A Collaborative e-Science Architecture towards a Virtual Research Environment

1
A Collaborative e-Science Architecture towards a
Virtual Research Environment

• Tran Vu Pham1, Dr. Lydia MS Lau1, Prof. Peter M
  Dew2 Prof. Michael J Pilling3
• 1 School of Computing, University of Leeds,
  Leeds, UK
• 2 Informatics Institute, School of Computing,
  University of Leeds, Leeds UK
• 3 School of Chemistry, University of Leeds,
  Leeds, UK
• Fourth e-Science All Hands Meeting
• 19 - 22nd September 2005 Nottingham

2
Content

• What do the e-scientists do?
• A Case study The Reaction Kinetics research
  community
• Examples of current architectures for VREs
• Web-based, Grid-based
• Inclusion of P2P - The Collaborative e-Science
  Architecture
• Testing the architecture
• Conclusion and future work

3
Case Study The Reaction Kinetics Research
Community
Modelling

• REACTIONS MOLES KJOULES/MOLE
  
• H2O gt OHH
  5.120E04 2.67 26.27
• OHH gt H2O
  3.534E04 2.62 18.95
• H2OH gt H2OH
  1.020E08 1.60 13.80
• H2OH gt H2OH
  4.520E08 1.60 77.08
• O2HM gt HO2M
  2.100E18 -.80 .00
• N2/0.67/ O2/0.4/ H2O/0./ AR/0.28/
• HO2M gt O2HM
  1.159E20 -1.26 211.41
• N2/0.67/ O2/0.4/ H2O/0./ AR/0.28/
• O2HH2O gt HO2H2O
  6.890E15 .00 -8.73
• HO2H2O gt O2HH2O
  3.801E17 -.46 202.68
• O2H gt OHO
  9.756E13 .00 62.11
• OHO gt O2H
  1.450E13 .00 2.94
• H2O2H gt HO2H2
  1.690E12 .00 15.71
• HO2H2 gt H2O2H
  1.507E09 .78 83.91
• H2O2H gt OHH2O
  1.020E13 .00 14.97
• OHH2O gt H2O2H
  6.724E07 1.28 295.88
• H2O2O gt OHHO2
  6.620E11 .00 16.63
• OHHO2 gt H2O2O
  4.073E08 .72 77.51

Chemical Reaction Mechanisms
Laboratory Experiments
Application in Environmental Chemistry
Simulations
Applications in Atmospheric Chemistry
Sensitivity Analysis
Applications in Combustion Chemistry
4
Case Study Challenges

• The requirements of Reaction Kinetics Research
  Community are related to two challenging issues
  in e-Science community
• How to provide the scientists with an integrated
  collaborative research environment for user
  collaborations
• How to provide the scientists easy access to
  computationally intensive resources from a
  desktop computer

5
Examples of Current Architectures Web-Based

1. The Virtual Knowledge Park, http//vkp.leeds.ac.uk
   
2. British Atmospheric Data Centre,
   http//badc.nerc.ac.uk

6
Examples of Current Architectures Grid-based

1. The White Rose Grid, http//www.wrgrid.org.uk
2. Collaboratory for Multi-Scale Chemical Science,
   http//cmcs.org

7
Examples of Current Architectures Summary

• Grid architecture is good for dealing with the
  need for computational resources and data storage
• Collaborations amongst users use web portal
  (basically a centralised web-based architecture).
  This approach has a few limitations
• Direct collaborations within such a community are
  limited (it is possible to use email, but this
  method is not suitable for sharing large data
  files)
• Cross-community collaborations are limited
• It is hard to form ad hoc working groups, which
  consist of members from different communities

8
Potential of P2P Computing
A P2P network

• Direct communication of peer users
• Bring end users closer to their communities and
  shared resources
• Sense of privacy and ownership over shared
  resources
• Ad hoc group can be formed easily to support
  collaborative work

A peer can easily connect to other peers in the
network
9
The Collaborative e-Science Architecture (CeSA)

• An collaborative architecture for an integrated
  collaborative research environment
• to better support user collaborations in
  distributed communities
• to provide scientists with easy access to
  computation intensive resources and storage

10
The CeSA An Example

1. The White Rose Grid, http//www.wrgrid.org.uk
2. Collaboratory for Multi-Scale Chemical Science,
   http//cmcs.org

11
The Collaborative e-Science Architecture (CeSA)
Collaborative Research Environment
Service Oriented Architecture
Resource Providers
12
The CeSA P2P Environment

• An integrated environment for user lightweight
  collaborations chatting, file sharing, discovery
  of shared resources, etc
• Tools for users to form virtual working groups
• User interfaces for executing services from grids
• Publication and discovery of services from the
  grids

13
The CeSA Components of P2P Application
14
The CeSA Grids

• Grids are providers of computation and data
  intensive resources (e.g. large datasets)
• These resources and data are provided to P2P
  environment in form of High Level Services
• Management of user community in P2P environment
  is separated from the management within grids

15
The CeSA High Level Services

• Can be built by wrapping resources on the grid
• Can be composite services or workflows available
  on the grids
• Conform to a unified service interface

16
Testing the architecture

• Approach
• Prototyping building a prototype system based on
  the CeSA using real requirements data from
  reaction kinetics research community, using JXTA
  and GT3
• Experiment conducting user evaluation on the
  prototype system and collecting their feedback

17
Prototype An Example Screen Shot
A working group
Services for Reaction Kinetics
A Peer member
Browsing service registry from a peer-to-peer
application
18
User Evaluation Objectives

• To evaluate the effect of using P2P
  collaborative environment provided by the CeSA in
  a realistic user working environment
• To assess how users can benefit from the access
  to remote simulations and analyses provided by
  grids via Grid Services
• To capture user general attitudes to the new
  collaborative research environment

19
User Evaluation Method Process

• Data collection method
• Questionnaire
• Participants were provided with mixture of closed
  and open questions
• A collaborative scenario was also provided
• The experiment process
• Three scientists involved in the experiment at
  the same time
• They used the prototype as guided by the scenario
  to collaborate with each other
• Their feedback was recorded in the questionnaire

20
User Evaluation The Results

• Generally positive
• Participants expressed their interests on using
  the prototype system, here is some feedback
• 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
• However, there were also some worries about
  security

21
Conclusion Future Work

• The positive result has shown the potential of
  the CeSA for collaborations in a scientific
  community
• The CeSA is a suitable architecture for Virtual
  Research Environments but need
• to investigate scalability issue of P2P
• to address security requirements
• more testings/ evaluations.

22
Questions?

• Email tranp_at_comp.leeds.ac.uk