R. Pordes, I Brazilian LHC Computing Workshop 1 What is Open Science Grid? High Throughput...
-
date post
19-Dec-2015 -
Category
Documents
-
view
216 -
download
2
Transcript of R. Pordes, I Brazilian LHC Computing Workshop 1 What is Open Science Grid? High Throughput...
R. Pordes, I Brazilian LHC Computing Workshop 1
What is Open Science Grid? High Throughput Distributed Facility
Shared opportunistic access to existing clusters, storage and networks.
Owner controlled resources and usage policies.
Supports Science Funded by NSF and DOE projects. Common technologies & cyber-infrastructure.
Open and Heterogeneous Research groups transitioning from & extending (legacy)
systems to Grids: Experiments developing new systems. Application Computer Scientists Real life use of technology, integration, operation.
R. Pordes, I Brazilian LHC Computing Workshop 2
Who is OSG: a Consortium US DOE HENP Laboratory Facilities + Universities
(US) LHC Collaborations + offshore sites LIGO Condor Project Running HENP Experiments - CDF, D0, STAR… Globus/CDIGS LBNL SDM
Collaboration of users, developers, grid technologists, facility administrators.
Training & help for administrators and users
R. Pordes, I Brazilian LHC Computing Workshop 3
OSG 1 day last week:
50 Clusters : used locally as well as through the grid
5 Large disk or tape stores 23 VOs >2000 jobs running through Grid;
BioinformaticsBioinformatics
Routed from Local UWisconsin Campus Grid
Routed from Local UWisconsin Campus Grid
2000 running jobs2000 running jobs
500 waiting jobs500 waiting jobs
LHCLHC
Run IIRun II
R. Pordes, I Brazilian LHC Computing Workshop 4
Broad Engagement
Executive Director Ruth Pordes Fermilab, US CMS Facility Coordinator Miron Livny Condor Project Applications Torre Wenaus &
Frank Wuerthwein US ATLAS US CMS
Council Chair Bill Kramer LBNL NERSC Resource Managers Paul Avery &
Albert Lazzarini iVDGL LIGO
Education & Outreach
Mike Wilde Globus
Deputy Directors Rob Gardner & Doug Olson
US ATLAS LBNL, STAR
Engagement Alan Blatecky RENCI Middleware Alain Roy VDT Operations Leigh Grundhoefer Grid Operations Security Officer Don Petravick Fermilab Liaison to EGEE John Huth US ATLAS Liaison to Teragrid Mark Green University at Buffalo
R. Pordes, I Brazilian LHC Computing Workshop 5
The OSG World: Partnerships
Campus Grids: GRid Of IoWa, Grid Laboratory Of Wisconsin, Crimson Grid, Texas Advanced Computer Center, Center for Computational Research /Buffalo, TIGRE, FermiGrid
Grid Projects DISUN CDIGS
National Grids: TeraGrid, HEP-Brazil International Grids: EGEE
R. Pordes, I Brazilian LHC Computing Workshop 6
What is an OSG Job? “work done” accomplished by and delivered as “benefit received”; accountable to multiple organizations
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
OSG EGEE
Job Counted on Campus Grid, OSG and EGEE.
Job Counted on Campus Grid, OSG and EGEE.
MyApplication, EGEE RB, , VDS, OSG RESS
MyApplication, EGEE RB, , VDS, OSG RESS
Job SubmissionCondor-G
Job SubmissionCondor-G
Job does work benefiting WLCG.
Job does work benefiting WLCG.
R. Pordes, I Brazilian LHC Computing Workshop 7
Common Middleware provided through Virtual Data ToolkitDomain science requirements.
OSG stakeholders and middleware developer (joint) projects.
Integrate into VDT Release. Deploy on OSG integration grid
Include in OSG release & deploy to OSG production.
Globus, Condor,EGEE etc
Test on “VO specific grid”
R. Pordes, I Brazilian LHC Computing Workshop 8
Reliable: Central Operations Activities
Automated validation of basic services and site configuration Configuration of HeadNode and Storage to reduce errors: Remove dependence on
Shared File System Condor-managed GRAM
fork queue Scaling tests of WS-GRAM
and GridFTP.
Daily Grid Exerciser:
R. Pordes, I Brazilian LHC Computing Workshop 9
OSG Drivers: Research groups transitioning from & extending (legacy) systems to Grids:
US LHC Collaborations Contribute to & depend on milestones, functionality,
capacity of OSG. Commitment to general solutions, sharing resources &
technologies; Application Computer Scientists
Real life use of technology, integration, operation. Federations with Campus Grids
Bridge & interface Local & Wide Area Grids.
Interoperation & partnerships with national/ international infrastructures Ensure transparent and ubiquitous access. Work towards standards.
LIGO- gravitational wave physics;STAR - nuclear physics, CDF, D0, - high energy physics,SDSS - astrophysicsGADU - bioinformaticsNanohub
NMI, Condor, Globus, SRM
GLOW, FermiGrid, GROW, Crimson, TIGRE
EGEE, TeraGrid, INFNGrid
R. Pordes, I Brazilian LHC Computing Workshop 10
LHC Physics drive schedule and performance envelope Beam starts in 2008: Distributed System must serve 20PB of
data in served across 30PB disk distributed across 100 sites worldwide to be analyzed by 100MSpecInt2000 of CPU.
Service Challenges give steps to full system
1 GigaByte/sec1 GigaByte/sec
R. Pordes, I Brazilian LHC Computing Workshop 11
Bridging Campus Grid Jobs - GLOW
Dispatch jobs from local security, job, storage infrastructure and “uploading” to wide-area infrastructure.
Fast ramp up in last week. Currently running the football pool problem which has application in data compression, coding theory, and statistical designs.
R. Pordes, I Brazilian LHC Computing Workshop 12
Genome Analysis and Database Update system Request: 1000 CPUs for 1-2 weeks. Once a month.
3 different applications: BLAST, Blocks, Chisel.
Currently ramping up on OSG and receiving 600 CPUs and 17,000 jobs a week.
R. Pordes, I Brazilian LHC Computing Workshop 13
Common Middleware provided through Virtual Data ToolkitDomain science requirements.
OSG stakeholders and middleware developer (joint) projects.
Integrate into VDT Release. Deploy on OSG integration grid
Include in OSG release & deploy to OSG production.
Globus, Condor,EGEE etc
Test on “VO specific grid”
Condor project
R. Pordes, I Brazilian LHC Computing Workshop 14
of course a special grid …
it’s the people…(some of them at the consortium meeting in Jan 06)
R. Pordes, I Brazilian LHC Computing Workshop 15
TeraGridThrough high-performance network connections, TeraGrid integrates high-performance computers, data resources and tools, and high-end experimental facilities around the (US) country.
CDF MonteCarlo jobs running on Purdue TeraGrid resource; able to access OSG data areas and be accounted to both Grids.
http://www.nsf.gov/news/news_images.jsp?cntn_id=104248&org=OLPA
R. Pordes, I Brazilian LHC Computing Workshop 16
OSG: More than a US Grid
Taiwan - (CDF, LHC)
Brazil - (D0, STAR,
LHC)
Korea
R. Pordes, I Brazilian LHC Computing Workshop 17
OSG: Where to find information: OSG Web site: www.opensciencegrid.org
Work in progress: http://osg.ivdgl.org/twiki/bin/view/Integration/OverviewGuide
Virtual Data Toolkit: http://vdt.cs.wisc.edu//index.html
News about Grids in Science in “Science Grid This Week”: www.interactions.org/sgtw
OSG Consortium meeting Seattle Aug 21st.Thank you!
R. Pordes, I Brazilian LHC Computing Workshop 18
BDIILDAPURLs
OSG - EGEE Interoperation for WLCG Jobs
GRAMGRAMGRAM
GRAM
GRAM
T2
GRAM
T2
GRAM
T2
SRM
GRAM
T2T2
SRM
Site
SRM T2
GRAM
T2
SRM
GRAM
T2
SRM
T2T2
SRM
Site
SRM
VO RBVO RBVO RB
VO UI
BDII
Data Stores
Picture thanks to I. FiskPicture thanks to I. Fisk
R. Pordes, I Brazilian LHC Computing Workshop 19
Open Science Grid in 1 minute: OSG Resources - use and
policy under owner control. Clusters and storage shared across local, Campus intra-grid, Regional Grid and large federated Inter-Grids.
OSG Software Stack - based on Virtual Data Toolkit. Interfaces: Condor-G job submission interface; GridFTP data movement SRM storage management; Glue Schema V1.2; easy to
configure GIPs;, CEMON coming in 3 months.
OSG Use - Register VO with with Operations Center; Provide URL for VOMS service - this must
be propagated to sites. Contact for Support Center. Join operations groups.
OSG Job Brokering, Site Selection - no central or unique service. LIGO uses Pegasus; SDSS uses VDS; STAR uses Star-schedule; CMS uses EGEE-RB; ATLAS uses Panda; CDF uses CDF GlideCAF; D0 uses SAM-JIM; GLOW uses “condor-schedd on the side”. Nano-hub uses application portal.
OSG Storage & Space Management shared file systems; persistent VO application areas; SRM interfaces.
OSG Operations - Distributed including each VO, Campus Grid. Operations is also a WLCG ROC.
OSG Accounting & Monitoring -MonaLisa; can support rGMA; OSG meters/probes for Condor being released soon. US Tier-1s reporting monthly to WLCG APEL.
Services to the US Tier-1 Sites LHCOPNApril 4th, 2006Joe Metzger
ESnet Engineering GroupLawrence Berkeley National Laboratory
R. Pordes, I Brazilian LHC Computing Workshop 21
ESnet Target Architecture: High-reliability IP Core
Chi
cago
Atl
anta
Seattle
Albuquerque
IP Core
LA
Denver
Primary DOE Labs
Possible hubs
SDN hubs
IP core hubs
Washington DCS
un
nyv
ale
New York
San Diego
Cle
vela
nd
R. Pordes, I Brazilian LHC Computing Workshop 22
ESnet Target Architecture: Science Data Network
New York
Chi
cago
Atl
anta
Seattle
Albuquerque
Science Data Network Core
San Diego
LA
Su
nn
yval
e
Denver
Primary DOE Labs
Possible hubs
SDN hubs
IP core hubs
Cle
vela
nd
Washington DC
R. Pordes, I Brazilian LHC Computing Workshop 23
10-50 Gbps circuitsProduction IP coreScience Data Network coreMetropolitan Area NetworksInternational connections
MetropolitanArea Rings
ESnet Target Architecture: IP Core+Science Data Network Core+Metro Area Rings
New York
Chi
cago
Washington DC
Atl
anta
Seattle
AlbuquerqueSan
Diego
LA
Su
nn
yval
e
Denver
Loop off Backbone
SDN Core
IP Core
Primary DOE Labs
Possible hubs
SDN hubs
IP core hubs
international connections international
connections
international connections
international connections
international connectionsin
tern
atio
nal
co
nn
ecti
on
s
Cle
vela
nd