The White Rose Grid Peter Dew e-Science Centre of Excellence

advertisement
e-Science Centre of Excellence
The White Rose Grid
Peter Dew
Chair of the White Rose Grid Executive
1
e-Science Centre of Excellence
Overview
•
•
•
•
About the White Rose Grid
WRG activities
WRG plans
Concluding remarks
2
e-Science Centre of Excellence
•
•
•
•
About the WRG
The White Rose Grid (WRG) is an association of the three research
Universities - Leeds, York & Sheffield
WRG works under the auspices of the White Rose University Consortium
(featured as a model of collaboration and enterprise in the HEFCE White
Paper)
It employs complementary skill bases to support both: larger projects than
can be delivered by any one University, and a broad research agenda
It aims:
– to strengthen e-Science research (using experience gained from eScience projects such as DAME, or gViz)
• initial focus
– decision support (engineering, health, social science)
– scientific visualisation
– to access inter-enterprise computing resources via grid portals
– to assess, in collaboration with Yorkshire Forward, regional demand
for Grid technology
3
e-Science Centre of Excellence
Commitment
• Senior staff
• White Rose Grid Executive
– White Rose
• Martin Doxey: md21@cs.york.ac.uk
– University of Leeds
• Peter Dew: dew@comp.leeds.ac.uk
• Ken Brodlie kwb@comp.leeds.ac.uk
– University of York
• Jim Austin: jim.austin@cs.york.ac.uk
– University of Sheffield
• Peter Fleming: p.fleming@sheffield.ac.uk
• IT Vendors
– Esteem
– Sun & Streamline
• WRG Project teams
4
e-Science Centre of Excellence
WRG systems
• Purposely acquired - with over £3M investment - 4 HPC
nodes (in total nearly 500 CPUs)
• To be extended with the additional £1.8M investment
• A heterogeneous facility comprising 3 clusters of Sun
shared-memory systems and 2 Intel processor-based
Beowulf clusters
25%
• To offer both:
– local HPC services (75% resources)
– the Grid infrastructure (25% resources)
• Each node specialises in the provision of a
WRG
distinct service
75%
resource allocation
5
e-Science Centre of Excellence
The WRG architecture
General
Purpose
HPC node
CFD node
Engineering
Application
node
Computer
Science
node
6
e-Science Centre of Excellence
WRG project teams
•Joint support teams across the three universities
Architecture Team
Globus, MyProxy, portals
Authentication, Authorisation user management, usage account.
& Accounting Team
X.509 digital certificates
Technical Team
stable service
Training Team
HPC techniques, Grid access &
applications
Business Outreach Team
WRG
USERS
working with regional companies
& universities
7
e-Science Centre of Excellence
WRG research projects
The White Rose Grid system
is used to compute 3D
convection in the Earth’s
mantle using a numerical
model. This figure shows a
snapshot of the temperature
field from one such
calculation; depicting a cold
isosurface in blue and a hot
one in yellow.
Courtesy of Julian P
Lowman, School of Earth
Science, University of Leeds.
8
e-Science Centre of Excellence
Grid Optimisation Software for Problems
of Elastohydrodynamic Lubrication
•
•
•
•
•
Roughness
The GOSPEL e-Science project combines the
numerical solver and the optimisation work into
one Grid-enabled application run in parallel
Built inside a Problem Solving Environment
(PSE) in IRIS Explorer
Achievements include:
– largest EHL calculations undertaken with
meshes of 8000x8000 leading to 64 million
dense nonlinear equations
Research undertaken by: M Berzins, C Goodyer,
P Jimack & L Scales
Funded by Shell and the e-Science Core
Programme
Film Thickness
9
e-Science Centre of Excellence
Visualization middleware
for e-Science
The gViz project, led
by Prof K Brodlie,
aims to provide today’s
e-Scientist with
visualization software
that works within
modern Grid
environments
Figure: Grid-enabled distributed
IRIS Explorer
10
e-Science Centre of Excellence
DAME
Grid Services
11
e-Science Centre of Excellence
12
e-Science Centre of Excellence
Struts-based
Portal using
OGSI
Browser
View X
Configuration
Controller
View Y
•
View Z
•
Action X
Action Y
Action Z
•
Workflow
Manager
Grid
Service
A
Grid
Service
B
Grid
Service
C
Grid
Service
D
•
Centralise request handling and
decision points, e.g. security
Avoid “cut & paste” code
distribution
Separate business logic from
views
Easier to maintain, update and
extend functionality
13
e-Science Centre of Excellence
• Dependable, secure quality service secure
quality service
14
e-Science Centre of Excellence
Business benefits
• The WRG underpins a variety of e-Science projects e.g.
DAME, HYDRA, gViz
• The WRG has engaged on a two-year outreach project
(funded by YF) to assess the value of a regional Grid
infrastructure. These activities include:
– assessment of regional interest
– development of a business plan identifying the type of
services, and the role – supported by both accountants
Deloitte & Touche and solicitors Hammonds
– provision of a trial infrastructure for company incubators
at the Innovation Centres/Science Parks of the three
Universities
15
e-Science Centre of Excellence
WRG Evolution
e-Science Grid &
National Grid Service
WRG
C
EDG/EGEE
WUN Grid
C
C
Academic Service
Infrastructure
WRG
Outreach
Companies
Buy
On-demand
Services”
16
e-Science Centre of Excellence
Concluding remarks
• The WRG focus is on :
– distributed diagnostics
– collaborative visualisation
– High Energy Physics
• Underpinned by
– Research: dependable, secure quality service
• Future tasks:
– increase the cross-site usage
– to develop further Grid portals
– to continue to work with regional companies
17
e-Science Centre of Excellence
References
• White Rose University Consortium http://www.whiterose.ac.uk
• WRG web site - http://www.wrgrid.org.uk/
• DAME XTO portal –
http://iri02.leeds.ac.uk:8080/damexto/damexto
• P M Dew, J G Schmidt, M Thompson, P Morris
The White Rose Grid: practice and experience –
in the proceedings of the All Hands conference
18
e-Science Centre of Excellence
Thank you for your attention
19
Download