The Reading e-Science Centre

1
The Reading e-Science Centre
http//www.resc.rdg.ac.uk resc_at_rdg.ac.uk
(plus many more colleagues http//www.resc.rdg.ac
.uk/personnel.php)
About the ReSC The Reading e-Science Centre
(ReSC) was formed in November 2003 as a Centre of
Excellence for promoting and developing e-Science
methods within the environmental sciences. The
ReSC is hosted in the NERC Environmental Systems
Science Centre (ESSC) at the University of
Reading and is a collaboration between ESSC and
the School of Systems Engineering. We work with
many partners in academia, government agencies
and industry and this poster introduces some of
our current projects. We are also very active
within the University of Reading and are working
to improve the e-Science facilities available to
Reading researchers in all disciplines. Central
to this effort is the construction of a Campus
Grid facility.
Search and Rescue British Maritime Technology
Ltd. (BMT) have developed an application (SARIS,
lower picture) that is used by the Coastguard for
predicting the location of a person that has
fallen overboard from a vessel at sea. The
software has been very successful in reducing the
time to rescue, and the ReSC has worked with BMT
and the Met Office to further improve the
accuracy of SARISs predictions by allowing it to
ingest up-to-date Met Office forecasts of ocean
currents and surface winds via the GADS Web
Service (see Data Serving). A new prototype
version of SARIS has been produced. The prototype
system will be further developed in the DEWS
project (see New projects). There are many
more uses for real-time met-ocean data such as
oil spill mitigation and predicting occurrences
of respiratory problems.
Data serving Modern computer simulations and
satellite observations of the oceans and
atmosphere produce very large amounts of data.
The ReSC provides access to around 2TB of such
datasets through a Web Service, known as GADS
(Grid Access Data Service) 1, which was
produced as part of the GODIVA project. A Web
interface to GADS is available see picture
opposite. We are investigating 2 how these
datasets can be managed more effectively using
modern database management systems (DBMSs).
Barrodale Computing Services (BCS) have recently
developed a software module (the BCS Grid
DataBlade), that plugs into IBMs Informix
Dynamic Server (IDS) and PostgreSQL, for storage
and efficient retrieval of gridded data. The
DataBlade is capable of subsetting,
interpolating, rotating, aggregating and
re-projecting data and thus allows users to
download data in the format required, without any
knowledge of how the data are stored. The
picture opposite illustrates how data can be
extracted along arbitrary, user-specified paths.

• Campus Grid
• Led by the School of Systems Engineering and with
  the support of IT Services, the ReSC is
  developing a Campus Grid infrastructure at
  Reading. We currently have a small number of
  users but are encouraging other groups to get
  involved. Key features of the Grid include
• Based on Condor/SRB with Globus interface.
  Currently 150 Linux nodes, but expanding.
• Future plans to extend to cluster facilities.
• Use of CoLinux (www.colinux.org) to run emulated
  version of Linux under Windows allowing use of
  large number of machines in teaching laboratories
  and computer rooms.
• Evaluation of Inferno Grid software and
  comparison with Condor.
• The Campus Grid is being designed to be
  compatible with the National Grid Service and
  Reading aims to be an NGS partner in future.

GODIVA web interface to data holdings at ReSC
(http//www.nerc-essc.ac.uk/godiva) and sample of
extracted data (inset sea surface temperature
from Met Office FOAM model)
Grid Flight Demo (www.barrodale.com). This
application demonstrates the extraction of
temperature data along an irregular flight path.

• Work within The University of Reading
• As well as working within the wider environmental
  science community, the ReSC supports and promotes
  e-Science activities within the University of
  Reading. These activities have included
• Promoting use of the Campus Grid and the
  National Grid Service. We have identified
  several local groups across many disciplines that
  we can help to use these facilities.
• Working with the Arts and Humanities Research
  Council to advise on e-Science methods. The ReSC
  has a consultancy role on the OGHAM project,
  which will build a Virtual Research Environment
  at the Silchester late Roman archaeological dig
  site near Reading. (http//www.silchester.rdg.ac.u
  k)
• Providing the Access Grid (below) for use by
  various groups.
• Working with Deafax, a Reading-based charity
  for the deaf. Deafax believe that the Access
  Grid, with its high-quality video capabilities,
  could be a valuable tool for communication and
  distance learning for deaf people (the video
  quality is good enough to support the use of sign
  language and lip reading).

Research The ReSC is involved in research
activities in many areas of environmental
science, currently including meteorology,
oceanography and volcanology. Here are some
examples.
Numerical simulations of flow through porous
media (E. Llewellin and J. Blower) The flow of
gases through volcanic rock is a major influence
on the style of a volcanic eruption. If the rock
is highly permeable, gas can escape and the
internal pressure of the volcano can dissipate.
If gas becomes trapped, the chances of an
explosive eruption occurring are increased.
Finite element (FEM) and Lattice Boltzmann
simulations and will help to understand this
process.
Water budget analysis (C. Old, C. Liu and K.
Haines) Budgetary analysis of North Atlantic and
Arctic waters is being carried out on 100 years
of output from the HADCM3 model. The persistence
of anomalies of mode water and mass
transformation are of particular interest. The
calculations are distributed across a network
using Web Services from the GODIVA project,
reducing typical calculation times from 3-4 hours
to 30 minutes.
FEM model of inviscid gas flow through a
simplified volcanic rock
Distributed system diagnostics of weather and
climate (L. Froude and K. Hodges) A Web interface
has been created to provide access to the TRACK
program for identifying and tracking storm
systems in meteorological data. The interface
allows users to run TRACK on large, distributed
datasets (using the OPeNDAP protocol) across a
heterogeneous Condor pool of desktop workstations.
Styx Grid Services (J. Blower, K. Haines and E.
Llewellin) The ReSC has created the Styx Grid
Service (SGS), a remote service type that allows
data to be streamed directly from service to
service 3. It also allows the asynchronous
monitoring of progress and status data without
requiring clients to open incoming firewall
holes. The framework can be used in
collaborative visualization and computational
steering. See http//jstyx.sf.net.
By wrapping SGSs as Web Services, the Triana
environment can be used to assemble workflows in
which data are streamed from service to service
Plot of storm tracks in the Northern Hemisphere
in Winter 1948 from NCEP re-analysis data
New projects and future work We have recently
started, or are about to start, several new
projects DEWS (Delivering Environmental Web
Services) This is a 2.2M DTI Inter-Enterprise
Computing project, led by the Met Office and
involving the ReSC, BMT, IBM, BADC and Lost Wax.
DEWS will produce a functional demonstrator for
the secure online delivery of Met Office data via
Web Services, initially to clients in the
maritime and health sectors. GCEP (Grid for
Coupled Ensemble Prediction) This is a NERC
e-Science project, led by the ReSC and involving
the British Antarctic Survey (BAS) and CCLRC.
GCEP will create and use a Grid of PC clusters to
run ensemble climate predictions over decadal
timescales, investigating the sensitivity of the
predictions to different initial conditions.
GCOM (Global Coastal-Ocean Modelling) This is
a NERC e-Science project, led by Proudman
Oceanographic Laboratory, that aims to couple
high-resolution shelf seas models with global
models of ocean circulation in a Grid
environment, in order to improve climate
modelling. Grid toolkit for environmental
e-Science The ReSC will develop tools to enable
environmental scientists to use Grid technology
with little effort. Initially, this will be
implemented as extensions to the popular CDAT
(Climate Data Analysis Tools) software.
References 1 Woolf, A., Haines, K and Liu, C, A
Web Service Model for Climate Data Access on the
Grid, International Journal of High Performance
Computing Applications, 17(3) 281-295 (2003) 2
A. Santokhee, C.L. Liu, J.D. Blower, K. Haines,
I. Barrodale, E.Davies, Managing and serving
large volumes of gridded spatial environmental
data, Proceedings of the 11th EC GI GIS
Workshop, ESDI Setting the framework 29 June-1st
July 2005 3 Jon Blower, Keith Haines, Ed
Llewellin, Data streaming, workflow and
firewall-friendly Grid Services with Styx,
Proceedings of the UK e-Science All Hands
Meeting, September 2005