聚类算法在无标度网络和云资源管理中的应用

Applications to clustering algorithms for scale-free networks and cloud resource management

基于最优云资源管理策略,提出了一种构建无标度网络模型的方法。网络建模为一个完全联通图的聚类应用,聚类算法实现了近似理论的度分布。能够重现现实网络权重和节点度呈幂律分布的统计特性,支持自组织,允许单个实体基于本地信息做出决策。实验结果表明,具有自我组织和自我管理属性的无标度网络,其各个核心节点可以作为计算机云的接入点,一旦形成这种集群可以实现最佳的云资源管理政策。

Based on the optimal cloud resource management strategies, a building scale-free network model is proposed. Unicom network is modeled as complete graph clustering applications, clustering algorithm to achieve a degree of approximation theory of distribution. The model is able to reproduce real network weights and node degree is the statistical characteristics of the powerlaw distribution, it also provides support for self-organization, allowing a single entity to make decisions based on local informa tion. Experimental results show that, various core nodes of the self-organization and self-management attributes scale-free net work can be used as cloud computer access points, once the formation of such clusters achieves, the optimal cloud resource management policy is realized.