A clustering analysis and agent-based trust model in a grid environment supporting virtual organisations

ISSN： 1741-1106

Source： International Journal of Web and Grid Services, Vol.5, Iss.1, 2009-03, pp. : 3-16

Disclaimer: Any content in publications that violate the sovereignty, the constitution or regulations of the PRC is not accepted or approved by CNPIEC.

Previous Menu Next

Abstract

Grid computing enables coordinated resource sharing in dynamic, distinct Virtual Organisations (VOs), which consist of different grid entities from traditional organisations. Recently, the main solution of grid security is applying trust as an important factor in grid computing to realise cooperation between unknown entities, especially for VOs. However, most of the research work inherits the concept of trust from Peer-to-Peer (P2P) networks and does not accurately satisfy the requirements in a grid environment. In this paper, we present a clustering analysis and agent-based trust model that introduces VO trust relationships into traditional grid entities, with which we can more effectively and dynamically share grid resources. We also propose a clustering analysis method to evaluate grid entities' trust in VOs and a weighted path to compute the transitivity of trust. In addition, two kinds of agents are introduced to the trust evaluation process to collaborate on and accomplish trust computation between VOs. The VO agent is a resident agent that stays in a VO which initiates search agent to collect information. A search agent is a mobile agent and is usually dispatched to many VO nodes. The simulation result shows that our model improves the overall performance of grid systems.