A decentralized quickest response algorithm for grid service discovery
Proceedings of the 2nd international conference on Scalable information systems
Hi-index | 0.01 |
Distributed Hash table (DHT) is widely used in Peerto- peer networks for its scalability, adaptability, and faulttolerance in large and dynamic environments. In this paper, we study on the application of DHT for Grid service discovery. Grid service is represented by multiple (attribute, value) pairs. The virtual organization (VO) is also selforganized into a 2-dimensional network overlays - one for attributes, and the other for values - as nodes join and publish services in the VO. Service discovery in a VO is firstly directed to the attribute overlay to find the nodes with the discovered attributes, then to the value overlay until the nodes which have the same (attribute, value) are reached. Gateway servers are deployed on the border to interconnect different VO's. The service discovery request can be redirected to other VO's in case that there is no provider in the requester's own VO. Our approach is highly efficient, scalable and quick responsive.