A deployment strategy of service registry in complex network

Many characteristics of complex networks, such as small-world, scale-free, aggregation properties, and so on have been thoroughly studied. Super-node structure is an important way to resolve the problem of service node selection in complex network. Two key issues of this method are how to define the scope of super-nodes in their respective communities and how to choose an appropriate super-node from so many nodes. This paper proposes a UDDI (Universal Description, Discovery, and Integration) node deployment model CSPS-UDDI (Community-based Super Peer Selection Model) in complex network. Using the communities’ structural characteristics, this model divides the whole network community according to the shortest time-delay of nodes between the community and the idea of load balancing in community structure proposed in this paper. During the process of division it real-time selects the super node in each community and applies to specific applications of UDDI service registry. Experimental results show that when compared with other algorithms of community structure division, the algorithm of CSPS-UDDI model is significantly more efficient in UDDI applications, especially for its strong adaptability to the network which has a high requirement in real-time of UDDI node selection. Keywords-component: complex network; community structure; UDDI deployment; load balancing