grid computing in israel
DESCRIPTION
A presentation from the HP-CAST 9 conference, Singapore, May 2008.TRANSCRIPT
Grid Computing and
High-Performance Computing
in Israel
Guy Tel-Zur, Ph.D.The Israeli Association of Grid Technologies
[email protected]://www.Grid.org.il
Topics■ Background
About the country Infrastructure
■ The Academy IAG The Technion, The Hebrew Univ., Ben-Gurion Univ.
■ The Industry IGT
■ The SEPAC Collaboration
Background
This means~500,000PCs
Overall Computing Power: ~500 TFLOPS
Network Infrastructure
ILAN Statistics
2006
2008
Israel
IUCC – The Inter University Computation Center
IUCC
The Academy
The Israel Academic Grid (IAG)
• http://iag.iucc.ac.il/• Funded by the MOST• Steering & Technical
Committees• Coordinates the Israeli activity in
EGEE, EGI and IsraGrid
IUCC is the CA for the IAG
EGEE III
■ May 1st 2008 to April 30th 2010
■ Budget reduced by about 50%
■ Subject to severe FP7 regulations
■ SA1 to be handled through ISRAGRID
Vision of EGI Formation of National Grid Initiatives (NGIs) which unite
efforts within each country, providing a single point of
contact for coordinated efforts
IsraGrid■ A national committee recommended last July to
establish a National Grid Computing Infrastructure.
■ Waiting for final approval by the Government■ To be used only for R&D purposes■ To be used by all the Academic institutes and
the Israeli High-Tech industry■ Secured access■ Managed by the IUCC
MOSIX
A management system targeted for HPC onx86 Linux clusters and multi-clusterorganizational grids
Main features:– supports parallel processes and batch jobs – Automatic resource discovery – Adaptive workload distribution by process process
migrationmigrationOutcome: the grid and each cluster performs like a
single computer with multiple processors
Guest processes can’t modify resources in hosting nodes
The Hebrew University Organizational Grid
• 15 MOSIX clusters ~400 nodes• In life-sciences, medical school, chemistry and computer
science• Applications:
Nano-technology, Molecular dynamics, Protein folding, Genomics (BLAT, BLAST, SW), Meteorological weather forecast (WRF), Navier-Stokes equations and turbulence (CFD) , CPU simulator of new hardware design (SimpleScalar)
More information at http://www.MOSIX.org
Nanco- a cluster for Nanotechnology
TechnionCenter for Computation in Nanotechnology, Russell-Berrie Nanotechnology InstituteTaub Computer Center
64 dual processor dual core compute nodes (total 256 cores), Opteron Rev. F
8GB RAM memory/node
2 master nodes for H/A , also Opterons for redundency
Fast DDR Infiniband Interconnect
Netapp storage
P P PP P PMM M
Infiniband Switch
Operational since summer 2007
Provided by Sun – integrated by EMET and Voltaire
SUN and GNU compilers
Voltaire MPI and OpenMPI for parallelization
Most of codes are MPI codes – either commercial or self developed
More info on http://phycomp.technion.ac.il/~nanco
Grid Computing at the Technion
Israel Institute of Technology
• Distributed Systems Laboratory
• Prof. Assaf Schuster – Head
• Projects:– GMS– Super-Link Online– The Dependable Grid– EGEE– …and more
http://dsl.cs.technion.ac.il/index.html
GMS – Grid Monitoring System
Distributively store all logs of a large batch system in local databases
Apply distributed data mining on logs Implementation using Condor Taken up by Intel NetBatch team: started a $3M
project
SuperLink Online
http://bioinfo.cs.technion.ac.il/superlink-online/
a production portal for geneticists working at hospitals
Submitted tasks contain gene mapping results from lab experiments
Portal user sees a single computer (!)
Implemented using a hierarchy of Condor pools
− Highest/smallest pool in Technion (DSL)
− Lowest/largest in Madison (GLOW). In progress: linkage@home and EGEE BioMed
implementations.
The Dependable Grid Provide a High Availability (HA) Library as a service for
any Grid component
HA for Condor matchmaker with zero loc changes (!!!)
Part of Condor 6.8 distribution
Deployed in many large Condor production pools
Plans to develop and support an open-source distribution
Ben Gurion University of the Negev
• Inter campus Condor pool
• Grid Computing
The BGU Condor Pool
• Started in 2000• Today: ~200 processors• Linux & Windows• Campus-wide project• Non-dedicated resources
We plan to build a new Condor pool installation at the Soroka Medical Center in Beer-Sheva
Grid Computing in the Negev - BGU, NRCN
BGU:• A Certified EGEE-II Production site• A Pre-Production EGEE site
NRCN:A small Condor pool40 processors, Part of the IGT Grid Lab.A member of the SEPAC Grid
Collaboration
Parallel Processing Education
■ Cluster made of Virtual Machines (Xen)
■ “Classic” tools: MPI, OpenMP
■ “Modern” tools: Star-P, Grid Mathematica
■ Grid Computing practice: Condor, Gilda and UNICORE
■ Final projects on a variety of subjects: Parallel Image Processing, Parallel Game of Life, Map/Reduce, Monte Carlo…
Scientific Computing,
Optimization and
Data Analysis
Applied Imaging
Science
Physics and
Engineering
The IDIP GroupThe IDIP Group
• BGU Members belong to variety of departments from the faculties of Engineering, Exact Sciences and the School of Medicine
• Collaborations with various parties in the academia and Industry
• More than 20 research students
Inter-Disciplinary Digital Image Processing
Multi-scale Geometric methods for Filaments detection in 3D
Development of state of the art tools Due to the large typical size of real 3D images and
the high dimensionality of the coefficients space the computational and storage complexity are very high
The Israeli Association of Grid Technologies (IGT)
IGT Members
IGT Work Groups
•Grid-Data Centers & Labs UtilizationPeter Weinstein, IGT Lab Manager
•Grid-SOARonen Yochpaz, CTO VeNotion
•Grid-HPCDr. Guy Tel-Zur, NRCN
•Grid-Application ServerNati Shalom, CTO GigaSpaces
•Grid-RDMAAsaf Somekh, Voltaire
•Grid-VirtualizationNiran Even Chen, BenefIT
IGT WEB SiteKnowledge Sharing and Networking
16,500 Visitors per Month/ 75% from the US
1GigaByte Downloads per Month
Cristophe Bisciglia Creator of Google's Academic Cloud Computing Initiative (ACCI) Senior Software Engineer, Google
Simone Brunozzi Web Services Evangelist,Amazon Web Services
Paul Strong Distinguished Research Scientist, eBay
Dr. Owen O'MalleyOwen O’Malley, Yahoo!Hadoop Architect and Apache VP for Hadoop
IGT2008 – World Summit of Cloud ComputingDecember 1-2, 2008, Hertzelia, Israel
Steve Rubinow, CIO, NYSE Euronext, The largest exchange in the world.
Dr. Yaron WolfsthalSenior Manager, Reliable System TechnologiesIBM Research Lab in Haifa (HRL) IBM and EU Joint Research Initiative for Cloud Computing - RESERVOIR
IGT2008 – World Summit of Cloud ComputingDecember 1-2, 2008, Hertzelia, Israel
Dr. Frank BaetkeGlobal HPC-TechnologyProgram ManagerHPCD Richardson / Munich
The SEPAC Collaboration
SEPAC, The Southern European Partnership for Advanced Computing, is a multi-national Grid-cooperation initiated by major South-European High Performance Computing Centers.
The objective is to build a Grid as a highly reliable application framework based on open interfaces facilitating a consistent and easy-to-use user interface for scientists and researchers in distributed heterogeneous environments.
The SEPAC Grid
SEPAC Future
We are looking for more sites to join ushttp://www.sepac-grid.org
Cloud Computing Layer
http://www.facebook.com/
group.php?gid=8450870046
Thanks to…■ Prof. David Horn, TAU, Head of the IAG
■ Mr. Avner Agom, IGT General Manager
■ Prof. Amnon Barak, CS Dept., HUJI.
■ Mr. Eddie Aharonovich, CS. Dept., TAU
■ Dr. Anne Weill,The Technion
■ Dr. Ofer Levi, The Ben-Gurion Univ.
■ The SEPAC Collaboration
Questions ?
Condor at the BGU:
http://www.ee.bgu.ac.il/~tel-zur/condor/
"An Introduction to Parallel Processing” course at the BGU:
http://www.ee.bgu.ac.il/~tel-zur/teaching/2008B
Grid Computing at the BGU:
http://www.ee.bgu.ac.il/~tel-zur/grid.html
IGT: http://www.grid.org.il
EGI: http://web.eu-egi.eu/
References: