Holding slide prior to starting show

1
Holding slide prior to starting show
2
e-Science and Grid Computing- Raising Awareness
Professor David WalkerExecutive Director
3
Schedule

• Welcome
• e-Science and the Welsh e-Science Centre
• Grid Technology
• Immersive Visualization
• High Performance Computing
• Tea / coffee break
• Presentation by SGI
• Demonstration
• e-Science Potential
• Close

4
Schedule

• Welcome
• e-Science and the Welsh e-Science Centre
• Grid Technology
• Immersive Visualization
• High Performance Computing
• Tea / coffee break
• Presentation by SGI
• Demonstration
• e-Science Potential
• Close

5
e-Science and theWelsh e-Science Centre Alex
HardistyGrid Centre Manager
6
Synopsis

• What is e-Science?
• UK e-Science Programme
• Role of the Welsh e-Science Centre
• Concluding remarks

7
Collaborative Scientific Experiments
Large Hadron Collider (CERN) raw data rate 1
Petabyte/sec Filtered rate 100Mbyte/sec (1
Petabyte/year) 1 Million CD
ROMs ( 200m3!)

• Physicists collaborating in an international
  experiment need to share
• Experimental data and storage resources
• Computers and software for extracting information
  from this data
• Computers and software for large-scale computer
  simulations

8
Engineering Design
A new aircraft may involve 10,000 collaborating
engineers

• Collaborating organisations need to share
• Digital blueprints of the design
• Supercomputers for performing multi-disciplinary
  simulations
• Software and data for performing those simulations

9
Post-Genomic Bioinformatics
Wales Gene Park

• Increasing size, complexity and diversity of
  databases of biological information
• Linkage and optimal exploitation
• Filtering results of micro-array experiments
• 30 or 40 results out of 1 million
• Simulation of large molecules
• Protein folding affects how drugs dock with
  receptors

10
Elements in Common

• Coordinated problem solving
• Beyond client-server distributed data analysis,
  computation, collaboration,
• Problem Solving Environments
• Resource sharing
• Computers, data, instruments, networks
• Multi-institutional virtual organisations
• Overlying traditional organisational structures
• Large or small, static or dynamic

11
e-Science

• science increasingly done through distributed
  global collaborations enabled by the Internet,
  using very large data collections, tera-scale
  computing resources and high performance
  visualization.

12
The Computing Foundationsfor e-Science

• Grid computing technology
• Immersive visualization
• High performance computing
• High speed networking broadband access
• Collaboration tools

13
UK e-Science Programme

• Spending Reviews
• 2000 120m for 3 years
• 2002 Further 115m for years 4 5
• Development of key IT infrastructure to support
  e-Science
• Managed by Research Councils DTI
• Application specific Pilot Projects
• Core programme to identify and develop generic
  Grid middleware

14
UK e-Science Network

• National Centre in Edinburgh/Glasgow
• 8 regional centres
• Grid support centre

Edinburgh
Glasgow
Newcastle
Belfast
Manchester
DL
Cambridge
Oxford
Hinxton (EBI)
RAL
Cardiff
London
Southampton
15
Welsh e-Science Centre

• Based at Cardiff University
• Department of Computer Science
• Funded by DTI, WDA and CU
• Role
• Promote e-Science research and development in
  Wales and South-west of England
• Accelerate the adoption of e-Science Grid
  capabilities
• http//www.wesc.ac.uk/

16
Our Role in Practice

• Provision of infrastructure
• Development of the technology
• Outreach to encourage
• Use of e-Science technologies by researchers
• Collaborative research projects
• Technology transfer to industry

17
Resources

• Monster Computing Power
• Locally SUN, SGI, storage, visualization
• Resources of the national grid !
• Access via Broadband
• Grid expertise much of it free
• Full-time staff (4)
• Related Researchers (20)

18
Training and User Support

• Grid technologies
• Installation troubleshooting
• Java Programming for the Grid (Using the Globus
  Toolkit)
• High Performance Computing
• Parallel Computing, MPI, etc.
• Visualization
• Techniques for visualization
• Using AVS

19
Research Foci

• Data and knowledge management
• Interoperation of heterogeneous distributed
  information resources
• Techniques to improve accessibility to data
• Problem-solving environments
• To provide assistance in formulating and solving
  problems in specific domains e.g., bio-molecular
  modelling
• Distributed visualization
• Collaborative 3-D immersive visualization

20
Projects

• Collaborative virtual teams
• Resource-aware visualization environment
• Visual workflow composition environment
• Workflow execution engine and resource management
• Agents negotiating to form consortia
• Mechanisms for trusted Grid services

21
Status of e-Science Grid

• Level 1
• Skeleton Grid operational now
• Load-balanced work distribution
• Level 2
• Operational April 2003
• National Grid Information Service
• Simple tools for user management
• Local accounting information
• Testbed applications

22
Concluding Remarks

• Vision of e-Science and the Grid is ambitious and
  far-reaching
• This is the start of the Grid era. Its a long
  term programme
• Significant commercial benefits to be gained
• WeSC exists to help scientists and industrialists
  use this new technology

23
(No Transcript)
24
Schedule

• Welcome
• e-Science and the Welsh e-Science Centre
• Grid Technology
• Immersive Visualization
• High Performance Computing
• Tea / coffee break
• Presentation by SGI
• Demonstration
• e-Science Potential
• Close

25
Grid Technology Jonathan GiddyGrid
Technologies Coordinator
26
What is a Grid?

• A Grid is
• a hardware and software infrastructure that
  provides dependable, consistent, and pervasive
  access to resources
• to enable sharing of computational resources,
  utility computing, autonomic computing,
  collaboration among virtual organizations, and
  distributed data processing, among others

Wolfgang Gentzsch, Director Grid Computing, Sun
Microsystems
27
The electric power grid
28
Virtual Organisation
Organisation A
Organisation B
Organisation C
Engine Data
Wing Data
29
More processors, more administrators, more
accounts, more filespaces, more
Grid computing
Distributed computing
Cluster computing
Single node computing
Less centralization, less reliability, less
trivially scalable, less secure,
30
How to handle heterogeneity

• Hardwire code to known infrastructure
• Code for the lowest common denominator
• Make application adapt to available resources

31
Grid computing

• Make infrastructure adapt to application
  requirements
• Ask for resources by properties rather than by
  name!
• Although even a name is one type of attribute

Join together a Fish Species Database, a 32-node
Sun machine and a visualization suite within 400
metres of my location
Join the Fish Species Database at poisson.fr to
the 32-node cspace.cardiff.ac.uk, and visualize
the results in room 212
poisson down cspace busy northviz ok
Using fishbase.de Using bigsun.ac.uk Using
northviz
32
Benefits of grid computing

• Lower cost of computing due to improved resource
  utilization
• Better collaboration within widely-dispersed,
  intra- or extra-enterprises, allowing the
  creation of virtual organizations
• Increased flexibility and speed of resource
  deployment to tackle problems of a scale that is
  not feasible today

IBM web page
33
Challenges of grid computing

• Security
• authentication
• authorisation
• accounting
• integrity
• privacy
• Mutual comprehension
• semantics
• protocols
• policy

• Trust
• resource access
• confidence in results
• contract enforcement
• commutability of trust
• Decentralisation
• autonomy
• discovery
• non-trivial quality of service

34
Globus Toolkit

• Middleware providing services useful in a grid
• Common Software Services
• Security Infrastructure
• Secure Communication
• Process Management
• Data Management
• Information Services
• Resource Drivers

35
Applications
High-level Services and Tools
Cactus
Condor-G
DUROC
Nimrod/G
COG
MPI-CH
globusrun
GAT
Globus Core Services
GridFTP
GRAM
MDS
Replica
I/O
GASS
Grid Security Infrastructure
Grid Services
Local Services
sysconf
SSL
SASL
getloadavg
Condor
Sun Grid Engine
LSF
TCP/IP
MPI
36
ActiveSheets
Cactus
Triana
Nimrod/G
Nimrod/G Broker
Grid Application Toolkit
Globus Resource Allocation Manager
37
Application
Application
Condor-G
Globus services
Condor
Condor
LSF
38
Grid Middleware
39
Open Grid Services Architecture

• Developed within the Global Grid Forum (GGF), an
  IETF-like body
• Implements the Grid using standard open protocols
• Based on Web Services (HTTP, XML, SOAP, WSDL,
  UDDI,)
• Grid extensions (service lifetimes, service
  attributes, notification,)

40
What is a Grid?

• An advanced approach to distributed computing
• that allows
• Coordination in or between physically dispersed
  virtual organisations.
• Dependable, consistent, and pervasive
  availability of computers, data, software,
  storage and other resources.
• Controlled access by resource providers and
  consumers, which defines who can share, what is
  shared, and which conditions allow sharing.

Paraphrase of an IBM web page
41
(No Transcript)
42
Schedule

• Welcome
• e-Science and the Welsh e-Science Centre
• Grid Technology
• Immersive Visualization
• High Performance Computing
• Tea / coffee break
• Presentation by SGI
• Demonstration
• e-Science Potential
• Close

43
Immersive Visualization Ieuan
NicholasVisualization Coordinator
44
Immersive Visualization

• 1. Purpose of Visualization (Viz)
• 2. Need for parallel visualization
• 3. Need for grid-based visualization
• 4. Grid visualization service approaches
• 5. Viz hardware resources at WeSC
• 6. Viz software resources at WeSC
• 7. Potential Applications
• 8. Work done at the WeSC Viz facility
• 9. Conclusion

45
1. Purpose of Visualization

• Search for structure, features, patterns, trends,
  anomalies and relationships in data
• Provide a qualitative overview of large and
  complex data sets
• Assist in identifying regions of interest
• Selection of appropriate parameters for more
  focussed qualitative analysis

46
2. Need for parallel visualization

• Large amounts of data
• Computationally intensive rendering
• Interaction

47
3. Need for grid-based visualization

• Current high-end visualization hardware is
  expensive
• Current high-end hardware is not flexible
• The Grid offers an unprecedented opportunity to
  process huge and complex data sets

48
4. Grid visualization service approach 1

• Do everything on the remote server
• Pros straightforward to grid-enable existing
  packages
• Cons high latency
• eliminates possibility of proxying and caching
  at local site

ModerateBandwidthLink(100Mbps)
Local machine
High-end remote server
49
Grid visualization service approach 2

• Do everything but the rendering on the remote
  server
• Pros moderately straightfoward to grid-enable
  existing packages
• removes some load from the remote site
• Cons requires every local site to have good
  rendering power

Computer with good rendering power
High-endremote server
ModerateBandwidthLink(100Mbps)
50
Grid visualization service approach 3

• Use a local proxy for the rendering
• Pros offloads work from the remote site
• allows local sites to contribute additional
  resources
• Cons local sites may not have sufficiently
  powerful proxy resources application is more
  complex requires high bandwidth between local
  and remote sites

High bandwidth link(1Gbps)
Moderate bandwidth link(100Mbps)
Low-end local computer
Local proxy server
High-end remote server
51
Grid visualization service approach 4

• Do everything at the local site
• Pros low latency straightforward to
  grid-enable existing packages
• Cons requires high-bandwidth link between
  sites requires powerful computational and
  graphics resources at the local site

Ultra-high bandwidth link(10Gbps)
Low-end remote server
Powerful local server
52
5. Visualization hardware resources at WeSC

• 32-processor SGI Onyx300 visualization system for
  large shared-memory applications
• Four 8-processor SGI Origin300 servers connected
  through a Myrinet network for large scale
  distributed-memory applications
• 2TB Storage Area Network
• Fakespace Portico Workwall for large scale
  stereoscopic viewing(6x8 display)
• Two Fakespace Immersadesk R2 workbenches (4x6
  display)
• Three SGI Fuel workstations

53
6. Visualization software resources at WeSC

• AVS/Express visualization toolkit

Finite Element model of a car body computed from
NASTRAN
A 3D grid space showing residual engine
combustion chemical species from a fuel infection
numerical experiment
Wind arrows can be animated and shown at multiple
levels. Cloud animation can be enabled to show
changing humidity
courtesy of AVS Ltd
54
7. Potential Applications

• Collaborative, distributed and immersive data
  visualization
• Computational steering
• Directly manipulable 3D CAD systems
• Virtual product prototyping
• Virtual-reality product presentations
• CAVE-based visualization of scientific data
• Virtual manufacturing
• Architecture prototyping
• Augmented reality
• Interactive training systems
• True 3D representations in new environments are
  likely to provide the means to manage
  information flow.

courtesy of PTC Inc
55
8. Work done at the WeSC Viz facility

• Translocation modeling of DNA molecules through a
  membrane pore

56
Visual Molecular Dynamics
57
Computational Steering

• Allows researchers to change simulation
  parameters on the fly and immediately receive
  feedback on the effect

58
9. Conclusion

• Grid computing proposes to create access to
  distributed computing resources with the same
  ease as electrical power. Visualization tools
  can take advantage of grid computing
  environments. New methodologies and techniques
  that harness such resources for graphics and
  visualization applications are crucial for the
  success of grid environments.

59
(No Transcript)
60
Schedule

• Welcome
• e-Science and the Welsh e-Science Centre
• Grid Technology
• Immersive Visualization
• High Performance Computing
• Tea / coffee break
• Presentation by SGI
• Demonstration
• e-Science Potential
• Close

61
High Performance Computing Roger PhilpHPC
Officer
62
What is High Performance Computing (HPC)?

• High Performance computing is considered to be
  computing that requires at least 10 times the
  computing power of the machine sat on your desk.

63
HPC def WeSc integration Fitting into Grid uk
64
What HPC provision is there?
Hardware support
Machine usage support
Software support
Consultancy
Algorithmic
Optimisation/Profiling/Debugging
Analytic
65
Where has the HPC provision come from?
Since 1988 there have been successive
collaborations between
Physics and Astronomy
Civil and structural Engineering
Computing Science
In 1998 this led to
Cardiff Centre for Computational Science and
Engineering
Which is now merging with the newly formed
Welsh eScience
66
HPC at Cardiff

• nCube 1 nCube, 1989, Physics (128p)
  Comsc(32p)
• nCube II nCube, 1992 Physics (32p)
• Paramid Transtech, 1994 (48 x i860 48 x
  T805)
• Power Challenge SGI 1996 (8xR8000)
• CSPACE I SUN, 1997(14 x Ultra sparc IIi)
• AZTEC SUN, 1998 (16 x Ultra sparc IIi)
• CSPACE II SUN, 1998 (30 x Ultra sparc IIi)

67
Primary HPC facility typical usage
68
Who is currently supported by the HPC centre
Physics and Astronomy Engineering Computer
Science Chemistry Biology
69
HPC projects at Cardiff

• Galaxy/Star formation
• Black hole simulation
• Heat and mass transfer processes
• Nuclear and Toxic waste repositories
• Surface Physics
• Quantum and Monte Carlo simulation
• Structural Engineering
• Electrodynamics
• Fluid Dynamics

70
P. Williams and A. Nelson
71
HPC Output
http//www.cs.cf.ac.uk/Hpc-centre
Publications Projects PhDs
60 since 1998
50 since 1998
20 since 1998
72
HPC and Grid Computing
73
UK e-Science Network

• National Centre in Edinburgh/Glasgow
• 8 regional centres
• Grid support centre

Edinburgh
Glasgow
Newcastle
Belfast
Manchester
DL
Cambridge
Oxford
Hinxton (EBI)
RAL
Cardiff
London
Southampton
74
Future plans
New hardware Courses Introduction to High
Performance and Parallel Computing
(HPC) Introduction to MPI I, II,
III F90 Make Optimisation Debugging Tools
75
(No Transcript)
76
Schedule

• Welcome
• e-Science and the Welsh e-Science Centre
• Grid Technology
• Immersive Visualization
• High Performance Computing
• Tea / coffee break
• Presentation by SGI
• Demonstration
• e-Science Potential
• Close

77
Schedule

• Welcome
• e-Science and the Welsh e-Science Centre
• Grid Technology
• Immersive Visualization
• High Performance Computing
• Tea / coffee break
• Presentation by SGI
• Demonstration
• e-Science Potential
• Close

78
Presentation By SGI Andrew Grant, SGI
79
(No Transcript)
80
Schedule

• Welcome
• e-Science and the Welsh e-Science Centre
• Grid Technology
• Immersive Visualization
• High Performance Computing
• Tea / coffee break
• Presentation by SGI
• Demonstration
• e-Science Potential
• Close

81
Demonstration Jonathan Giddy
82
Nimrod/G

• Parameter sweep application
• Script language for describing parameter
  substitution, file transfer and execution
• Best-effort or deadline/cost scheduling
• ActiveSheets Microsoft Excel front-end

83
Triana

• Visual workflow composition environment
• Varied box of generic tools (audio, text and
  image processing, maths functions, signal
  processing)
• Workflow distribution and management engine

84
(No Transcript)
85
(No Transcript)
86
Schedule

• Welcome
• e-Science and the Welsh e-Science Centre
• Grid Technology
• Immersive Visualization
• High Performance Computing
• Tea / coffee break
• Presentation by SGI
• Demonstration
• e-Science Potential
• Close

87
e-Science PotentialWho may benefit? John
OliverCommercial Coordinator
88
Key Questions

• Which organisations/companies might be interested
  in Grid technology?
• In what ways would they be interested?
• What are the commercial benefits?

89
What is Grid Good for?

• Smallpox Research Project Oxford.2 million
  home computers linked, 35 million drug molecule
  interactions
• Physics Research1,000 physicists worldwide pool
  resources to analyse petabytes of data
• Civil engineering Engineers collaborate to
  design, execute, analyse shake table
  experiments
• Climate StudyScientists visualize, annotate,
  analyse terabyte simulation datasets

90
What is Grid Good for?

• Combating bio-terrorismA global team combines
  real time data, weather model and demographic
  data to plan the response to a crisis
• Aerospace DevelopmentMultidisciplinary analysis
  in aerospace couples code and data in four
  companies
• ArchitectureA home user invokes architectural
  design functions at an application service
  provider
• Video on-demand management

91
What is Grid Good for?

• Product DevelopmentScientists working for a
  multinational soap company design and trial a new
  product
• Community ProjectsA Residents Association pools
  members PCs to analyse alternative designs for a
  local traffic calming scheme
• Medical DiagnosisUsing an adaptive display
  environment, medical staff collaborate using 3-D
  MRI scan data linked to patient records
• On-Line Games and Entertainment

92
Which Organisations/Industries?

• Aerospace
• Automotive
• Bioscience
• Defence
• Digital Media / Entertainment
• Engineering

• Financial
• ICT
• Manufacturing
• Medical / Healthcare
• Oil / Gas
• Pharmaceutical
• Universities Research Centres

93
Typical Organisations

• Commercial enterprises
• High tech with an RTD need
• Specialist suppliers contracted to a consortium
• Spin-outs / start-ups in Grid based product /
  services
• e-GovernmentAll UK local authorities to provide
  100 of services on-line by 2005

94
Typical Organisations

• Universities Research
• Further Education Institutes
• Research Centres/Parks

95
Typical Grid Opportunities

• Distributed information sharing
• Simulation/massive scale computation
• Integrated Design-Build-Test
• Global Crisis Management capability
• Application Service Provider

96
Typical Grid Opportunities

• Collaborative working
• Problem solving environments
• Improved knowledge sharing
• Resource sharing/flexibility
• Remote working

97
Take-up of Grid Technologies
98
Commercial Benefits

• Reduced capital costs
• Development cost savings
• Shorter development times
• Reduction in time-to market
• Globalisation enabler
• e-Commerce acceleration

99
Summary

• Which organisations/companies might be interested
  in Grid technology?
• In what ways would they be interested?
• What are the commercial benefits?

100
Thank you for coming
101
Thank you for coming