P2P网络负载调度的超图划分算法

发布时间：2018-11-01 20:34

【摘要】：目的解决P2P用户数量持续增多时,网络节点异质性造成的负载不均衡问题,提高节点和带宽利用率.方法基于超图的动态划分思想和排队论的理论证明,将网络划分为若干个小区域的集合,利用定义的超级节点对本区域内的节点进行管理.每个区域随机选取另一个区域共同进行超图的2路划分过程,使划分后的区域总负载保持一致.通过每个小区域的节点负载平衡来获得整个网络负载的相对平衡.结果实验结果表明,采用负载均衡算法的网络节点利用率有了明显改善,而且能在更短的时间内调整至均匀分布状态.在连接数较高的情况下,网络仍然能够保持较好的稳定性.结论基于超图划分的负载调度方案可以使系统的局部范围处于平衡状态,进而实现P2P网络整体的负载均衡.
[Abstract]:Aim to solve the problem of load imbalance caused by heterogeneity of network nodes and improve the utilization of nodes and bandwidth when the number of P2P users continues to increase. Methods based on the theory of hypergraph dynamic partitioning and queuing theory, the network is divided into several sets of small regions, and the defined super nodes are used to manage the nodes in the region. Each region randomly selects another region to carry out the 2-way partition process of hypergraph, so that the total load of the divided area is consistent. The relative balance of the network load is obtained by balancing the nodes in each small area. Results the experimental results show that the network node utilization rate using load balancing algorithm has been significantly improved and can be adjusted to a uniform distribution state in a shorter time. In the case of high number of connections, the network can still maintain better stability. Conclusion the load scheduling scheme based on hypergraph partition can balance the local scope of the system and realize the load balance of P2P network as a whole.
【作者单位】：沈阳建筑大学网络中心;沈阳建筑大学信息与控制工程学院;
【基金】：国家自然科学基金项目(61272253)
【分类号】：TP393.02

【相似文献】