TECHNOLOGY BRIEF Xcellis Technical Overview: A deep dive into the latest hardware designed for StorNext 5 ABSTRACT Xcellis represents the culmination of over 15 years of file system and data management
More information<Insert Picture Here> Oracle Storage
Oracle Storage Jennifer Feng Principal Product Manager IT Challenges Have Not Slowed Increasing Demand for Storage Capacity and Performance 2010 New Digital Data ( Replicated (¼ Created,
More informationSpanish Tier-2. Francisco Matorras (IFCA) Nicanor Colino (CIEMAT) F. Matorras N.Colino, Spain CMS T2,.6 March 2008"
Spanish Tier-2 Francisco Matorras (IFCA) Nicanor Colino (CIEMAT) Introduction Report here the status of the federated T2 for CMS basically corresponding to the budget 2006-2007 concentrate on last year
More informationTINE Video System. A Modular, Well-Defined, Component-Based and Interoperable TV System. Proceedings On Redesign VSv3
TINE Video System A Modular, Well-Defined, Component-Based and Interoperable TV System Proceedings On Redesign VSv3 Stefan Weisse, David Melkumyan, Philip Duval DESY, Germany Where That Comes From: PITZ
More informationChallenges in Storage
Challenges in Storage or : A random walk Presented by Patrick Fuhrmann with contributions from many experts. Contributions and thoughts provided by Oxana Smirnova, NeIC, Lund Markus Schulz, CERN IT Steven
More informationParallel File Systems. John White Lawrence Berkeley National Lab
Parallel File Systems John White Lawrence Berkeley National Lab Topics Defining a File System Our Specific Case for File Systems Parallel File Systems A Survey of Current Parallel File Systems Implementation
More informationDetectors for Future Light Sources. Gerhard Grübel Deutsches Elektronen Synchrotron (DESY) Notke-Str. 85, Hamburg
Detectors for Future Light Sources Gerhard Grübel Deutsches Elektronen Synchrotron (DESY) Notke-Str. 85, 22607 Hamburg Overview Radiation from X-Ray Free Electron lasers (XFEL, LCLS) Ultrafast detectors
More informationFeedback on BeeGFS. A Parallel File System for High Performance Computing
Feedback on BeeGFS A Parallel File System for High Performance Computing Philippe Dos Santos et Georges Raseev FR 2764 Fédération de Recherche LUmière MATière December 13 2016 LOGO CNRS LOGO IO December
More informationDesigning Next Generation FS for NVMe and NVMe-oF
Designing Next Generation FS for NVMe and NVMe-oF Liran Zvibel CTO, Co-founder Weka.IO @liranzvibel Santa Clara, CA 1 Designing Next Generation FS for NVMe and NVMe-oF Liran Zvibel CTO, Co-founder Weka.IO
More informationHTCondor with KRB/AFS Setup and first experiences on the DESY interactive batch farm
HTCondor with KRB/AFS Setup and first experiences on the DESY interactive batch farm Beyer Christoph & Finnern Thomas Madison (Wisconsin), May 2018 HTCondor week The Team and the Outline The Team Outline
More information10GE network tests with UDP. Janusz Szuba European XFEL
10GE network tests with UDP Janusz Szuba European XFEL Outline 2 Overview of initial DAQ architecture Slice test hardware specification Initial networking test results DAQ software UDP tests Summary 10GE
More informationTuning I/O Performance for Data Intensive Computing. Nicholas J. Wright. lbl.gov
Tuning I/O Performance for Data Intensive Computing. Nicholas J. Wright njwright @ lbl.gov NERSC- National Energy Research Scientific Computing Center Mission: Accelerate the pace of scientific discovery
More informationImproved Solutions for I/O Provisioning and Application Acceleration
1 Improved Solutions for I/O Provisioning and Application Acceleration August 11, 2015 Jeff Sisilli Sr. Director Product Marketing jsisilli@ddn.com 2 Why Burst Buffer? The Supercomputing Tug-of-War A supercomputer
More informationMonte Carlo Production on the Grid by the H1 Collaboration
Journal of Physics: Conference Series Monte Carlo Production on the Grid by the H1 Collaboration To cite this article: E Bystritskaya et al 2012 J. Phys.: Conf. Ser. 396 032067 Recent citations - Monitoring
More informationBenoit DELAUNAY Benoit DELAUNAY 1
Benoit DELAUNAY 20091023 Benoit DELAUNAY 1 CC-IN2P3 provides computing and storage for the 4 LHC experiments and many others (astro particles...) A long history of service sharing between experiments Some
More informationStefan Koestner on behalf of the LHCb Online Group ( IEEE - Nuclear Science Symposium San Diego, Oct.
Stefan Koestner on behalf of the LHCb Online Group (email: Stefan.Koestner@cern.ch) IEEE - Nuclear Science Symposium San Diego, Oct. 31 st 2006 Dedicated to B-physics : single arm forward spectrometer
More informationDetector Control LHC
Detector Control Systems @ LHC Matthias Richter Department of Physics, University of Oslo IRTG Lecture week Autumn 2012 Oct 18 2012 M. Richter (UiO) DCS @ LHC Oct 09 2012 1 / 39 Detectors in High Energy
More informationExperiences using a multi-tiered GPFS file system at Mount Sinai. Bhupender Thakur Patricia Kovatch Francesca Tartagliogne Dansha Jiang
Experiences using a multi-tiered GPFS file system at Mount Sinai Bhupender Thakur Patricia Kovatch Francesca Tartagliogne Dansha Jiang Outline 1. Storage summary 2. Planning and Migration 3. Challenges
More informationvsan Mixed Workloads First Published On: Last Updated On:
First Published On: 03-05-2018 Last Updated On: 03-05-2018 1 1. Mixed Workloads on HCI 1.1.Solution Overview Table of Contents 2 1. Mixed Workloads on HCI 3 1.1 Solution Overview Eliminate the Complexity
More informationAdvanced Photon Source Data Management. S. Veseli, N. Schwarz, C. Schmitz (SDM/XSD) R. Sersted, D. Wallis (IT/AES)
Advanced Photon Source Data Management S. Veseli, N. Schwarz, C. Schmitz (SDM/XSD) R. Sersted, D. Wallis (IT/AES) APS Data Management - Globus World 2018 Growing Beamline Data Needs X-ray detector capabilities
More informationIBM V7000 Unified R1.4.2 Asynchronous Replication Performance Reference Guide
V7 Unified Asynchronous Replication Performance Reference Guide IBM V7 Unified R1.4.2 Asynchronous Replication Performance Reference Guide Document Version 1. SONAS / V7 Unified Asynchronous Replication
More informationThe RAMDISK Storage Accelerator
The RAMDISK Storage Accelerator A Method of Accelerating I/O Performance on HPC Systems Using RAMDISKs Tim Wickberg, Christopher D. Carothers wickbt@rpi.edu, chrisc@cs.rpi.edu Rensselaer Polytechnic Institute
More informationComissioning of the IR beamline at FLASH. DESY.DE
Comissioning of the IR beamline at FLASH Michael.Gensch @ DESY.DE IR Undulator Beamline at FLASH First light 2 FLASH experimental hall before installation of IR beamline BL1 PG2 BL2 BL3 fs Laser 3 FLASH
More informationGPFS for Life Sciences at NERSC
GPFS for Life Sciences at NERSC A NERSC & JGI collaborative effort Jason Hick, Rei Lee, Ravi Cheema, and Kjiersten Fagnan GPFS User Group meeting May 20, 2015-1 - Overview of Bioinformatics - 2 - A High-level
More informationNFS around the world Tigran Mkrtchyan for dcache Team dcache User Workshop, Umeå, Sweden
NFS around the world Tigran Mkrtchyan for dcache Team dcache User Workshop, Umeå, Sweden The NFS community History v1 1984, SUN Microsystems intern 16 ops, 1:1 mapping to vfs 1986 First Connectathon! v2
More informationParallel File Systems for HPC
Introduction to Scuola Internazionale Superiore di Studi Avanzati Trieste November 2008 Advanced School in High Performance and Grid Computing Outline 1 The Need for 2 The File System 3 Cluster & A typical
More informationSummary of the LHC Computing Review
Summary of the LHC Computing Review http://lhc-computing-review-public.web.cern.ch John Harvey CERN/EP May 10 th, 2001 LHCb Collaboration Meeting The Scale Data taking rate : 50,100, 200 Hz (ALICE, ATLAS-CMS,
More informationCRFS: A Lightweight User-Level Filesystem for Generic Checkpoint/Restart
CRFS: A Lightweight User-Level Filesystem for Generic Checkpoint/Restart Xiangyong Ouyang, Raghunath Rajachandrasekar, Xavier Besseron, Hao Wang, Jian Huang, Dhabaleswar K. Panda Department of Computer
More informationCSD3 The Cambridge Service for Data Driven Discovery. A New National HPC Service for Data Intensive science
CSD3 The Cambridge Service for Data Driven Discovery A New National HPC Service for Data Intensive science Dr Paul Calleja Director of Research Computing University of Cambridge Problem statement Today
More informationMellanox InfiniBand Solutions Accelerate Oracle s Data Center and Cloud Solutions
Mellanox InfiniBand Solutions Accelerate Oracle s Data Center and Cloud Solutions Providing Superior Server and Storage Performance, Efficiency and Return on Investment As Announced and Demonstrated at
More informationOpenStack SwiftOnFile: User Identity for Cross Protocol Access Demystified Dean Hildebrand, Sasikanth Eda Sandeep Patil, Bill Owen IBM
OpenStack SwiftOnFile: User Identity for Cross Protocol Access Demystified Dean Hildebrand, Sasikanth Eda Sandeep Patil, Bill Owen IBM 2015 Storage Developer Conference. Insert Your Company Name. All Rights
More informationHIGH-PERFORMANCE STORAGE FOR DISCOVERY THAT SOARS
HIGH-PERFORMANCE STORAGE FOR DISCOVERY THAT SOARS OVERVIEW When storage demands and budget constraints collide, discovery suffers. And it s a growing problem. Driven by ever-increasing performance and
More informationSAP HANA IBM x3850 X6
SAP HANA IBM x3850 X6 Miklos Farkas SAP HANA IBM x3850 X6 IBM Workload Optimized Solution for SAP HANA appliance Applications Data Center Ready SUSE SAP HANA GPFS FPO functionality OS SUSE Linux Enterprise
More informationDDN s Vision for the Future of Lustre LUG2015 Robert Triendl
DDN s Vision for the Future of Lustre LUG2015 Robert Triendl 3 Topics 1. The Changing Markets for Lustre 2. A Vision for Lustre that isn t Exascale 3. Building Lustre for the Future 4. Peak vs. Operational
More informationUpdate on PRad GEMs, Readout Electronics & DAQ
Update on PRad GEMs, Readout Electronics & DAQ Kondo Gnanvo University of Virginia, Charlottesville, VA Outline PRad GEMs update Upgrade of SRS electronics Integration into JLab DAQ system Cosmic tests
More informationSC Series: Performance Best Practices. Brad Spratt Performance Engineering Midrange & Entry Solutions
SC Series: Performance Best Practices Brad Spratt Performance Engineering Midrange & Entry Solutions What s New with the SC Series New Dell EMC SC5020 Hybrid Array Optimized for Economics and Efficiency
More informationRed Hat Gluster Storage performance. Manoj Pillai and Ben England Performance Engineering June 25, 2015
Red Hat Gluster Storage performance Manoj Pillai and Ben England Performance Engineering June 25, 2015 RDMA Erasure Coding NFS-Ganesha New or improved features (in last year) Snapshots SSD support Erasure
More informationProof of Concept TRANSPARENT CLOUD TIERING WITH IBM SPECTRUM SCALE
Proof of Concept TRANSPARENT CLOUD TIERING WITH IBM SPECTRUM SCALE ATS Innovation Center, Malvern PA Joshua Kwedar The ATS Group October November 2017 INTRODUCTION With the release of IBM Spectrum Scale
More informationStorage Evaluations at BNL
Storage Evaluations at BNL HEPiX at DESY Spring 2007 Robert Petkus RHIC/USATLAS Computing Facility Brookhaven National Laboratory State of Affairs Explosive disk storage growth trajectory over the next
More information2008 JINST 3 S Online System. Chapter System decomposition and architecture. 8.2 Data Acquisition System
Chapter 8 Online System The task of the Online system is to ensure the transfer of data from the front-end electronics to permanent storage under known and controlled conditions. This includes not only
More informationShared Parallel Filesystems in Heterogeneous Linux Multi-Cluster Environments
LCI HPC Revolution 2005 26 April 2005 Shared Parallel Filesystems in Heterogeneous Linux Multi-Cluster Environments Matthew Woitaszek matthew.woitaszek@colorado.edu Collaborators Organizations National
More informationDatabase Services at CERN with Oracle 10g RAC and ASM on Commodity HW
Database Services at CERN with Oracle 10g RAC and ASM on Commodity HW UKOUG RAC SIG Meeting London, October 24 th, 2006 Luca Canali, CERN IT CH-1211 LCGenève 23 Outline Oracle at CERN Architecture of CERN
More informationGPFS on a Cray XT. Shane Canon Data Systems Group Leader Lawrence Berkeley National Laboratory CUG 2009 Atlanta, GA May 4, 2009
GPFS on a Cray XT Shane Canon Data Systems Group Leader Lawrence Berkeley National Laboratory CUG 2009 Atlanta, GA May 4, 2009 Outline NERSC Global File System GPFS Overview Comparison of Lustre and GPFS
More informationAUTOMATING IBM SPECTRUM SCALE CLUSTER BUILDS IN AWS PROOF OF CONCEPT
AUTOMATING IBM SPECTRUM SCALE CLUSTER BUILDS IN AWS PROOF OF CONCEPT By Joshua Kwedar Sr. Systems Engineer By Steve Horan Cloud Architect ATS Innovation Center, Malvern, PA Dates: Oct December 2017 INTRODUCTION
More informationHall D and IT. at Internal Review of IT in the 12 GeV Era. Mark M. Ito. May 20, Hall D. Hall D and IT. M. Ito. Introduction.
at Internal Review of IT in the 12 GeV Era Mark Hall D May 20, 2011 Hall D in a Nutshell search for exotic mesons in the 1.5 to 2.0 GeV region 12 GeV electron beam coherent bremsstrahlung photon beam coherent
More informationExtraordinary HPC file system solutions at KIT
Extraordinary HPC file system solutions at KIT Roland Laifer STEINBUCH CENTRE FOR COMPUTING - SCC KIT University of the State Roland of Baden-Württemberg Laifer Lustre and tools for ldiskfs investigation
More informationSystem Management and Infrastructure
System Management and Infrastructure WP 28 Accelerator Controls Conceptual Design Report CDR Meeting December, 14 th 2009, DESY - Tim Wilksen System Management 2 Hardware Management Management xtca Systems
More informationPaNSIG (PaNData), and the interactions between SB-IG and PaNSIG
PaNSIG (PaNData), and the interactions between SB-IG and PaNSIG Erica Yang erica.yang@stfc.ac.uk Scientific Computing Department STFC Rutherford Appleton Laboratory Structural Biology IG 27 March 2014
More informationGrid Computing a new tool for science
Grid Computing a new tool for science CERN, the European Organization for Nuclear Research Dr. Wolfgang von Rüden Wolfgang von Rüden, CERN, IT Department Grid Computing July 2006 CERN stands for over 50
More informationIME (Infinite Memory Engine) Extreme Application Acceleration & Highly Efficient I/O Provisioning
IME (Infinite Memory Engine) Extreme Application Acceleration & Highly Efficient I/O Provisioning September 22 nd 2015 Tommaso Cecchi 2 What is IME? This breakthrough, software defined storage application
More informationMission-Critical Lustre at Santos. Adam Fox, Lustre User Group 2016
Mission-Critical Lustre at Santos Adam Fox, Lustre User Group 2016 About Santos One of the leading oil and gas producers in APAC Founded in 1954 South Australia Northern Territory Oil Search Cooper Basin
More informationGlobal Collaboration on Accelerator Operations and Experiments
Global Collaboration on Accelerator Operations and Experiments Globalization in the Financial World Has a bad taste. Socializing risk? Privatizing win? in the HEP Community Is key to build the next big
More informationPocket: Elastic Ephemeral Storage for Serverless Analytics
Pocket: Elastic Ephemeral Storage for Serverless Analytics Ana Klimovic*, Yawen Wang*, Patrick Stuedi +, Animesh Trivedi +, Jonas Pfefferle +, Christos Kozyrakis* *Stanford University, + IBM Research 1
More informationImproving Packet Processing Performance of a Memory- Bounded Application
Improving Packet Processing Performance of a Memory- Bounded Application Jörn Schumacher CERN / University of Paderborn, Germany jorn.schumacher@cern.ch On behalf of the ATLAS FELIX Developer Team LHCb
More informationUniversity at Buffalo Center for Computational Research
University at Buffalo Center for Computational Research The following is a short and long description of CCR Facilities for use in proposals, reports, and presentations. If desired, a letter of support
More informationExtremely Fast Distributed Storage for Cloud Service Providers
Solution brief Intel Storage Builders StorPool Storage Intel SSD DC S3510 Series Intel Xeon Processor E3 and E5 Families Intel Ethernet Converged Network Adapter X710 Family Extremely Fast Distributed
More informationData storage services at KEK/CRC -- status and plan
Data storage services at KEK/CRC -- status and plan KEK/CRC Hiroyuki Matsunaga Most of the slides are prepared by Koichi Murakami and Go Iwai KEKCC System Overview KEKCC (Central Computing System) The
More informationMAHA. - Supercomputing System for Bioinformatics
MAHA - Supercomputing System for Bioinformatics - 2013.01.29 Outline 1. MAHA HW 2. MAHA SW 3. MAHA Storage System 2 ETRI HPC R&D Area - Overview Research area Computing HW MAHA System HW - Rpeak : 0.3
More informationMID: Materials Imaging and Dynamics Instrument
MID: Materials Imaging and Dynamics Instrument A. Madsen 1,*, J. Hallmann 1, T. Roth 1, G. Ansaldi 1, W. Lu 1,2 1 European XFEL 2 Technische Universität Berlin * anders.madsen@xfel.eu XFEL User Meeting
More information