| Popis: |
In a grid environment, resources and services are distributed with dynamic and heterogeneous characteristics. Efficient service discovery is one challenging issue in a grid environment. In this paper, we propose a new distributed and hierarchical mechanism for improving fault tolerance and speed of service discovery in a grid environment. This approach has five layers that in previous works was presented in four layers. A root layer is added as coverage for the service index of under layers. For improvement fault tolerance in the proposed approach, after the requested service is found, the action of service discovery will not stop. It continues to find several instances of the requested service. These services will send to the cache of the institution layer that will be replaced immediately if necessary. Also for speed-up service discovery, nodes in same level, which have the same parent, send queries to each together directly. Performance evaluation shows that our approach achieves good efficiency, is fault tolerant and consistent. The proposed mechanism for service discovery improves the fault tolerance of services more than 28%. |
Full Text from ScienceDirect