Transcript spacegridoverview_marchetti

SpaceGRID
EO GRID Workshop
ESRIN 6-7 May 2002
Pier Giorgio Marchetti
[email protected]
see also http://esagrid.esa.int/spacegrid
EOP-A Research and Technology Development Section
SpaceGRID Goals
• Assess how GRID technology can serve requirements
across a large variety of space disciplines and applications.
• Foster collaboration and enable shared efforts across
space applications
• Sketch the design of an ESA-wide (horizontal and
common) GRID infrastructure
• Proof of concept through prototyping
• Involve both industry and research centres
• Keep Europe up with GRID efforts !!
EOP-A Research and Technology Development Section
SpaceGRID Study
•
Financed under Agency’s General Study Programme 800KEuro
•
Datamat, main contractor, subcontractors: Alactel Space
Industries, CS Systémes d’Information , Qinetic, Rutherford
Appleton Laboratories, Science Systems (Space) Ltd
•
1 Technical Officer + 4 Domain experts from Agency:
– Eamonn Daly, Hugh Evans (Space Weather)
– Hans-Peter De Koning (Spacecraft Engineering)
– Gerhard Kreiner, Salim Ansari (Solar System Research)
– Patrick Grimont (Future EO Ground Segments)
•
September 2001-July2003
•
Co-operation with EC financed projects DataGRID, EGSO
EOP-A Research and Technology Development Section
Radiation Transport
Geant4 Application
• Monte-Carlo simulation of
high energy particle
interactions inside a
spacecraft system,
components or detector.
• Large numbers (>107) of
source particles are
required to ensure good
statistics.
• Easily adaptable to a
parallel processing system
(independent particle
histories).
EOP-A Research and Technology Development Section
Spacecraft/Plasma Simulation
•
Monte-Carlo 3-d time
dependent electrostatic
kinetic Particle-in-Cell Method.
•
Large Number of particles
(>109) to be tracked per
iteration.
•
Strong coupling between cells
required to solve the field.
•
Challenging to develop
sufficiently independent
subtasks for parallel
processing.
EOP-A Research and Technology Development Section
Space Weather Simulation
•
Simulation of a Coronal Mass
Ejection and interplanetary
shock from the Sun to the Earth
and subsequent effects on the
Earth’s magnetosphere
•
Magneto Hydrodynamic
Simulation (MHD).
•
Particle simulation of shock
effects on trapped particle belt
population (adiabatic responses).
•
Trade-off GRID vs LCMPP and
“@home” solutions
EOP-A Research and Technology Development Section
SpaceGRID - Objectivies for ESA Science
Evaluation of GRID technology in support of
Solar System Research activities:
• access to multiple distributed data
archives
• online data manipulation, analysis and
visualisation of complex data sets
• interoperability between legacy systems
and heterogeneous databases
• collaborative environment (e.g. sharing of
algorithms for analysis and visualisation)
EOP-A Research and Technology Development Section
ENVISAT
• launched 28 February 2002
•
•
•
•
•
10 instruments on board
200 Mbps data rate to ground
400 Tbytes data archived/year
~100 “standard” products
10+ dedicated facilities in Europe
• ~700 approved
science
user
projects
EOP-A Research
and Technology
Development
Section
Space Communication - ARTEMIS
Ka-Band
X-Band
Ka-Band
KIRUNA
ESOC
FOCC
ESRIN PDS
EOP-A Research and Technology Development Section
USERS
Number crunching: interferometry and
DEM generation
EOP-A Research and Technology Development Section
Knowledge Driven
Information Mining
EOP-A Research and Technology Development Section
EOP-A Research and Technology Development Section
EOP-A Research and Technology Development Section
Mechnical Engineering Usage Scenario
GRID-based analysis and test support
Test team at test facility
(ESTEC, Intespace, IABG)
Test results data


Video/audio teleconferencing
Updated analysis predicitions
Engineering analysis team
at (sub)contractor site (anywhere in Europe)



Remote monitoring of test
Near-real-time correlation analysis model with test results
Near-real-time investigation of test anomalies
EOP-A Research and Technology Development Section
Perform test scenario
Make test data available
Why Service Chaining
Service 5: Meteo info
Service 1: NDVI
Service 4: Land use
Service A3: LST
MASS project
Service 2: LST
SC permits
 exploitation of multiple
Services 1,2,3
 new improved Service:
Risk Index corrected
with meteo
 new Service: precision
farming (NDVI + land
use + meteo)
 …
Service 3=1+2 Risk Index
P1/R meeting September 11-12, 2001
EOP-A Research and Technology Development Section
Simple Service Chaining
Service C
Request
Receive
Request
Service C
Status
Process
Service
C=A+B
Service C
Delivery Info
Deliver
Service
C=A+B
Portal
Service C
Receive
Request A
Process
Service A
Deliver
Service A
Service A
EOP-A Research and Technology Development Section
Receive
Request B
Process
Service B
Deliver
Service B
Service B
High Level GRID Requirements
• Distributed access to
– computing and storage resources
– real-time data and archives
• Seamless integration with W3C Web Services
• Easy access to metadata repositories
• Seamless integration with flow control and management
• Domain based and geographic based Portals and Service
directories
• Network availability
• Group to Group Collaborative Environment
EOP-A Research and Technology Development Section