基于智能配网业务的WRED算法改进及仿真

摘要

In consideration of the complexity and diversity of the smart distribution business, a Quality of Service(QoS) mechanism that is more flexible and efficient than that of tradition is needed. The smart distribution service characteristic is analyzed. WRED_gentle algorithm, an improved Weighted Randomly Early Detection(WRED) algorithm, is proposed in active queue management. By the modeling of meter reading application and the power distribution automation application, noWRED, WRED, and WRED_gentle algorithm are simulated on OPNET. Results indicate that the performance of network delay, throughput and link rate is obviously improved after adopting the WRED_gentle, which guarantees the QoS requirements of the smart distribution communication network.%由于智能配网通信业务的复杂多样性,要求其通信网提供比传统服务质量(QoS)机制更具有针对性的灵活高效的QoS机制保证.为此,分析智能配网的业务特征,提出在主动队列管理中加权随机早检测(wRED)的改进算法WRED_gentle,通过在OPNET上建立集抄应用和配电自动化应用的业务模型,对NoWRED、WRED和WRED_gentle 3种算法进行仿真实验.结果证明WRED_ gentle算法能够明显改善网络时延、吞吐量和链路利用率等性能,保证智能配电通信网的QoS要求.