CS 267: Applications of Parallel Computers Grid Computing

1
CS 267 Applications of Parallel
ComputersGrid Computing

• Kathy Yelick
• Material based on lectures by
• Ian Foster, Carl Kesselman

2
Grid Computing is NOT
Slide source Jim Napolitano _at_ RPI
3
Problem Overview

• Many activities in research and education are
  collaborative
• sharing of data and code
• sharing of computing and experimental facilities
• Goal of grids is to simplify these activities
• Approach
• Create advanced middleware services that enable
  large-scale, flexible resource sharing on a
  national and international scale.

Slide source Ian Foster _at_ ANL
4
Outline

• Problem statement What are Grids?
• Architecture and technologies
• Projects
• Future

5
The Grid Problem

• Resource sharing coordinated problem solving
  in dynamic, multi-institutional virtual
  organizations

Slide source Ian Foster _at_ ANL
6
What is a Grid?

• The "Grid problem is defined by Ian Foster as
• flexible, secure, coordinated resource sharing
• among dynamic collections of individuals,
  institutions, and resources (referred to as
  virtual organizations)
• Terminology from analogy to the electric power
  grid
• Computing as a (public) utility
• The key technical challenges in grid computing
• authentication
• authorization
• resource access
• resource discovery
• Grid sometime categorized
• Data grids, computational grids, business grids,

Slide derived from Sara Murphy _at_ HP
7
The evolving notion of the Grid

• A computational grid is a hardware and software
  infrastructure that provides dependable,
  consistent, pervasive, and inexpensive access to
  high-end computing capabilities.
• Ian Foster and Carl Kesselman, editors, The
  GRID Blueprint for a New Computing
  Infrastructure (Morgan-Kaufmann Publishers, SF,
  1999) 677 pp. ISBN 1-55860-8
• The Grid is an infrastructure to enable virtual
  communities to share distributed resources to
  pursue common goals
• The Grid infrastructure consists of protocols,
  application programming interfaces, and software
  development kits to provide authentication,
  authorization, and resource location/access
• Foster, Kesselman, Tuecke The Anatomy of the
  Grid Enabling Scalable Virtual Organizations
  http//www.globus.org/research/papers.html
• The Grid integrates services across distributed,
  heterogeneous, dynamic virtual organizations
  formed from the disparate resources within a
  single enterprise and/or from external resource
  sharing and service provider relationships in
  both e-business and e-science
• Foster, Kesselman, Nick, Tuecke The Physiology
  of the Grid http//www.globus.org/research/papers
  /ogsa.pdf

8
Elements of the Problem

• Resource sharing
• Computers, storage, sensors, networks,
• Sharing always conditional issues of trust,
  policy, negotiation, payment,
• Coordinated problem solving
• Beyond client-server distributed data analysis,
  computation, collaboration,
• Dynamic, multi-institutional virtual
  organizations
• Community overlays on classic org structures
• Large or small, static or dynamic

Slide source Ian Foster _at_ ANL
9
Why Grids?

• A biochemist exploits 10,000 computers to screen
  100,000 compounds in an hour
• 1,000 physicists worldwide pool resources for
  petaflop analyses of petabytes of data
• Civil engineers collaborate to design, execute,
  analyze shake table experiments
• Climate scientists visualize, annotate, analyze
  terabyte simulation datasets
• A home user invokes architectural design
  functions at an application service provider
• Service provider buys cycles from compute cycle
  providers

Slide source Dave Angulo _at_ U Chicago
10
Why Grids Now?
Conventional Wisdom, General Terms

• Deployed internet bandwidth is increasing at a
  faster rate than either CPU speed or memory or
  data storage size
• Therefore, it makes more sense to plan for
  computing that is portable so that it can be
  distributed worldwide.

11
Why Grids Now?

• CPU speed doubles every 18 months
• Moores Law
• Data storage doubles every 12 months
• Deployed network bandwidth doubles every 9 months
• 1986-2001 x340,000
• Gilders Law

• Internet bandwidth increasing faster than CPU
  speed or memory or data storage size
• Therefore, plan for computing that is portable
  so that it can be distributed worldwide.

12
The Grid World Current Status

• Dozens of major Grid projects in scientific
  technical computing/research education
• Deployment, application, technology
• Considerable consensus on key concepts and
  technologies
• Globus Toolkit has emerged as de facto standard
  for major protocols services
• Although there are still competing alternatives
• Global Grid Forum is a significant force
• Cross project commnity

Slide derived from Dave Angulo _at_ U Chicago
13
Science Grids Really Big Science

• The process of Large-Scale Science is changing
• Large-scale science and engineering problems
• require collaborative use of many compute, data,
  and instrument resources all of which must be
  integrated with application components and
• efficient use of large resources is important
• data sets that are
• developed by independent teams of researchers
• or are obtained from multiple instruments
• at different geographic locations

Slide derived from Bill Johnston _at_ LBNL
14
Grid Applications in Physics

• GriPhyN
• CS and Physics collaboration to develop virtual
  data concept
• Physics ATLAS, CMS, LIGO, Sloan Digital Sky
  Survey
• CS build on existing technologies such as
  Globus, Condor, fault tolerance, storage
  management (SRB), plus new computer science
  research
• Main goal is to develop a virtual data catalog
  and data language (VDC, VDL, VDLI)
• iVDGL
• International Virtual Data Grid Laboratory
• Platform to design, implement, integrate and test
  grid software
• Infrastructure for ATLAS and CMS prototype Tier2
  centers
• Forum for grid interoperability collaborate
  with EU DataGrid, DataTag, etc.

Slide source Rob Gardner _at_ Indiana U
15
Data Grids for High Energy Physics
Compact Muon Spectrometer at CERN
PBytes/sec
100 MBytes/sec
Offline Processor Farm 20 TIPSz
There is a bunch crossing every 25 nsecs. There
are 100 triggers per second Each triggered
event is 1 MByte in size
100 MBytes/sec
Tier 0
CERN Computer Centre
622 Mbits/sec
or Air Freight
(deprecated)
Tier 1
FermiLab 4 TIPS
France Regional Centre
Italy Regional Centre
Germany Regional Centre
622 Mbits/sec
Tier 2
622 Mbits/sec
Institute 0.25TIPS
Institute
Institute
Institute
Physics data cache
1 MBytes/sec
Tier 4
Physicist workstations
Slide source Ian Foster _at_ ANL
Image courtesy Harvey Newman, Caltech
16
Network for Earthquake Eng. Simulation

• NEESgrid national infrastructure to couple
  earthquake engineers with experimental
  facilities, databases, computers, each other
• On-demand access to experiments, data streams,
  computing, archives, collaboration

Slide source Ian Foster _at_ ANL
NEESgrid Argonne, Michigan, NCSA, UIUC, USC
17
What is a Grid Architecture?

• Descriptive
• Provide a common vocabulary for use when
  describing Grid systems
• Guidance
• Identify key areas in which services are required
  
• Prescriptive
• Define standard Intergrid protocols and APIs to
  facilitate creation of interoperable Grid systems
  and portable applications

Slide source Ian Foster _at_ ANL
18
What Sorts of Standards?

• Need for interoperability when different groups
  want to share resources
• E.g., IP lets me talk to your computer, but how
  do we establish maintain sharing?
• How do I discover, authenticate, authorize,
  describe what I want to do, etc., etc.?
• Need for shared infrastructure services to avoid
  repeated development, installation, e.g.
• One port/service for remote access to computing,
  not one per tool/application
• X.509 enables sharing of Certificate Authorities

Slide source Ian Foster _at_ ANL
19
A Grid Architecture Must Address

• Development of Grid protocols services
• Protocol-mediated access to remote resources
• New services e.g., resource brokering
• On the Grid speak Intergrid protocols
• Mostly (extensions to) existing protocols
• Development of Grid APIs SDKs
• Facilitate application development by supplying
  higher-level abstractions
• The model is the Internet and Web

Slide source Ian Foster _at_ ANL
20
Grid Services (aka Middleware) and Tools
net
Slide source Ian Foster _at_ ANL
21
Layered Grid Architecture
A Grid architecture uses layers, analogous to IP
layers
Application
Slide source Ian Foster _at_ ANL
22
Where Are We With Architecture?

• No official standards exist
• But
• Globus Toolkit has emerged as the de facto
  standard for several important Connectivity,
  Resource, and Collective protocols
• GGF has an architecture working group
• Technical specifications are being developed for
  architecture elements e.g., security, data,
  resource management, information
• Internet drafts submitted in security area

Slide source Ian Foster _at_ ANL
23
Grid Services Architecture (2) Connectivity Layer

• Communication
• Internet protocols IP, DNS, routing, etc.
• Security Grid Security Infrastructure (GSI)
• Uniform authentication authorization mechanisms
  in multi-institutional setting
• Single sign-on, delegation, identity mapping
• Public key technology, SSL, X.509, GSS-API
  (several Internet drafts document extensions)
• Supporting infrastructure Certificate
  Authorities, key management, etc.

Slide source Ian Foster _at_ ANL
24
GSI in Action Create Processes at A and B that
Communicate Access Files at C
User
Site B (Unix)
Site A (Kerberos)
Computer
Computer
Site C (Kerberos)
Storage system
Slide source Ian Foster _at_ ANL
25
Grid Services Architecture (3) Resource Layer

• Resource Layer has Protocols and Services
• Resource management GRAM
• Remote allocation, reservation, monitoring,
  control of compute resources
• Data access GridFTP
• High-performance data access transport
• Information MDS (GRRP, GRIP)
• Access to structure state information
• others emerging catalog access, code
  repository access, accounting,
• All integrated with GSI

Slide source Ian Foster _at_ ANL
26
GRAM Resource Management Protocol

• Grid Resource Allocation Management
• Allocation, monitoring, control of computations
• Simple HTTP-based RPC
• Job request Returns opaque, transferable job
  contact string for access to job
• Job cancel, Job status, Job signal
• Event notification (callbacks) for state changes
• Protocol/server address robustness (exactly once
  execution), authentication, authorization
• Servers for most schedulers C and Java APIs

Slide source Ian Foster _at_ ANL
27
Resource Management

• Advance reservations
• As prototyped in GARA in previous 2 years
• Multiple resource types
• Manage anything storage, networks, etc., etc.
• Recoverable requests, timeout, etc.
• Build on early work with Condor group
• Use of SOAP (RPC using HTTP XML)
• First step towards Web Services
• Policy evaluation points for restricted proxies

Slide source Ian Foster _at_ ANL
Karl Czajkowski, Steve Tuecke, others
28
Data Access Transfer

• GridFTP extended version of popular FTP protocol
  for Grid data access and transfer
• Secure, efficient, reliable, flexible,
  extensible, parallel, concurrent, e.g.
• Third-party data transfers, partial file
  transfers
• Parallelism, striping (e.g., on PVFS)
• Reliable, recoverable data transfers
• Reference implementations
• Existing clients and servers wuftpd, nicftp
• Flexible, extensible libraries

Slide source Ian Foster _at_ ANL
29
Grid Services Architecture (4) Collective Layer

• Index servers aka metadirectory services
• Custom views on dynamic resource collections
  assembled by a community
• Resource brokers (e.g., Condor Matchmaker)
• Resource discovery and allocation
• Replica management and replica selection
• Optimize aggregate data access performance
• Co-reservation and co-allocation services
• End-to-end performance
• Etc.

Slide source Ian Foster _at_ ANL
30
The Grid Information Problem

• Large numbers of distributed sensors with
  different properties
• Need for different views of this information,
  depending on community membership, security
  constraints, intended purpose, sensor type

Slide source Ian Foster _at_ ANL
31
The Globus Toolkit Solution MDS-2

• Registration enquiry protocols, information
  models, query languages
• Provides standard interfaces to sensors
• Supports different directory structures
  supporting various discovery/access strategies

Slide source Ian Foster _at_ ANL
Karl Czajkowski, Steve Fitzgerald, others
32
GriPhyN/PPDG Data Grid Architecture
Application
initial solution is operational
DAG
Catalog Services
Monitoring
Planner
Info Services
Repl. Mgmt.
DAG
Executor
Policy/Security
Reliable Transfer Service
Compute Resource
Storage Resource
Slide source Ian Foster _at_ ANL
Ewa Deelman, Mike Wilde, others
www.griphyn.org
33
The Network Weather Service

• A distributed system for producing short-term
  deliverable performance forecasts
• Goal dynamically measure and forecast the
  performance deliverable at the application level
  from a set of network resources
• Measurements currently supported
• Available fraction of CPU time
• End-to-end TCP connection time
• End-to-end TCP network latency
• End-to-end TCP network bandwidth

Slide source Rich Wolski _at_ UCSB
34
NWS System Architecture

• Design objectives
• Scalability scales to any metacomputing
  infrastructure
• Predictive accuracy provides accurate
  measurements and forecasts
• Non-intrusiveness shouldnt load the resources
  it monitors
• Execution longevity available all time
• Ubiquity accessible from everywhere, monitors
  all resources

Slide source Rich Wolski _at_ UCSB
35
System Components

• Four different component processes
• Persistent State process handles storage of
  measurements
• Name Server process directory server for the
  system
• Sensor processes measure current performance of
  different resources
• Forecaster process predicts deliverable
  performance of a resource during a given time

Slide source Rich Wolski _at_ UCSB
36
NWS Processes
Slide source Rich Wolski _at_ UCSB
37
NWS Components

• Persistent State Management
• Naming Server
• Performance Monitoring NWS Sensors
• CPU Sensor
• Network Sensor
• Sensor Control
• Cliques hierarchy and contention
• Adaptive time-out discovery
• Forecasting
• Forecaster and forecasting models
• Sample forecaster results

Slide source Rich Wolski _at_ UCSB
38
Persistent State Management

• All NWS processes are stateless
• The system state (measurements) are managed by
  the PS process
• Storage retrieval of measurements
• Measurements are time-stamped plain-text strings
• Measurements are written to disk immediately and
  acknowledged
• Measurements are stored in a circular queue of
  tunable size

Slide source Rich Wolski _at_ UCSB
39
Naming Server

• Primitive text string directory service for the
  NWS system
• The only component known system-wide
• Information stored include
• Name to IP binding information
• Group configuration
• Parameters for various processes
• Each process must refresh its registration with
  the name server periodically
• Centralized

Slide source Rich Wolski _at_ UCSB
40
Performance Monitoring

• Actual monitoring is performed by a set of
  sensors
• Accuracy vs. Intrusiveness
• A sensors life

Register with the NS Query the NS for
parameters Generate conditional test Forever
if conditions are met then perform
test time-stamp results and send them to the
PS refresh registration with the NS
Slide source Rich Wolski _at_ UCSB
41
CPU Sensor

• Measures available CPU fraction
• Testing tools
• Unix uptime reports load average in the past x
  minutes
• Unix vmstat reports idle-, user- and system-time
• Active probes
• Accuracy
• Results assume a full priority job
• Doesnt know the priority of jobs in the queue

Slide source Rich Wolski _at_ UCSB
42
Active Probing Improvements
Measurements produced using vmstat
Measurements produced using uptime
Slide source Rich Wolski _at_ UCSB
43
Network Sensor

• Carries network-related measurements
• Testing using active network probes
• Establish and release TCP connections
• Moving large (small) data to measure bandwidth
  (delay)
• Measures connections with all peer sensors
• Problems
• Accuracy depends on socket interface
• Complexity N2-N tests, collisions (contention)

Slide source Rich Wolski _at_ UCSB
44
Network Sensor Control

• Sensors are organized into sensor sets called
  cliques
• Each clique is configurable and has one leader
• Clique sets are logical, but can be based on
  physical topology
• Leaders are elected using a distributed election
  protocol
• A sensor can participate in many cliques
• Advantages
• Scalability by organizing cliques in a hierarchy
• Reduce the N2-N
• Accuracy by more frequent tests

Slide source Rich Wolski _at_ UCSB
45
Clique Hierarchy
Slide source Rich Wolski _at_ UCSB
46
Contention

• Each leader maintains a clique token (and time
  between tokens)
• The sensor that has the token performs all its
  tests then passes the token to the next sensor in
  the list
• Adaptive time-out discovery
• Tokens have time-out field
• Tokens have sequence numbers
• The leader adaptively controls the time-out

Slide source Rich Wolski _at_ UCSB
47
Forecaster Process

• A forecasting driver compile-time prediction
  modules
• Forecasting process
• Fetching required measurements from the PS
• Passing the time series to each prediction module
• Choosing the best returned prediction
• Incorporate sophisticated prediction techniques?

UC Santa Barbara Kansas State U. Recorded
Bandwidth
UC Santa Barbara Kansas State U. Forecasted
Bandwidth
Slide source Rich Wolski _at_ UCSB
48
Sample Graph
49
Selected Major Grid Projects
Name URL Sponsors Focus
Access Grid www.mcs.anl.gov/FL/accessgrid DOE, NSF Create deploy group collaboration systems using commodity technologies
BlueGrid IBM Grid testbed linking IBM laboratories
DISCOM www.cs.sandia.gov/discomDOE Defense Programs Create operational Grid providing access to resources at three U.S. DOE weapons laboratories
DOE Science Grid sciencegrid.org DOE Office of Science Create operational Grid providing access to resources applications at U.S. DOE science laboratories partner universities
Earth System Grid (ESG) earthsystemgrid.orgDOE Office of Science Delivery and analysis of large climate model datasets for the climate research community
European Union (EU) DataGrid eu-datagrid.org European Union Create apply an operational grid for applications in high energy physics, environmental science, bioinformatics
New
New
Slide source Ian Foster _at_ ANL
50
Selected Major Grid Projects
Name URL/Sponsor Focus
EuroGrid, Grid Interoperability (GRIP) eurogrid.org European Union Create technologies for remote access to supercomputer resources simulation codes in GRIP, integrate with Globus
Fusion Collaboratory fusiongrid.org DOE Off. Science Create a national computational collaboratory for fusion research
Globus Project globus.org DARPA, DOE, NSF, NASA, Msoft Research on Grid technologies development and support of Globus Toolkit application and deployment
GridLab gridlab.org European Union Grid technologies and applications
GridPP gridpp.ac.uk U.K. eScience Create apply an operational grid within the U.K. for particle physics research
Grid Research Integration Dev. Support Center grids-center.org NSF Integration, deployment, support of the NSF Middleware Infrastructure for research education
New
New
New
New
New
Slide source Ian Foster _at_ ANL
51
Selected Major Grid Projects
Name URL/Sponsor Focus
Grid Application Dev. Software hipersoft.rice.edu/grads NSF Research into program development technologies for Grid applications
Grid Physics Network griphyn.org NSF Technology RD for data analysis in physics expts ATLAS, CMS, LIGO, SDSS
Information Power Grid ipg.nasa.gov NASA Create and apply a production Grid for aerosciences and other NASA missions
International Virtual Data Grid Laboratory ivdgl.org NSF Create international Data Grid to enable large-scale experimentation on Grid technologies applications
Network for Earthquake Eng. Simulation Grid neesgrid.org NSF Create and apply a production Grid for earthquake engineering
Particle Physics Data Grid ppdg.net DOE Science Create and apply production Grids for data analysis in high energy and nuclear physics experiments
New
New
Slide source Ian Foster _at_ ANL
52
Selected Major Grid Projects
Name URL/Sponsor Focus
TeraGrid teragrid.org NSF U.S. science infrastructure linking four major resource sites at 40 Gb/s
UK Grid Support Center grid-support.ac.uk U.K. eScience Support center for Grid projects within the U.K.
Unicore BMBFT Technologies for remote access to supercomputers
New
New
Also many technology RD projects e.g., Condor,
NetSolve, Ninf, NWS See also www.gridforum.org
Slide source Ian Foster _at_ ANL
53
The 13.6 TF TeraGridComputing at 40 Gb/s
Site Resources
Site Resources
26
HPSS
HPSS
4
24
External Networks
External Networks
8
5
Caltech
Argonne
External Networks
External Networks
NCSA/PACI 8 TF 240 TB
SDSC 4.1 TF 225 TB
Site Resources
Site Resources
HPSS
UniTree
Slide source Ian Foster _at_ ANL
TeraGrid/DTF NCSA, SDSC, Caltech, Argonne
www.teragrid.org
54
iVDGLInternational Virtual Data Grid Laboratory
Slide source Ian Foster _at_ ANL
U.S. PIs Avery, Foster, Gardner, Newman, Szalay
www.ivdgl.org
55
NSF GRIDS Center

• Grid Research, Integration, Deployment, Support
  (GRIDS) Center
• Develop, deploy, support
• Middleware infrastructure for national-scale
  collaborative science and engineering
• Integration platform for experimental middleware
  technologies
• UC, USC/ISI, UW, NCSA, SDSC
• Partner with Internet-2, SURA, Educause in NSF
  Middleware Initiative

Slide source Ian Foster _at_ ANL
www.grids-center.org
www.nsf-middleware-org
56
The State of Grids Some Case Studies

• Further, Grids are becoming a critical element of
  many projects e.g.
• The High Energy Physics problem of managing and
  analyzing petabytes of data per year has driven
  the development of Grid Data Services
• The National Earthquake Engineering Simulation
  Grid has developed a highly application oriented
  approach to using Grids
• The Astronomy data federation problem has
  promoted work in Web Services based interfaces

57
High Energy Physics Data Management

• Petabytes of data per year must be distributed to
  hundreds of sites around the world for analysis
• This involves
• Reliable, wide-area, high-volume data management
• Global naming, replication, and caching of
  datasets
• Easily accessible pools of computing resources
• Grids have been adopted as the infrastructure for
  this HEP data problem

58
High Energy Physics Data Management CERN / LHC
Data One of Sciences most challenging data
management problems
100 MBytes/sec
event simulation
Online System
PByte/sec
Tier 0 1
eventreconstruction
human2m
HPSS
CERN LHC CMS detector 15m X 15m X 22m, 12,500
tons, 700M.
2.5 Gbits/sec
Tier 1
German Regional Center
French Regional Center
FermiLab, USA Regional Center
Italian Center
0.6-2.5 Gbps
analysis
Tier 2
0.6-2.5 Gbps
Tier 3
CERN/CMS data goes to 6-8 Tier 1 regional
centers, and from each of these to 6-10 Tier 2
centers. Physicists work on analysis channels
at 135 institutes. Each institute has 10
physicists working on one or more channels. 2000
physicists in 31 countries are involved in this
20-year experiment in which DOE is a major player.
Institute 0.25TIPS
Institute
Institute
Institute
100 - 1000 Mbits/sec
Physics data cache
Tier 4
Courtesy Harvey Newman, CalTech
Workstations
59
High Energy Physics Data Management

• Virtual data catalogues and on-demand data
  generation have turned out to be an essential
  aspect
• Some types of analysis are pre-defined and
  catalogued prior to generation - and then the
  data products are generated on demand when the
  virtual data catalogue is accessed
• Sometimes regenerating derived data is faster and
  easier than trying to store and/or retrieve that
  data from remote repositories
• For similar reasons this is also of great
  interest to the EOS (Earth Observing Satellite)
  community

60
US-CMS/LHC Grid Data Services TestbedInternation
al Virtual Data Grid Laboratory
metadatadescriptionof analyzeddata
Interactive User Tools
Data GenerationRequestExecution Management
Tools
Data Generation RequestPlanning Scheduling
Tools
Virtual Data Tools

• Metadata catalogues
• Virtual data catalogues

Security andPolicy
Other GridServices
ResourceManagement
Core Grid Services
Transforms
Distributed resources(code, storage,
CPUs,networks)
Raw datasource
61
CMS Event Simulation Using GriPhyN

• Production Run on Integration Testbed (400 CPUs
  at 5 sites)
• Simulate 1.5 million full CMS events for physics
  studies
• 2 months continuous running across 5 testbed
  sites
• Managed by a single person at the US-CMS Tier
  1site
• 30 CPU years delivered 1.5 Million Events to
  CMS Physicists

62
National Earthquake Engineering Simulation Grid

• NEESgrid will link earthquake researchers across
  the U.S. with leading-edge computing resources
  and research equipment, allowing collaborative
  teams (including remote participants) to plan,
  perform, and publish their experiments
• Through the NEESgrid, researchers will
• perform tele-observation and tele-operation of
  experiments shake tables, reaction walls, etc.
  
• publish to, and make use of, a curated data
  repository using standardized markup
• access computational resources and open-source
  analytical tools
• access collaborative tools for experiment
  planning, execution, analysis, and publication

63
NEES Sites

• Large-Scale Laboratory Experimentation Systems
• University at Buffalo, State University of New
  York
• University of California at Berkeley
• University of Colorado, Boulder
• University of Minnesota-Twin Cities
• Lehigh University
• University of Illinois, Urbana-Champaign
• Field Experimentation and Monitoring
  Installations
• University of California, Los Angeles
• University of Texas at Austin
• Brigham Young University

• Shake Table Research Equipment
• University at Buffalo, State University of New
  York
• University of Nevada, Reno
• University of California, San Diego
• Centrifuge Research Equipment
• University of California, Davis
• Rensselaer Polytechnic Institute
• Tsunami Wave Basin
• Oregon State University, Corvallis, Oregon
• Large-Scale Lifeline Testing
• Cornell University

64
NEESgrid Earthquake Engineering Collaboratory
Instrumented Structures and Sites
Remote Users
Simulation Tools Repository
High-Performance Network(s)
Laboratory Equipment
Field Equipment
Curated Data Repository
Large-scale Computation
Global Connections
Remote Users (K-12 Faculty and Students)
Laboratory Equipment
65
NEESgrid Approach

• Package a set of application level services and
  the supporting Grid software in a singlepoint
  of presence (POP)
• Deploy the POP to a select set of earthquake
  engineering sites to provide the applications,
  data archiving, and Grid services
• Assist in developing common metadata so that the
  various instruments and simulations can work
  together
• Provide the required computing and data storage
  infrastructure

66
NEESgrid Multi-Site Online Simulation (MOST)

• A partnership between the NEESgrid team, UIUC and
  Colorado Equipment Sites to showcase NEESgrid
  capabilities
• A large-scale experiment conducted in multiple
  geographical locations which combines physical
  experiments with numerical simulation in an
  interchangeable manner
• The first integration of NEESgrid services with
  application software developed by Earthquake
  Engineers (UIUC, Colorado and USC) to support a
  real EE experiment
• See http//www.neesgrid.org/most/

67
NEESgrid Multi-Site Online Simulation (MOST)
UIUC Experimental Setup
U. Colorado Experimental Setup
68
Multi-Site, On-Line Simulation Test (MOST)
Colorado Experimental Model
UIUC Experimental Model
SIMULATION COORDINATOR

• UIUC MOST-SIM
• Dan Abrams
• Amr Elnashai
• Dan Kuchma
• Bill Spencer
• and others
• Colorado FHT
• Benson Shing
• and others

NCSA Computational Model
69
1994 Northridge Earthquake SimulationRequires a
Complex Mix of Data and Models
Pier 7
Pier 5
Pier 8
Pier 6
NEESgrid provides the common data formats,
uniform dataarchive interfaces, and
computational services needed to supportthis
multidisciplinary simulation
Amr Elnashai, UIUC
70
NEESgrid Architecture
Java Applet
Web Browser
User Interfaces
MultidisciplinarySimulations
Collaborations
Experiments
Curated Data Repository
Simulation Tools Repository
SIMULATION COORDINATOR
Data AcquisitionSystem
NEESpop
NEES Operations
E-Notebook Services
Metadata Services
CompreHensive collaborativE Framework (CHEF)
NEESgrid Monitoring
Video Services
GridFTP
NEESGrid StreamingData System
Accounts MyProxy
Grid Services
NEES distributed resources
Instrumented Structures and Sites
Large-scale Storage
Large-scale Computation
Laboratory Equipment
71
The Changing Face of Observational Astronomy

• Large digital sky surveys are becoming the
  dominant source of data in astronomy gt 100 TB,
  growing rapidly
• Current examples SDSS, 2MASS, DPOSS, GSC,
  FIRST, NVSS, RASS, IRAS CMBR experiments
  Microlensing experiments NEAT, LONEOS, and other
  searches for Solar system objects
• Digital libraries ADS, astro-ph, NED, CDS, NSSDC
• Observatory archives HST, CXO, space and
  ground-based
• Future QUEST2, LSST, and other synoptic surveys
  GALEX, SIRTF, astrometric missions, GW detectors
• Data sets orders of magnitude larger, more
  complex, and more homogeneous than in the past

72
The Changing Face of Observational Astronomy

• Virtual Observatory Federation of N archives
• Possibilities for new discoveries grow as O(N2)
• Current sky surveys have proven this
• Very early discoveries from Sloan (SDSS),2
  micron (2MASS), Digital Palomar (DPOSS)
• see http//www.us-vo.org

73
Sky Survey Federation
74
Mining Data is Often a Critical Aspect of Doing
Science

• The ability to federate survey data is enormously
  important
• Studying the Cosmic Microwave Background a key
  tool in studying the cosmology of the universe
  requires combined observations from many
  instruments in order to isolate the extremely
  weak signals of the CMB
• The datasets that represent the material
  between us and the CMB are collected from
  different instruments and are stored and curated
  at many different institutions
• This is immensely difficult without approaches
  like National Virtual Observatory in order to
  provide a uniform interface for all of the
  different data formats and locations

(Julian Borrill, NERSC, LBNL)
75
NVO Approach

• Focus is on adapting emerging information
  technologies to meet the astronomy research
  challenges
• Metadata, standards, protocols (XML, http)
• Interoperability
• Database federation
• Web Services (SOAP, WSDL, UDDI)
• Grid-based computing (OGSA)
• Federating data bases is difficult, but very
  valuable
• An XML-based mark-up for astronomical tables and
  catalogs - VOTable
• Developed metadata management framework
• Formed international registry, dm (data
  models), semantics, and dal (data access
  layer) discussion groups
• As with NEESgrid, Grids are helping to unify the
  community

76
NVO Image Mosaicking

• Specify box by position and size
• SIAP server returns relevant images
• Footprint
• Logical Name
• URL

Can choose standard URL http//....... SRB
URL srb//nvo.npaci.edu/..
77
Atlasmaker Virtual Data System
Metadata repositories Federated by OAI
Higher LevelGrid Services
Data repositories Federated by SRB
2d Store result return result
Core Grid Services
2c Compute on TG/IPG
Compute resources Federated by TG/IPG
78
Background Correction
Uncorrected
Corrected
79
NVO Components
Visualization
Resource/Service Registries
Web Services
Simple Image Access Services
Cone Search Services
VOTable
VOTable
Cross-Correlation Engine
UCDs
UCDs
Streaming
Grid Services
Data archives
Computing resources
80
International Virtual Observatory Collaborations

• German AVO
• Russian VO
• e-Astronomy Australia
• IVOA(International Virtual Observatory
  Alliance)

• Astrophysical Virtual Observatory (European
  Commission)
• AstroGrid, UK e-scienceprogram
• Canada
• VO India
• VO Japan
• (leading the work on VO query language)
• VO China

US contacts Alex Szalay szalay_at_jhu.edu, Roy
Williams roy_at_cacr.caltech.edu,Bob Hanisch
lthanisch_at_stsci.edugt
81
And Whats This Got To Do With

• CORBA?
• Grid-enabled CORBA underway
• Java, Jini, Jxta?
• Java CoG Kit. Jini, Jxta future uncertain
• Web Services, .NET, J2EE?
• Major Globus focus (GRAM-2 SOAP, WSDL)
• Workflow/choreography services
• Q What can Grid offer to Web services?
• Next revolutionary technology of the month?
• Theyll need Grid technologies too

Slide source Ian Foster _at_ ANL
82
The Future All Software is Network-Centric

• We dont build or buy computers anymore, we
  borrow or lease required resources
• When I walk into a room, need to solve a problem,
  need to communicate
• A computer is a dynamically, often
  collaboratively constructed collection of
  processors, data sources, sensors, networks
• Similar observations apply for software

Slide source Ian Foster _at_ ANL
83
And Thus

• Reduced barriers to access mean that we do much
  more computing, and more interesting computing,
  than today gt Many more components ( services)
  massive parallelism
• All resources are owned by others gt Sharing (for
  fun or profit) is fundamental trust, policy,
  negotiation, payment
• All computing is performed on unfamiliar systems
  gt Dynamic behaviors, discovery, adaptivity,
  failure

Slide source Ian Foster _at_ ANL
84
Acknowledgments

• Globus RD is joint with numerous people
• Carl Kesselman, Co-PI
• Steve Tuecke, principal architect at ANL
• Others to be acknowledged
• GriPhyN RD is joint with numerous people
• Paul Avery, Co-PI Newman, Lazzarini, Szalay
• Mike Wilde, project coordinator
• Carl Kesselman, Miron Livny CS leads
• ATLAS, CMS, LIGO, SDSS participants others
• Support DOE, DARPA, NSF, NASA, Microsoft

Slide source Ian Foster _at_ ANL
85
Summary

• The Grid problem Resource sharing coordinated
  problem solving in dynamic, multi-institutional
  virtual organizations
• Grid architecture Emphasize protocol and service
  definition to enable interoperability and
  resource sharing
• Globus Toolkit a source of protocol and API
  definitions, reference implementations
• See globus.org, griphyn.org, gridforum.org,
  grids-center.org, nsf-middleware.org

Slide source Ian Foster _at_ ANL