摘要
分布式哈希表(DHT)存在逻辑拓扑与实际网络拓扑不匹配的问题,造成查询操作时延过大。针对该问题,提出一种基于节点交换的DHT优化方法,通过对DHT节点逻辑位置的调整,改善DHT的查找性能。使用地标聚类方法对节点进行分类,物理位置相近的节点属于同一个地标聚类区,采用地理布局的思想,使DHT覆盖网络逻辑拓扑与底层物理拓扑尽量匹配,从而达到缩短DHT查询操作的时延、提高DHT路由性能的目的。
Due to the problem of Distributed Hash Table（DHT）, which is that the logical topology does not match the actual network topology, it results in the problem of excessive delay of lookup operation. Aiming at this problem, this paper proposes a DHT optimization method based on node exchange to solve this problem by adjusting the logical location of DHT node, and ultimately to improve the search performance of DHT. The nodes are classified by using the landmark clustering method, the nodes whose physical location close to each other belongs to the same landmark district. It uses the idea of geographic layout, makes the DHT overlay network logical topology match the underlying physical topology as much as possible, and thus reduces the DHT query delay, which improves the performance of DHT routing.
出处
《计算机工程》
CAS
CSCD
北大核心
2011年第8期70-73,共4页
Computer Engineering
关键词
分布式哈希表
节点标号
地标聚类
地理布局
Distributed Hash Tabte（DHT）
node label
landmark clustering
geographic layout
