Supporting Formation and Operation of Virtual Organisations in a Grid Environment

1
Supporting Formation and Operation of Virtual
Organisations in a Grid Environment
Jianhua Shao
(Demo - Patrick Stockreisser Gareth Shercliff)
2
Constraint Oriented Negotiation in Open
Information Seeking Environments for the Grid
Alun PreeceTim Norman Peter Gray Stuart
Chalmers Nir Oren
Alex Gray Nick Fiddian Jianhua Shao Vikas
DeoraGareth ShercliffPatrick Stockreisser
Nick Jennings Mike Luck Viet Dang Thuc
Nguyen Luke Teacy Jigar Patel
David GriffithsDave RohlfingSimon Thompson
3
Summary

• The CONOISEG Project
• Virtual Organisations
• Concept and Lifecycle
• Challenging Issues
• Agent-based Solutions
• Demonstration
• Conclusions

4
The CONOISE-G Project
Grid Applications
Grid Infra-Structure
5
The CONOISE-G Project
Agents
CONOISE-G
Grid Infra-Structure
6
Virtual Organisations (VO)

• Consist of autonomous entities
• Each entity has a range of resources and some
  problem solving capabilities

7
Challenges
VO Formation

• How to form a best possible VO?
• what is available?
• can a bid be trusted?
• what should be selected?
• is the quality good?
• etc.

• How to manage a VO effectively?
• honour your contract?
• new providers available?
• what is the quality delivered?
• etc.

8
Types of VO

• Establishing collaborations
• Forming supply chains
• Exploiting market gaps
• Reacting to market demands
• etc.

9
Types of VO

• Establishing collaborations
• Forming supply chains
• Exploiting market gaps
• Reacting to market demands
• etc.

10
Service Provider Behaviour

• An open market place where service providers can
  come and go.
• Service providers can compete against one another
  for orders.
• Service providers cannot entirely be trusted to
  honour their promises.

11
VO Lifecycle and Workflow
Discover services
Obtain bids
Select bids
Form the VO
Monitor provision
12
Discovering Services

• For a given service request, discover who can be
  a potential provider.

Potential Providers
RA
YP

• DAML-S was extended to include QoS attributes in
  service description 1.

13
Obtaining Bids

• RA must decide whether to invite an SP to bid.
• An SP must decide whether/what to offer.
• SP uses constraint reification in decision making
  2.

14
Selecting Bids

• Assessing the bids

Establishing utility for providers

• Conducting an auction

15
Assessing the Bids

• Expectation based QoS assessment 3

Conventional QoS is calculated from what is
promised and what is delivered.
Expectation Based QoS is calculated from what
is promised, what is delivered, and what the user
expects.

• A preliminary trust model 4

16
Conducting Auction

• Combinatorial reverse auction -- bidders may bid
  for arbitrary combinations of items.
• Bidders specify piece-wise linear functions of
  quantity against utility for each item.
• Polynomial clearing algorithm guarantees a
  solution within a finite bound of the optimal 5.

A supply/demand curve for two items
17
Forming VO

• Establishing a service level agreement with all
  parties.
• Setting up arrangements for service provision to
  be monitored.
• Reacting to changes within VO or in market.

VOM
etc
market
QoSC
18
Monitoring Service Provision

• QoS Monitoring and prediction

• Reactive policing

VOM
SPi
Policing
QoSC
Contract
19
CONOISE-G And Grid
VO formation management capability on Grid
Wrapping / enabling any existing capability /
resources
20
Conclusions

• The ability to form and operate virtual
  organisations in grid is important.
• We aim to support robust and resilient VO
  formation and operation.
• We are developing core technologies for
• Service discovery incorporating QoS
• Decision making mechanisms during VO formation
• Establishing trust reputation
• Policing within VO
• QoS provision monitoring and prediction

21
References

• V. Deora, J. Shao, P. J. Stockreisser, G.
  Shercliff, W. A. Gray and N. J. Fiddian.
  Incorporating QoS specifications in service
  discovery. to appear in Web Services Quality
  Workshop 2004.
• S. Chalmers, A. Preece, T. J. Norman, and P.M.D.
  Gray. Commitment management through constraint
  reification. Proc. 3rd Int. Joint Conf. on
  Autonomous Agents and Multiagent Systems, volume
  1, 2004, 430-437.
• V. D. Dang and N. R. Jennings. Polynomial
  algorithms for clearing multi-unit single item
  and multi-unit combinatorial reverse auctions.
  Proc. 15th European Conf. on AI (ECAI), 2002,
  23-27.
• V. Deora, J. Shao, W. A. Gray and N. J. Fiddian.
  A quality of service management framework based
  on user expectations. Proc. 1st Int. Conf. on
  Service Oriented Computing (ICSOC), 2003,
  104-114.
• L. Teacy, J. Patel, M. Luck and N. Jennings.
  Trust provision in CONOISE-G. CONOISE-G Technical
  Report, 2004.