【摘要】 使用云内容分发网络(Cloud Content Delivery Network,CCDN)对数据内容进行分发与传统的内容分发网络(Content Delivery Netw ork,CDN)相比可以节省大量运营成本,是当前进行内容分发的主要措施.内容副本放置优化问题是CCDN研究的关键问题之一.然而如何在考虑CCDN系统服务质量、降低能耗开销的情况下选择最优路径放置副本仍然是一个极有挑战性的问题.为解决这一问题,本文提出了一种基于多播树的能耗优化分发模型(Energy Efficient Delivery Model,EEDM).该模型首先使用K-Canopy算法确定关键节点个数k;其次通过K-Means聚类算法,将整个网络划分成k个区域;然后通过关键节点选择算法,找出这k个区域的关键节点作为多播组的成员;最后根据最小化分发代价的副本放置多播路由选择算法,构建一棵连接所有多播成员的分发树.实验表明,该模型减少了副本分发的路径长度,降低了整体能耗.更多还原

【Abstract】 The delivery of data content using cloud content delivery network( CCDN) can save a lot of operating costs compared to traditional content distribution networks( CDNs). It’s the main measure of the current content delivery. Content placement optimization is one of the key issues in CCDN research. However,howto choose the optimal path to deploy the replacement in consideration of QoS and reducing energy consumption in CCDN system is still a very challenging problem. To solve this problem,this paper presents an energy efficient delivery model( EEDM) based on the multicast tree. The model first uses the K-Canopy algorithm to determine the number of key nodes k; Then,we divide the entire network into k regions through the K-M eans clustering algorithm. Next,we identify k key nodes as members of the multicast group for each region through a key node selection algorithm; Finally,we build a delivery tree which connects all multicast members according to the replica replacement cost minimized multicast routing algorithm. Experiments showthat the proposed model reduces the path length of the replica distribution and thus reduces the overall energy consumption.更多还原