Enabling e-Research in Combustion Research Community

1
Enabling e-Research in Combustion Research
Community

• T.V Pham1, P.M. Dew1, L.M.S. Lau1 and M.J.
  Pilling2
• 1School of Computing
• 2School of Chemistry
• University of Leeds

2
Outline

• Combustion research community
• Research focus and process
• Current limitations and issues
• Requirements
• The Collaborative e-Science Architecture
• Early user evaluation results
• Application of the Collaborative e-Science
  Architecture to combustion research community
• Future work
• Conclusions

3
Combustion Research Community

• The central focus is on modelling of chemical
  reaction mechanisms
• Consist of members from around the world
• Related to reaction kinetics and atmospheric
  research communities

4
3-Stage Modelling Process in Combustion Research
5
Limitations and Issues

• Data necessary for generation of new models are
  scattered in the community
• Lack of coordination across research groups
  making the gathering and evaluation of data more
  difficult
• Use of many different custom built data formats
• Need support for computational capability

6
Requirements from Combustion Research Community

• A collaborative infrastructure to support their
  distributed collaborations
• Allow scientists who are working on the same or
  similar research activities to dynamically form
  working groups
• Provide efficient support for timely
  collaborations within and across working groups
• Sharing expert knowledge, day-to-day working data
  to speed up the data collection and evaluation
  process.
• Provide easy access to computational intensive
  resources
• Time and resource consuming simulations and
  analyses
• Storage of large amount of experimental data.

7
Grids Web-based Portal Approach
8
The Collaborative e-Science Architecture (CeSA) -
Goals

• To be scalable with respect to decentralised
  nature of scientific communities.
• Able to support scientific collaborations at
  different levels of granularity
• Able to provide access and enables back-end
  computationally intensive resources for complex
  computation and storage requirements.

9
The Collaborative e-Science Architecture (CeSA)
10
Potentials of P2P Computing
11
CeSA Service Oriented Architecture
12
CeSA System Architecture
User
P2P
Application Specific Services
WG Service (e.g. security)
Modelling Services
Data Services
Grid
OGSA Services
Computation and Date Resources
Model Repository
Chemical Reaction Data Library
13
Ontology-based Adaptive Approach to Resource
Discovery

• For resource discovery in the P2P collaborative
  environment of the CeSA
• Provide an efficient mechanism for query routing
  by exploiting user interests
• Try to forwards search queries to peers that most
  likely to have the answers
• Use ontology for classification of user interests
• Learn from past query results to know other
  peers interests in order to adaptively route
  query
• Simulation results showed significant improvement
  over the basic flooding approach

14
Early User Evaluation

• A prototype instance of the CeSA was built using
  JXTA P2P platform and Globus Toolkit version 3
• A number of simulation programs in for chemical
  reaction mechanism were wrapped into Grid
  Services
• The prototype was evaluated by potential users
  from reaction kinetics research group at The
  University of Leeds
• Initial results were positive
• A fully working system would benefit the
  atmospheric chemistry group provided it was
  widely accepted by the whole community
• I think that our group would certainly use such
  a system if it proved to be the way forward in
  e-Science (which I feel it is) and the community
  embraced the use of such a system

15
Application of the CeSA to Combustion Research
Community
16
Addressing the Limitations and Issues

• Making scattered data easily accessible through
  P2P resource discovery
• Identifying expertise for potential
  collaborations through P2P collaborative
  environment
• Supporting the modelling process with
  computational and data resources from the Grid
  environments using Grid/Web Services.

17
Conclusions and Future Work

• Early user evaluation has confirmed the potential
  of the CeSA, particular on the use of P2P
  collaborative environment to support distributed
  scientific collaborations
• CeSA can also potentially be used for the
  combustion research community, which is closely
  related to the reaction kinetics community
• Further user evaluation of the CeSA on the
  combustion research community is being planned
• Further work is also necessary on the management
  of ontology in the P2P environment
• Research on technical qualities include security,
  connectivity and scalability of resource
  discovery of P2P application.