摘要
在现实的P2P网络环境中,由于节点的计算能力和带宽等方面的异构性,网络负载不均衡现象非常突出。基于数据复制/转移策略,提出一种动态的平衡算法。根据节点的能力、当前节点负载状态、负载转移代价预估算,在整个系统范围内找到一组传输代价较小并且负载较轻的节点集合,从中随机选取较为适宜的节点进行负载转移或者数据复制。试验结果表明,该算法能够有效地均衡负载的分布以及降低负载的迁移率。
In a real P2P network environment, the heterogeneous of computational capacity and bandwidth among nodes could result in load imbalance. A dynamic load balancing solution with efficient online data migration/replication strategy is proposed. This strategy, keeping track ofcurrent workload, capacity statistics and load-transfer cost, pick randomly a proper peer to do data migration /replication. According to the experimental results, the solution can efficiently balance load distribution and decrease the ratio of load-transfer.
出处
《计算机工程与设计》
CSCD
北大核心
2009年第1期58-60,共3页
Computer Engineering and Design
