Abstract
Autonomic and distributed management requires the cooperation between the network nodes and the periodical exchange of network information between them. To increase scalability of the information exchange, network nodes need to be associated and grouped through an efficient set of rules, to minimize the amount of control messages flowing in the network. This paper proposes a community-based association between nodes, where nodes in a community are connected through social-based metrics. The proposed social-based metrics are able to join the network nodes through their characteristics and probable interactions, reducing the disconnection between nodes and optimizing the information exchange between them. The obtained results show that the social-based metrics increase stability of the links and provide better resource aggregation, decreasing the required number of messages exchanged between nodes in a distributed management approach
Authors: Lucas Guardalben, Tome Gomes, Paulo Salvador and Susana Sargento