Upload
giles
View
36
Download
1
Tags:
Embed Size (px)
DESCRIPTION
LHC Experiments and the PACI A Partnership for Global Data Analysis. Harvey B. Newman, Caltech Advisory Panel on CyberInfrastructure National Science Foundation November 29, 2001 http://l3www.cern.ch/~newman/LHCGridsPACI.ppt. Global Data Grid Challenge. - PowerPoint PPT Presentation
Citation preview
Harvey B. Newman, CaltechHarvey B. Newman, Caltech Advisory Panel on CyberInfrastructureAdvisory Panel on CyberInfrastructure
National Science FoundationNational Science Foundation November 29, 2001November 29, 2001
http://l3www.cern.ch/~newman/LHCGridsPACI.ppthttp://l3www.cern.ch/~newman/LHCGridsPACI.ppt
LHC Experiments and the PACILHC Experiments and the PACIA Partnership for Global Data AnalysisA Partnership for Global Data Analysis
Global Data Grid ChallengeGlobal Data Grid Challenge
““Global scientific communities, served by networks Global scientific communities, served by networks with bandwidths varying by orders of magnitude, with bandwidths varying by orders of magnitude, need to perform computationally demanding need to perform computationally demanding analyses of geographically distributed datasets analyses of geographically distributed datasets that will grow by at least 3 orders of magnitude that will grow by at least 3 orders of magnitude over the next decade, from the 100 Terabyte to over the next decade, from the 100 Terabyte to the 100 Petabyte scale [from 2000 to 2007]”the 100 Petabyte scale [from 2000 to 2007]”
The Large Hadron Collider (2006-)The Large Hadron Collider (2006-) The Next-generation Particle Collider The Next-generation Particle Collider
The largest superconductor The largest superconductor installation in the worldinstallation in the world
Bunch-bunch collisions at 40 MHz,Bunch-bunch collisions at 40 MHz,Each generating ~20 interactionsEach generating ~20 interactions
Only one in a trillion may lead Only one in a trillion may lead to a major physics discovery to a major physics discovery
Real-time data filtering: Real-time data filtering: Petabytes per second to Gigabytes Petabytes per second to Gigabytes per secondper second
Accumulated data of many Accumulated data of many Petabytes/YearPetabytes/Year
Large data samples explored and analyzed by thousands Large data samples explored and analyzed by thousands of globally dispersed scientists, in hundreds of teamsof globally dispersed scientists, in hundreds of teams
Four LHC Experiments: The Four LHC Experiments: The Petabyte to Exabyte ChallengePetabyte to Exabyte ChallengeATLAS, CMS, ALICE, LHCBATLAS, CMS, ALICE, LHCB
Higgs + New particles; Quark-Gluon Plasma; CP ViolationHiggs + New particles; Quark-Gluon Plasma; CP Violation
Data storedData stored ~40 Petabytes/Year and UP; ~40 Petabytes/Year and UP; CPU CPU 0.30 Petaflops and UP 0.30 Petaflops and UP
0.1 to 1 Exabyte (1 EB = 100.1 to 1 Exabyte (1 EB = 101818 Bytes) Bytes) (2007) (~2012 ?) for the LHC Experiments(2007) (~2012 ?) for the LHC Experiments
Evidence for the Higgs at LEP at M~115 GeV The LEP Program Has Now Ended
All charged tracks with pt > 2 GeV
Reconstructed tracks with pt > 25 GeV
(+30 minimum bias events)
109 events/sec, selectivity: 1 in 1013 (1 person in a thousand world populations)
LHC: Higgs Decay into 4 muons LHC: Higgs Decay into 4 muons 1000X LEP Data Rate1000X LEP Data Rate
LHC Data Grid HierarchyLHC Data Grid Hierarchy
Tier 1
Tier2 Center
Online System
CERN 700k SI95 ~1 PB Disk; Tape Robot
FNAL: 200k SI95; 600 TBIN2P3 Center INFN Center RAL Center
InstituteInstituteInstituteInstitute ~0.25TIPS
Workstations
~100-400 MBytes/sec
2.5 Gbps
100 - 1000
Mbits/sec
Physicists work on analysis “channels”Each institute has ~10 physicists working on one or more channels
Physics data cache
~PByte/sec
~2.5 Gbits/sec
Tier2 CenterTier2 CenterTier2 Center~2.5 Gbps
Tier 0 +1
Tier 3
Tier 4
Tier2 Center Tier 2
Experiment
CERN/Outside Resource Ratio ~1:2Tier0/( Tier1)/( Tier2) ~1:1:1
TeraGrid:TeraGrid:NCSA, ANL, SDSC, CaltechNCSA, ANL, SDSC, Caltech
NCSA/UIUC
ANL
UIC Multiple Carrier HubsStarlight / NW Univ
Ill Inst of TechUniv of Chicago
Indianapolis (Abilene NOC)
I-WIRE
Pasadena
San Diego
DTF Backplane(4x: 40 Gbps)
Abilene
Chicago
IndianapolisUrbana
OC-48 (2.5 Gb/s, Abilene)Multiple 10 GbE (Qwest)Multiple 10 GbE (I-WIRE Dark Fiber)
Solid lines in place and/or available in 2001 Dashed I-WIRE lines planned for Summer 2002
Source: Charlie Catlett, Argonne
StarLight: Int’l Optical Peering Point(see www.startap.net)
A Preview of the Grid Hierarchyand Networks of the LHC Era
Current Grid Challenges: Resource Current Grid Challenges: Resource Discovery, Co-Scheduling, TransparencyDiscovery, Co-Scheduling, Transparency
Discovery and Efficient Co-Scheduling of Computing, Discovery and Efficient Co-Scheduling of Computing, Data Handling, and Network ResourcesData Handling, and Network Resources
Effective, Consistent Replica ManagementEffective, Consistent Replica Management Virtual Data: Recomputation Versus Data Transport Virtual Data: Recomputation Versus Data Transport
DecisionsDecisions Reduction of Complexity In a “Petascale” WorldReduction of Complexity In a “Petascale” World
““GA3”: Global Authentication, Authorization, AllocationGA3”: Global Authentication, Authorization, Allocation VDT: Transparent Access to Results VDT: Transparent Access to Results
(and Data (and Data When Necessary)When Necessary) Location Independence of the User Analysis, Grid,Location Independence of the User Analysis, Grid,
and Grid-Development Environmentsand Grid-Development Environments Seamless Multi-Step Data Processing and Analysis:Seamless Multi-Step Data Processing and Analysis:
DAGMan (Wisc), MOP+IMPALA(FNAL)DAGMan (Wisc), MOP+IMPALA(FNAL)
CMS Production: Event Simulation CMS Production: Event Simulation and Reconstructionand Reconstruction
““Grid-Enabled”Grid-Enabled” AutomatedAutomated
Imperial Imperial CollegeCollege
UFLUFL
Fully operationalFully operational
CaltechCaltech
PUPUNo PUNo PU
Not Op.Not Op.Not Op.Not Op.
In progressIn progress
Common Common Prod. toolsProd. tools(IMPALA)(IMPALA)
GDMPGDMPDigitizationDigitizationSimulationSimulation
Not Op.Not Op.HelsinkiHelsinkiIN2P3IN2P3WisconsinWisconsinBristolBristol
UCSDUCSD
INFNINFNMoscowMoscowFNALFNALCERNCERN
Worldwide Production
at 12 Sites
US CMS TeraGrid Seamless US CMS TeraGrid Seamless PrototypePrototype
Caltech/Wisconsin Condor/NCSA ProductionCaltech/Wisconsin Condor/NCSA Production Simple Job Launch from CaltechSimple Job Launch from Caltech
Authentication Using Globus Security Infrastructure (GSI)Authentication Using Globus Security Infrastructure (GSI) Resources Identified Using Globus Information Resources Identified Using Globus Information
Infrastructure (GIS)Infrastructure (GIS) CMSIM Jobs (Batches of 100, 12-14 Hours, 100 GB Output)CMSIM Jobs (Batches of 100, 12-14 Hours, 100 GB Output)
Sent to the Wisconsin Condor Flock Using Condor-G Sent to the Wisconsin Condor Flock Using Condor-G Output Files Automatically Stored in NCSA Unitree (Gridftp)Output Files Automatically Stored in NCSA Unitree (Gridftp)
ORCA Phase: Read-in and Process Jobs at NCSAORCA Phase: Read-in and Process Jobs at NCSA Output Files Automatically Stored in NCSA UnitreeOutput Files Automatically Stored in NCSA Unitree
Future: Multiple CMS Sites; Storage in Caltech HPSS Also,Future: Multiple CMS Sites; Storage in Caltech HPSS Also,Using GDMP (With LBNL’s HRM).Using GDMP (With LBNL’s HRM).
Animated Flow Diagram of the DTF Prototype:Animated Flow Diagram of the DTF Prototype: http://cmsdoc.cern.ch/~wisniew/infrastructure.htmlhttp://cmsdoc.cern.ch/~wisniew/infrastructure.html
Baseline BW for the US-CERN Link:Baseline BW for the US-CERN Link: HENP Transatlantic WG (DOE+NSF) HENP Transatlantic WG (DOE+NSF)
US-CERN Plans: 155 Mbps to 2 X 155 Mbps this Year;US-CERN Plans: 155 Mbps to 2 X 155 Mbps this Year; 622 Mbps in April 2002;622 Mbps in April 2002;
DataTAG 2.5 Gbps Research Link in Summer 2002;DataTAG 2.5 Gbps Research Link in Summer 2002;10 Gbps Research Link in ~200310 Gbps Research Link in ~2003
Transoceanic Networking
Integrated with the TeraGrid,
Abilene, Regional Nets
and Continental Network
Infrastructuresin US, Europe,
Asia, South America
2001 2002 2003 2004 2005 2006CMS 100 200 300 600 800 2500
ATLAS 50 100 300 600 800 2500BaBar 300 600 1100 1600 2300 3000CDF 100 300 400 2000 3000 6000D0 400 1600 2400 3200 6400 8000
BTeV 20 40 100 200 300 500DESY 100 180 210 240 270 300
CERNBW
155-310
622 1250 2500 5000 10000
[*] [*] Installed BW. Maximum Link Occupancy 50% AssumedInstalled BW. Maximum Link Occupancy 50% AssumedThe Network Challenge is Shared by Both Next- and The Network Challenge is Shared by Both Next- and
Present Generation ExperimentsPresent Generation Experiments
Transatlantic Net WG (HN, L. Price)Transatlantic Net WG (HN, L. Price) Bandwidth Requirements [*] Bandwidth Requirements [*]
Internet2 HENP Networking WG [*]Internet2 HENP Networking WG [*]MissionMission
To help ensure that the requiredTo help ensure that the required National and international network infrastructuresNational and international network infrastructures Standardized tools and facilities for high performance Standardized tools and facilities for high performance
and end-to-end monitoring and tracking, andand end-to-end monitoring and tracking, and Collaborative systemsCollaborative systems
are developed and deployed in a timely manner, are developed and deployed in a timely manner, and used effectively to meet the needs of the US LHC and and used effectively to meet the needs of the US LHC and other major HENP Programs, as well as the general needs other major HENP Programs, as well as the general needs of our scientific community.of our scientific community.
To carry out these developments in a way that is broadly To carry out these developments in a way that is broadly applicable across many fields, within and beyond the applicable across many fields, within and beyond the scientific communityscientific community
[*] Co-Chairs: S. McKee (Michigan), H. Newman (Caltech); [*] Co-Chairs: S. McKee (Michigan), H. Newman (Caltech); With thanks to R. Gardner and J. Williams (Indiana)With thanks to R. Gardner and J. Williams (Indiana)
Grid R&D: Focal Areas for Grid R&D: Focal Areas for NPACI/HENP PartnershipNPACI/HENP Partnership
Development of Grid-Enabled User Analysis EnvironmentsDevelopment of Grid-Enabled User Analysis Environments CLARENS (+IGUANA) CLARENS (+IGUANA) Project for Portable Grid-EnabledProject for Portable Grid-Enabled
Event Visualization, Data Processing and Analysis Event Visualization, Data Processing and Analysis Object Integration:Object Integration: backed by an ORDBMS, and backed by an ORDBMS, and
File-Level Virtual Data CatalogsFile-Level Virtual Data Catalogs Simulation Toolsets for Systems Modeling, OptimizationSimulation Toolsets for Systems Modeling, Optimization
For example: theFor example: the MONARC MONARC System System Globally Scalable Agent-Based Realtime Information Globally Scalable Agent-Based Realtime Information
Marshalling SystemsMarshalling Systems To face the next-generation challenge of DynamicTo face the next-generation challenge of Dynamic
Global Grid design and operationsGlobal Grid design and operations Self-learning (e.g. SONN) optimization Self-learning (e.g. SONN) optimization Simulation (Now-Casting) enhanced: to monitor, track and Simulation (Now-Casting) enhanced: to monitor, track and
forward predict site, network and global system stateforward predict site, network and global system state 1-10 Gbps Networking development and global deployment1-10 Gbps Networking development and global deployment
Work with the TeraGrid, STARLIGHT, Abilene, the iVDGL Work with the TeraGrid, STARLIGHT, Abilene, the iVDGL GGGOC, HENP Internet2 WG, Internet2 E2E, and DataTAGGGGOC, HENP Internet2 WG, Internet2 E2E, and DataTAG
Global Collaboratory Development: e.g. VRVS, Access GridGlobal Collaboratory Development: e.g. VRVS, Access Grid
CLARENS: a Data AnalysisCLARENS: a Data AnalysisPortal to the Grid: Steenberg (Caltech)Portal to the Grid: Steenberg (Caltech)
A highly functional graphical interface, A highly functional graphical interface, Grid-enabling the working environment for Grid-enabling the working environment for “non-specialist” physicists’ data analysis“non-specialist” physicists’ data analysis
Clarens consists of a server communicating with Clarens consists of a server communicating with various clients via the commodity XML-RPC protocol. various clients via the commodity XML-RPC protocol. This ensures implementation independence.This ensures implementation independence.
The server is implemented in C++ to give access The server is implemented in C++ to give access to the CMS OO analysis toolkit.to the CMS OO analysis toolkit.
The server will provide a remote API to Grid tools:The server will provide a remote API to Grid tools: Security services provided by the Grid (GSI)Security services provided by the Grid (GSI) The Virtual Data Toolkit: Object collection accessThe Virtual Data Toolkit: Object collection access Data movement between Tier centers using GSI-FTPData movement between Tier centers using GSI-FTP CMS analysis software (ORCA/COBRA)CMS analysis software (ORCA/COBRA)
Current prototype is running on the Caltech Proto-Tier2Current prototype is running on the Caltech Proto-Tier2 More information at More information at http://heppc22.http://heppc22.hephep..caltechcaltech..eduedu, ,
along with a web-based demoalong with a web-based demo
Modelling and understanding current systems, their performance and limitations, is essential for the design of the future large scale distributed processing systems.
The simulation program developed within the MONARC (MModels odels OOf f NNetworked etworked AAnalysis At nalysis At RRegional egional CCenters) enters) project is based on a process oriented approach for discrete event simulation. It is based on the on Java(TM) technology and provides a realistic modelling tool for such large scale distributed systems.
Modeling and Simulation:Modeling and Simulation:MONARC SystemMONARC System
SIMULATION of Complex Distributed Systems
MONARC SONN: 3 Regional Centres MONARC SONN: 3 Regional Centres Learning to Export Jobs (Day 9)Learning to Export Jobs (Day 9)
NUST20 CPUs
CERN30 CPUs
CALTECH25 CPUs
1MB/s ; 150 ms RTT
1.2 MB/s
150 ms RTT
0.8
MB/s
200 m
s RTT
Day = 9
<E> = 0.73
<E> = 0.66
<E> = 0.83
TCP Protocol Study: Limits We determined Precisely
The parameters which limit the throughput over a high-BW, long delay (170 msec) network
How to avoid intrinsic limits; unnecessary packet loss
Methods Used to Improve TCP Linux kernel programming in order
to tune TCP parameters We modified the TCP algorithm A Linux patch will soon be
available
Result: The Current State of the Art for Reproducible Throughput
125 Mbps between CERN and Caltech
135 Mbps between CERN and Chicago
Status: Ready for Tests at Higher BW (622 Mbps) in Spring 2002
Congestion window behavior of a TCP connection over the transatlantic line
3) Back to slow start(Fast Recovery couldn’t repair the lostThe packet lost is detected by timeout => go back to slow start cwnd = 2 MSS)
2) Fast Recovery (Temporary state to repair the lost)1) A packet is
lost New loss
Losses occur when the cwnd is larger than 3,5 Mbyte
TCP performance between CERN and Caltech
0
20
40
60
80
100
120
140
1 2 3 4 5 6 7 8 9 10 11
Connection number
Mbp
s
Without tunning
By tunning theSSTHRESHparameter
Maximizing US-CERN TCP Maximizing US-CERN TCP Throughput (S.Ravot, Caltech)Throughput (S.Ravot, Caltech)
Reproducible 125 Mbps Between
CERN and Caltech/CACR
Agent-Based Distributed System: Agent-Based Distributed System: JINI Prototype (Caltech/PakistanJINI Prototype (Caltech/Pakistan))
Includes “Station Servers” (static) that host mobile “Dynamic Services”
Servers are interconnected dynamically to form a fabric in which mobile agents travel, with a payload of physics analysis tasks
Prototype is highly flexible and robust against network outages
Amenable to deployment on leading edge and future portable devices (WAP, iAppliances, etc.) “The” system for the
travelling physicist The Design and Studies with this
prototype use the MONARC Simulator, and build on SONN studies See http://home.cern.ch/clegrand/lia/ See http://home.cern.ch/clegrand/lia/
StationServer
StationServer
StationServer
LookupService
LookupService
Proxy Exchange
Registration
Service Listener
Lookup Discovery
Service
Remote Notification
RCMonitorService
Farm Monitor
Client(other service)
LookupService
LookupService Registration
Farm Monitor
Discovery
Proxy
Component Factory
GUI marshaling Code Transport RMI data access
Push & Pullrsh & ssh existing scripts
snmp
Globally Scalable Monitoring ServiceGlobally Scalable Monitoring Service
ExamplesExamples GLASTGLAST meeting meeting
10 participants connected via VRVS (and 16 participants in Audio only)10 participants connected via VRVS (and 16 participants in Audio only)
VRVS7300 Hosts; 4300 Registered Users In 58 Countries34 Reflectors; 7 In I2Annual Growth 250% US CMS will use the CDF/KEK remote control room concept for Fermilab Run II
as a starting point. However, we will (1) expand the scope to encompass a US based physics group and US LHC accelerator tasks, and (2) extend the concept to a Global Collaboratory for realtime data acquisition + analysis
Next Round Grid Challenges: Global Workflow Next Round Grid Challenges: Global Workflow Monitoring, Management, and OptimizationMonitoring, Management, and Optimization
Workflow Management, Balancing Policy Versus Workflow Management, Balancing Policy Versus Moment-to-moment Capability to Complete TasksMoment-to-moment Capability to Complete Tasks Balance High Levels of Usage of Limited Resources Balance High Levels of Usage of Limited Resources
Against Better Turnaround Times for Priority JobsAgainst Better Turnaround Times for Priority Jobs Goal-Oriented; According to (Yet to be Developed) Goal-Oriented; According to (Yet to be Developed)
MetricsMetrics Maintaining a Global View of Resources and System StateMaintaining a Global View of Resources and System State
Global System Monitoring, Modeling, Quasi-realtime Global System Monitoring, Modeling, Quasi-realtime
simulation; feedback on the Macro- and Micro-simulation; feedback on the Macro- and Micro-ScalesScales
Adaptive Learning: new paradigms for execution Adaptive Learning: new paradigms for execution optimization and Decision Support (eventually optimization and Decision Support (eventually automated)automated)
Grid-enabled User EnvironmentsGrid-enabled User Environments
PACI, TeraGrid and HENPPACI, TeraGrid and HENP The scale, complexity and global extent of the LHC Data The scale, complexity and global extent of the LHC Data
Analysis problem is unprecedentedAnalysis problem is unprecedented The solution of the problem, using globally distributed Grids, The solution of the problem, using globally distributed Grids,
is mission-critical for frontier science and engineeringis mission-critical for frontier science and engineering HENP has a tradition of deploying new highly functional HENP has a tradition of deploying new highly functional
systems (and sometimes new technologies) to meet its systems (and sometimes new technologies) to meet its technical and ultimately its scientific needstechnical and ultimately its scientific needs
HENP problems are mostly “embarrassingly” parallel; but HENP problems are mostly “embarrassingly” parallel; but potentially “overwhelming” in their data- and network potentially “overwhelming” in their data- and network intensivenessintensiveness
HENP/Computer Science synergy has increased dramatically HENP/Computer Science synergy has increased dramatically over the last two years, focused on Data Gridsover the last two years, focused on Data Grids Successful collaborations in GriPhyN, PPDG, EU Data GridSuccessful collaborations in GriPhyN, PPDG, EU Data Grid
The TeraGrid (present and future) and its development program The TeraGrid (present and future) and its development program is scoped at an appropriate level of depth and diversityis scoped at an appropriate level of depth and diversity to tackle the LHC and other “Petascale” problems, to tackle the LHC and other “Petascale” problems,
over a 5 year time span over a 5 year time span matched to the LHC time schedule, with full ops. In 2007matched to the LHC time schedule, with full ops. In 2007
Some Extra Some Extra
Slides FollowSlides Follow
Computing Challenges: Computing Challenges: LHC ExampleLHC Example
Geographical dispersion:Geographical dispersion: of people and resources of people and resources Complexity:Complexity: the detector and the LHC environment the detector and the LHC environment Scale: Scale: Tens of Petabytes per year of dataTens of Petabytes per year of data
5000+ Physicists 250+ Institutes 60+ Countries
Major challenges associated with:Major challenges associated with:Communication and collaboration at a distanceCommunication and collaboration at a distance
Network-distributed computing and data resources Network-distributed computing and data resources Remote software development and physics analysisRemote software development and physics analysisR&D: New Forms of Distributed Systems: Data GridsR&D: New Forms of Distributed Systems: Data Grids
Why Worldwide Computing? Why Worldwide Computing? Regional Center Concept GoalsRegional Center Concept Goals
Managed, fair-shared access for Physicists everywhereManaged, fair-shared access for Physicists everywhere Maximize total funding resources while meeting the Maximize total funding resources while meeting the
total computing and data handling needstotal computing and data handling needs Balance proximity of datasets to large central resources, Balance proximity of datasets to large central resources,
against regional resources under more local controlagainst regional resources under more local control Tier-N ModelTier-N Model
Efficient network use: higher throughput on short pathsEfficient network use: higher throughput on short paths Local > regional > national > internationalLocal > regional > national > international
Utilizing all intellectual resources, in several time zonesUtilizing all intellectual resources, in several time zones CERN, national labs, universities, remote sitesCERN, national labs, universities, remote sites Involving physicists and students at their home institutionsInvolving physicists and students at their home institutions
Greater flexibility to pursue different physics interests, Greater flexibility to pursue different physics interests, priorities, and resource allocation strategies by regionpriorities, and resource allocation strategies by region
And/or by Common Interests (physics topics, And/or by Common Interests (physics topics, subdetectors,…)subdetectors,…)
Manage the System’s ComplexityManage the System’s Complexity Partitioning facility tasks, to manage and focus resourcesPartitioning facility tasks, to manage and focus resources
HENP Related Data Grid HENP Related Data Grid ProjectsProjects
Funded ProjectsFunded Projects PPDG IPPDG I USAUSA DOEDOE $ 2M$ 2M 1999-20011999-2001 GriPhyNGriPhyN USAUSA NSFNSF $ 11.9M + $1.6M$ 11.9M + $1.6M 2000-20052000-2005 EU DataGridEU DataGrid EUEU ECEC € € 10M10M 2001-20042001-2004 PPDG II (CP)PPDG II (CP) USAUSA DOEDOE $ 9.5M$ 9.5M 2001-20042001-2004 iVDGLiVDGL USAUSA NSFNSF $ 13.7M + $2M$ 13.7M + $2M 2001-20062001-2006 DataTAGDataTAG EUEU ECEC € € 4M4M 2002-20042002-2004
About to be Funded ProjectAbout to be Funded Project GridPPGridPP** UKUK PPARCPPARC >$15M?>$15M? 2001-20042001-2004
Many national projects of interest to HENPMany national projects of interest to HENP Initiatives in US, UK, Italy, France, NL, Germany, Japan, …Initiatives in US, UK, Italy, France, NL, Germany, Japan, … EU networking initiatives (GEU networking initiatives (Gééant, SURFNet)ant, SURFNet) US Distributed Terascale Facility: US Distributed Terascale Facility:
($53M, 12 TFL, 40 Gb/s network)($53M, 12 TFL, 40 Gb/s network)
* = in final stages of approval
Network Progress andNetwork Progress andIssues for Major ExperimentsIssues for Major Experiments
Network backbones are advancing rapidly to the 10 Gbps Network backbones are advancing rapidly to the 10 Gbps range: “Gbps” end-to-end data flows will soon be in demandrange: “Gbps” end-to-end data flows will soon be in demand These advances are likely to have a profound impactThese advances are likely to have a profound impact
on the major physics Experiments’ Computing Models on the major physics Experiments’ Computing Models We need to work on the technical and political network issuesWe need to work on the technical and political network issues
Share technical knowledge of TCP: Windows, Share technical knowledge of TCP: Windows, Multiple Streams, OS kernel issues; Provide User ToolsetMultiple Streams, OS kernel issues; Provide User Toolset
Getting higher bandwidth to regions outside W. Europe and Getting higher bandwidth to regions outside W. Europe and US: China, Russia, Pakistan, India, Brazil, Chile, Turkey, etc.US: China, Russia, Pakistan, India, Brazil, Chile, Turkey, etc. Even to enable their collaborationEven to enable their collaboration
Advanced integrated applications, such as Data Grids, rely onAdvanced integrated applications, such as Data Grids, rely onseamless “transparent” operation of our LANs and WANsseamless “transparent” operation of our LANs and WANs With reliable, quantifiable (monitored), high performanceWith reliable, quantifiable (monitored), high performance Networks need to become part of the Grid(s) designNetworks need to become part of the Grid(s) design New paradigms of network and system monitoringNew paradigms of network and system monitoring
and use need to be developed, in the Grid contextand use need to be developed, in the Grid context
Grid-Related R&D Projects in CMS: Grid-Related R&D Projects in CMS: Caltech, FNAL, UCSD, UWisc, UFlCaltech, FNAL, UCSD, UWisc, UFl
Installation, Configuration and Deployment of Prototype Installation, Configuration and Deployment of Prototype Tier2 Centers at Caltech/UCSD and FloridaTier2 Centers at Caltech/UCSD and Florida
Large Scale Automated Distributed Simulation ProductionLarge Scale Automated Distributed Simulation Production DTF “TeraGrid” (Micro-)Prototype: CIT, Wisconsin DTF “TeraGrid” (Micro-)Prototype: CIT, Wisconsin
Condor, NCSACondor, NCSA Distributed MOnte Carlo Production (MOP): FNALDistributed MOnte Carlo Production (MOP): FNAL
““MONARC” Distributed Systems Modeling;MONARC” Distributed Systems Modeling; Simulation system applications to Grid Hierarchy Simulation system applications to Grid Hierarchy managementmanagement Site configurations, analysis model, workloadSite configurations, analysis model, workload Applications to strategy development; e.g. inter-siteApplications to strategy development; e.g. inter-site
load balancing using a “Self Organizing Neural Net” load balancing using a “Self Organizing Neural Net” (SONN)(SONN)
Agent-based System Architecture for DistributedAgent-based System Architecture for DistributedDynamic ServicesDynamic Services
Grid-Enabled Object Oriented Data AnalysisGrid-Enabled Object Oriented Data Analysis
MONARC Simulation System ValidationMONARC Simulation System Validation
CMS Proto-Tier1 Production Farm at FNAL
Mean measured Value ~48MB/sMeasurement
SimulationJet
<0.52>
Muon<0.90>
CMS Farm at CERN
MONARC SONN: 3 Regional Centres MONARC SONN: 3 Regional Centres
Learning to Export Jobs (Day 0)Learning to Export Jobs (Day 0)
Day = 0
NUST20 CPUs
CERN30 CPUs
CALTECH25 CPUs
1MB/s ; 150 ms RTT
1.2 MB/s
150 ms RTT 0.8
MB/s
200 m
s RTT
US CMS Remote Control RoomUS CMS Remote Control RoomFor LHCFor LHC
Using local Tag event database, user plots event parameters of interest User selects subset of events to be fetched for further analysis Lists of matching events sent to Caltech and San Diego Tier2 servers begin sorting through databases extracting required events For each required event, a new large virtual object is materialized in the server-side cache, this object contains all tracks in the event. The database files containing the new objects are sent to the client using Globus FTP, the client adds them to its local cache of large objects The user can now plot event parameters not available in the Tag Future requests take advantage of previously cached large objects in the client
Full Event Database of
~100,000 large objects
Full Event Database of
~40,000 large objects
“Tag” database
of ~140,000
small objects
Bandwidth Greedy Grid-enabled Object Collection Analysis for Particle Physics (SC2001 Demo)
Julian Bunn, Ian Fisk, Koen Holtman, Harvey Newman, James Patton
RequestRequest
Parallel tuned GSI FTP
Parallel tuned GSI FTP
The object of this demo is to show grid-supported interactive physics analysis on a set of 144,000 physics events.Initially we start out with 144,000 small Tag objects, one for each event, on the Denver client machine. We also
have 144,000 LARGE objects, containing full event data, divided over the two tier2 servers.
http://pcbunn.cacr.caltech.edu/Tier2/Tier2_Overall_JJB.htm