Grid computing using Alina Bejan University of Chicago.
-
Upload
elfrieda-riley -
Category
Documents
-
view
220 -
download
0
Transcript of Grid computing using Alina Bejan University of Chicago.
Open Science Grid (OSG)
• takes High Throughput Computing to the next level, to transform data-intensive science through a cross-domain, self-managed nationally distributed cyber-infrastructure.
• brings together campuses and communities, and facilitates the needs of Virtual Organizations at all scales.
• The OSG Consortium includes– universities– national laboratories– scientific collaborations – software developers
working together to meet these goals
What is a grid?
• Grid is a system that:– coordinates resources that are not subject to
centralized control,– using standard, open, general-purpose protocols
and interfaces,– to deliver nontrivial qualities of service (based on Ian Foster’s definition in http://www.gridtoday
.com/02/0722/100136.html)
Grids consist of distributed clusters
Grid Client
Application& User Interface
Grid ClientMiddleware
Resource,Workflow
& Data Catalogs
4
Grid Site 2:Sao Paolo
GridService
Middleware
ComputeCluster
GridStorage
Grid
Protocols
Grid Site 1:Fermilab
GridService
Middleware
ComputeCluster
GridStorage
…Grid Site N:UWisconsin
GridService
Middleware
ComputeCluster
GridStorage
• Do you have a project that takes too long when running on a single processor ?
• Do you deal with large amounts of data from simulations or experiments ?
Scaling up Science:Citation Network Analysis in Sociology
2002
1975
1990
1985
1980
2000
1995
Work of James Evans, University of Chicago,
Department of Sociology
6
Scaling up the analysis• Query and analysis of 25+ million citations• Work started on desktop workstations• Queries grew to month-long duration• With data distributed across
U of Chicago TeraPort cluster:– 50 (faster) CPUs gave 100 X speedup– Many more methods and hypotheses can be
tested!
• Higher throughput and capacity enables deeper analysis and broader community access.
7
Mining Seismic data for hazard analysis (Southern Calif. Earthquake Center).
InSAR Image of theHector Mine Earthquake
• A satellitegeneratedInterferometricSynthetic Radar(InSAR) image ofthe 1999 HectorMine earthquake.
• Shows thedisplacement fieldin the direction ofradar imaging
• Each fringe (e.g.,from red to red)corresponds to afew centimeters ofdisplacement.
SeismicHazardModel
Seismicity Paleoseismology Local site effects Geologic structure
Faults
Stresstransfer
Crustal motion Crustal deformation Seismic velocity structure
Rupturedynamics 88
Grids work like a CHARMM for molecular dynamics
• Understanding the mathematics of molecular movement helps researchers simulate slices of the atomic world
• But when accurate nanosecond simulations pose a serious challenge, how can you simulate full microseconds of complex molecular dynamics?
Designing Proteins from Scratch
• Scientists use OSG to design proteins that adopt specific 3D structures and more ambitiously bind and regulate target proteins important in cell biology and pathogenesis
Genetics
• Grid computing is helping microbiologists solve the mysteries of mapping new genomes using GADU (Genome Analysis and Database Update)
Genome Analysis and Database Update (GADU)
• Runs across OSG and TeraGrid. Uses the Virtual Data System (VDS) workflow & provenance.
• Pass through public DNA and protein databases for new and newly updated genomes of different organisms and runs BLAST, Blocks, Chisel. 1200 users of resulting DB.
• Request: 1000 CPUs for 1-2 weeks. Once a month, every month. On OSG at the moment >600CPUs and 17,000 jobs a week.
Stormy weather: grid computing powers fine-scale climate modeling
• Why run individual models when you can run models in combination?
• When it comes to climate modeling, meteorologists are showing 16 forecasts are better than one.
Which sciences can benefit ?
• particle and nuclear physics• astrophysics• bioinformatics• gravitational-wave science• computer science• mathematics• medical imaging • nanotechnology• potentially any other science …
• Research Participation
Majority from physics : Tevatron, LHC, STAR, LIGO. Used by 10 other (small) research groups. 90 members, 30 VOs,
Contributors:- 80 sites / 50 organizations 5 DOE Labs : BNL, Fermilab, NERSC, ORNL, SLAC. 65 Universities. 5 partner campus/regional grids.
Accessible resources: 43,000+ cores 6 Petabytes disk cache 10 Petabytes tape stores 14 internetwork partnership
Usage 15,000 CPU WallClock days/day 1 Petabyte data distributed/month. 100,000 application jobs/day. 20% cycles through resource sharing, opportunistic usage.
• Research Participation
Majority from physics : Tevatron, LHC, STAR, LIGO. Used by 10 other (small) research groups. 90 members, 30 VOs,
Contributors:- 80 sites / 50 organizations 5 DOE Labs : BNL, Fermilab, NERSC, ORNL, SLAC. 65 Universities. 5 partner campus/regional grids.
Accessible resources: 43,000+ cores 6 Petabytes disk cache 10 Petabytes tape stores 14 internetwork partnership
Usage 15,000 CPU WallClock days/day 1 Petabyte data distributed/month. 100,000 application jobs/day. 20% cycles through resource sharing, opportunistic usage.
OSG
• We are successfully operating a large, shared, very distributed system for many users.
• The size & capabilities continue to increase:– software functionality– Users– Sites– partners– cycles used– data storage capabilities.
The OSG project research
Has two components:– To enable scientific discovery by providing a state
of the art production distributed infrastructure for science
– To advance the state of the art in distributed computing through experimental computer science through a large scale production quality distributed system
OSG’s grids
• Grids can be– Campus, Community, Regional, National,
International
• OSG scope includes bridging, and interfacing between them
Grid Resources in the US
• Research Participation
Majority from physics : Tevatron, LHC, STAR, LIGO.
Used by 10 other (small) research groups. 90 members, 30 VOs,
Contributors: 5 DOE Labs
BNL, Fermilab, NERSC, ORNL, SLAC. 65 Universities. 5 partner campus/regional grids.
Accessible resources: 43,000+ cores 6 Petabytes disk cache 10 Petabytes tape stores 14 internetwork partnership
Usage 15,000 CPU WallClock days/day 1 Petabyte data distributed/month. 100,000 application jobs/day. 20% cycles through resource sharing,
opportunistic use.
• Research Participation
Majority from physics : Tevatron, LHC, STAR, LIGO.
Used by 10 other (small) research groups. 90 members, 30 VOs,
Contributors: 5 DOE Labs
BNL, Fermilab, NERSC, ORNL, SLAC. 65 Universities. 5 partner campus/regional grids.
Accessible resources: 43,000+ cores 6 Petabytes disk cache 10 Petabytes tape stores 14 internetwork partnership
Usage 15,000 CPU WallClock days/day 1 Petabyte data distributed/month. 100,000 application jobs/day. 20% cycles through resource sharing,
opportunistic use.
• Research Participation Support for Science Gateways over 100 scientific data
collections (discipline specific databases)
Contributors: 11 Supercomputing centers
Indiana, LONI, NCAR, NCSA, NICS, ORNL, PSC, Purdue, SDSC, TACC and UC/ANL
• Computational resources: – > 1 Petaflop computing capability– 30 Petabytes of storage (disk and
tape)– Dedicated high performance
internet connections (10G( TFLOPS (161K-cores) in 750
parallel computing systemsand growing
• Research Participation Support for Science Gateways over 100 scientific data
collections (discipline specific databases)
Contributors: 11 Supercomputing centers
Indiana, LONI, NCAR, NCSA, NICS, ORNL, PSC, Purdue, SDSC, TACC and UC/ANL
• Computational resources: – > 1 Petaflop computing capability– 30 Petabytes of storage (disk and
tape)– Dedicated high performance
internet connections (10G) 750 TFLOPS (161K-cores) in
parallel computing systems and growing
TeraGridOSG
• grid service providers:– middleware developers– cluster, network and storage administrators– local-grid communities
• the grid consumers:– global collaborations – single researchers– campus communities – under-served science domains
into a cooperative infrastructure to share and sustain a common heterogeneous distributed facility in the US and beyond.
The Open Science Grid Consortium brings:
96 Resources across production & integration infrastructures
30 Virtual Organizations +6 operations Includes 25% non-physics.
~30,000 CPUs (from 30 to 4000)~6 PB Tapes
~4 PB Shared Disk
Snapshot of Jobs on OSGs
Sustaining through OSG submissions:3,000-4,000 simultaneous jobs .
~100K jobs/day~50K CPUhours/day.
Peak test jobs of 15K a day.
Using production & research networks
OSG Snapshot
The Grid Middleware Stack (and course modules)
Grid Security Infrastructure (M4)
JobManagement (M4)
DataManagement (M5)
Grid InformationServices (M4)
Core Globus Services (M4)
Standard Network Protocols and Web Services
Workflow system (explicit or ad-hoc) (M2)
Grid Application (M2)(often includes a Portal)
25
Globus and Condor play key roles
• Globus Toolkit provides the base middleware– Client tools which you can use from a command line– APIs (scripting languages, C, C++, Java, …) to build
your own tools, or use direct from applications– Web service interfaces– Higher level tools built from these basic components,
e.g. Reliable File Transfer (RFT)
• Condor provides both client & server scheduling– In grids, Condor provides an agent to queue, schedule
and manage work submission
26
Organization A Organization B
Compute Server C1Compute Server C2
Compute Server C3
File server F1 (disks A and B)
Person C(Student)
Person A(Faculty)
Person B(Staff) Person D
(Staff)Person F(Faculty)
Person E(Faculty)
Virtual Community C
Person A(Principal Investigator)
Compute Server C1'
Person B(Administrator)
File server F1 (disk A)
Person E(Researcher)
Person D(Researcher)
Virtual Organization (VO) Concept
• VO for each application or workload• Carve out and configure resources for a particular
use and set of users
To efficiently use a Grid, you must
locate and monitor its resources.
• Check the availability of different grid sites
• Discover different grid services
• Check the status of “jobs”
• Make better scheduling decisions with information maintained on the “health” of sites
Virtual Organization Resource Selector - VORS
http://vors.grid.iu.edu/
• Custom web interface to a grid scanner that checks services and resources on:– Each Compute Element – Each Storage Element
• Very handy for checking:– Paths of installed tools on Worker Nodes.– Location & amount of disk space for planning a workflow.– Troubleshooting when an error occurs.
Grid School Syllabus• Intro to distributed computing and the Grid• Grid security and basic Grid access• Grid resource and job management• Grid data management• Building, monitoring, maintaining & using Grids• Grid applications and frameworks• Workflow and related issues
Conclusion: Why Grids?
• New approaches to inquiry based on– Deep analysis of huge quantities of data– Interdisciplinary collaboration– Large-scale simulation and analysis– Smart instrumentation– Dynamically assemble the resources to tackle a
new scale of problem
• Enabled by access to resources & services without regard for location & other barriers
35
Grids:Because Science needs community …
• Teams organized around common goals– People, resource, software, data, instruments…
• With diverse membership & capabilities– Expertise in multiple areas required
• And geographic and political distribution– No location/organization possesses all required skills
and resources
• Must adapt as a function of the situation– Adjust membership, reallocate responsibilities,
renegotiate resources
36
Getting Started with OSG
• I want to use OSG resources
• I want to get my application running on OSG
• I want information about adapting my campus IT facility to form a campus grid
• I want to federate or partner my grid with OSG
• I want to make resources available to OSG
• I want to help build OSG
I want to use OSG resources
• Must join a VO:– Individual/small for independent research
• Join OSG VO
– Join an existing member VO (see the list of current OSG VOs )
– Form new VO for your research community
• Run VO specific applications
Want to get my app running on OSG
• Engagement team– Dedicated effort– Genetics, library science, earthquake simulation, video
processing, physics; Examples:• Production running using 20,000+ CPU hours of the CHARMM molecular dynamic simulation to
the problem of water penetration in staphylococcal nuclease using opportunistically available resources across 10+ OSG sites ( see Grids work like a CHARMM for molecular dynamics)
• Improvement of the performance of the nanoWire application from the nanoHub project on OSG/TeraGrid, such that stable running of batches of 500 jobs across more than 5 sites is routine; (see Keeping up with Moore's Law)
• Adaptation and production running opportunistically using 100,000+ CPU hours of the Rosetta application from the Kuhlman Laboratory in North Carolina across more than 13 OSG sites (see Designing proteins from scratch)
• Production runs of the Weather Research and Forecast (WRF) application using more 150,000 CPUhours on the NERSC OSG site at Lawrence Berkeley National Laboratory (LBNL)
Want to form a campus grid
• You have a campus IT facility– Want to make it a campus grid– And federate it with OSG
• OSG is committed to including US universities in the national cyberinfrastructure. – The OSG middleware and operational framework enables any
site to participate as an OSG resource, provided it is a well maintained resource that users can count on
• Technically there are no hurdles in having every US university and college contribute resources to OSG and use OSG resources in return. – See module on Friday
• Several campuses have done so very well: Purdue University, University of Wisconsin- Madison, and Clemson University
• Several other universities participate in OSG through individual research groups.
I want to federate or partner my grid with OSG
• OSG Consortium envisions a world-wide grid formed of a number of different federations of grids (analogous to the Internet as a network of networks)
• Federation, therefore, is a natural concept within OSG
• We are interested in partnerships with other grids trying to develop richer methods and tools for federation.
I want to make resources accessible to OSG
• Recommended that you join a VO• Minimal requirements
– sufficient to assure interoperability, stability– set of "standard" services which define the
requirements on interfaces and capabilities.
• Register resource with GOC– See module on Friday
OSG - Education, Training and
Outreach
OpenScienceGrid.org/Education
OpenScienceGrid.org/About/Outreach
OSG EOT Mission• Organize and deliver training for OSG
– OSG End Users – Site Administrators – Support new communities / VOs joining OSG
• Engage young people in (e)Science and CS– Primary focus: graduate students and faculty– Promote and train in interdisciplinary collaboration– Reach high schools through I2U2 (QuarkNet follow-on)
• Reach out – To under-represented communities
• Engage and assist minority students and minority serving institutions by providing resources and opportunities.
– internationally• Strengthen and assist emerging, underserved regions of strategic importance to
form bonds to US science and Grid communities• Focus (for outreach) is on Latin America and Africa• OISE focus on engagement in Europe and Asia
OSG EOT Program Overview
• End User Education– In-person workshops– Online training– EOT VO for student engagement, access and support
• Community Outreach– International student/faculty exchange via OISE– Supporting under-represented and under-resourced communities in US,
Latin America and Africa through workshops, technical assistance and grid access
– High School Education – I2U2 support - http://ed.fnal.gov/uueo/i2u2.html• Site Admin Training
– Training grid administrators in setup and support of OSG sites using the OSG/VDT software stack
2007-08 Workshop Programwww.opensciencegrid.org/workshops
• Georgetown University Grid School 2008, April 15-17, DC
• Tuskegee University Grid School 2008, Feb 6-8 - Tuskegee AL
• Florida International Grid School 2008, Jan 23-25, at Florida International University, Miami, Florida
• Supercomputing ’07 tutorials, Nov 11 & 13, at Reno, Nevada
• Great Plains Grid School (GPGS’07), Aug 8-10, at the U. of Nebraska-Lincoln
• Rio Grande Grid School (RGGS’07), Jun 8-10, at the U. of Texas at Brownsville, coordinated with UT-Pan American
• TeraGrid Conference tutorials, Jun 4-8, at the U. of Wisconsin-Madison
• South Africa Workshop, Mar 26-30, at the IFIP School on Software (ISS’07), Gordon's Bay, South Africa
• Midwest Grid Workshop (MGW’07), Mar 24-25 at the U. of Illinois at Chicago
• Argentine Grid Workshop, Mar 12-14 at Santa Fe, Argentina
Self-paced / online instruction
• opensciencegrid.org/OnlineGridCourse
• Flexible roadmaps for navigating the material
• Lectures and labs
• Access to online community to provide support
• Online office hours
I2U2 Interactions In Understanding the
Universe• The Grid for Secondary Science
Education “educational virtual organization”• creates an infrastructure to develop
– hands-on laboratory course content and– an interactive learning experience that
• brings tangible aspects of each experiment into a “virtual laboratory.”
• These labs use the Grid for education in the same way that science uses the Grid.
• www.i2u2.org
I2U2
• "e-Labs”– delivered as Web-based portals accessible in the classroom and at
home– implemented with of Web-based media capabilities
• "i-Labs”– delivered as interactive interfaces typically located within science
museums and similar public venues– leverage the latest advances in
• display technology and • human-computer interaction,
– and bring the experiences and appreciation of scientific investigation and inquiry to the wide audience of informal education
List of e-Labs– Cosmic Ray e-Lab
• High school students investigate data from a cosmic ray detector array. (not necessary to have a detector to participate.)
• Possible investigations: ・Muon Lifetime ・ Diurnal changes in flux ・ Effects of shielding ・ High-energy showers ・ Altitude effects
– CMS Test Beam e-Lab (Beta Version)• High school students analyze CMS test beam data in an online graphical
ROOT environment.
• Shower Depth ・ Lateral Shower Size ・ Beam Purity ・ Detector Resolution
– LIGO e-Lab (Beta version) • High school and middle school students investigate seismic behavior with
data from LIGO ( Laser Interferometer Gravitational-wave Observatory).
• Earthquake Studies ・ Frequency Band Studies ・Microseismic Studies ・Studies of Human-induced Seismic Activity
– ATLAS e-Lab– STAR e-Lab
Cooperation with EGEE
International Schools on Grid Computing
– OSG as co-organizer for ISSGC’07/08 and ISSGC’09
International Summer School on Grid Computing www.issgc.org
• sponsor alumni of US Grid Schools to attend the International Summer school.
– Joint lectureships and material sharing / development efforts
– Content sharing
Cooperation with TeraGrid
• Another major national cyberinfrastructure
• Use of TG and OSG resources
• Contribute content
• Joint training
Education VO
• Interested in getting started with OSG ?
• Join OSGEDU VO – Use OSG resources– Contribute resources
• Wiki, email lists, follow-up discussions– Support, engagement– Postings of opportunities for students
Students2004-2008 facts:
• International participation:– Argentina , Brazil, Canada, Colombia, India, Mexico, New Zealand, Russia,
South Africa, Uruguay• Women
– Approx. 15%• Minorities
– Approx 15%
Try to improve these statistics
Participants’ domains
Computer Science Image processing Communications Networking
Physics Astrophysics High Energy Nuclear Physics Optical Networks Theoretical solid state physics Atomic Physics Computational Physics
Chemistry Computational Chemistry Molecular Dynamics & Simulation
Applied Mathematics
Geosciences
Computational Multibody Dynamics for Distributed computing
Judicial Administration
Engineering Materials Science
Quantum theory
…and others …
Summary of OSG
• Provides core services, software and a distributed facility for an increasing set of research communities.
• Helps VOs access resources on many different infrastructures.
• Interested in collaborating and contributing our experience and efforts.
Acknowledgments
Various OSG members and contributors (Alain Roy, Mike Wilde, Ruth Pordes,
Gabrielle Allen and many others …)
Joining OSG
• Assumption:– You have a campus grid
• Question: – What changes do you need to make to join OSG?
Your Campus Grid
• assuming that you have a cluster with a batch system:
– Condor– Sun Grid Engine– PBS/Torque– LSF
Administrative Work
• You need a security contact– Who will respond to security concerns
• You need to register your site
• You should have a web page about your site.– This will be published– People can learn about your site.
Big Picture
• Compute Element (CE)– OSG jobs submitted to CE, which gives
them to batch system– Also has information services and lots of
support software
• Shared file system– OSG requires a couple of directories to be
mounted on all worker nodes
• Storage Element (SE)– How do you manage your storage at your
site
Installing Software
• The OSG Software Stack– Based on the VDT
• The majority of the software you’ll install• It is grid independent
– OSG Software Stack:• VDT + OSG-specific configuration
• Installed via Pacman
What is installed?
• GRAM: – Allows job submissions
• GridFTP: – Allows file transfers
• CEMon/GIP: – Publishes site information
• Some authorization mechanism– grid-mapfile: file that lists authorized users, or– GUMS (grid identity mapping service)
• And a few other things…
OSG Middleware
Infr
astr
uctu
reA
ppli
cati
ons
VO Middleware
Core grid technology distributions: Condor, Globus, Myproxy: shared with TeraGrid and
others
Virtual Data Toolkit (VDT) core technologies + software needed by
stakeholders: many components shared with EGEE
OSG Release Cache: OSG specific configurations, utilities etc.
HEP
Data and workflow management etc
Biology
Portals, databases etc
User Science Codes and Interfaces
Existing Operating, Batch systems and Utilities.
Astrophysics
Data replication etc
Shared file system
• OSG_APP– For users to store applications
• OSG_DATA– A place to store data– Highly recommended, not required
• OSG_GRID– Software needed on worker nodes– Not required– May not exist on non-Linux clusters
• Home directories for users– Not required, but often very convenient
Storage Element
• Some folks require more sophisticated storage
management
– How do worker nodes access data?
– How do you handle terabytes (petabytes?) of data
• Storage Elements are more complicated
– More planning needed
– Some are complex to install and configure
• Two OSG supported options of SRMs:
– dCache
– Bestman