Heterogeneity-Aware Group-Based Semantic Overlay Network for P2P Systems

In the classical semantic overlay network, all the nodes are regarded as the same, without considering the differences of nodes' sharing resources and the network distance between them. Thus the resource classification maintenance and search costs in such network are very high. In this paper, by analyzing the impact of nodes' shared resources and the network distance between them on the resource classification maintenance and search costs in semantic overlay network, we propose a novel semantic overlay network based on unstructured P2P network. In our proposed network, by P2P network topology matching method and limited use of resource classification, the resource classification maintenance and search costs in the classical semantic overlay network can be greatly reduced, and thus the network performance can be improved.