摘要
在引入优先级因素的流量预测模型基础上,提出了一种多优先级队列的缓存管理算法,该算法通过固定时刻的缓存预分配,结合时刻间隔中的缓存竞争,使缓存的分配更加合理而有预见性,仿真实验表明它能够明显降低高优先级数据包的丢包率,同时兼顾低优先级数据包的转发,并且不产生额外的延迟,表明该算法有很好的应用价值。
Based on traffic forecasting which imports priority, a buffer management algorithm of multiple priority queues is presented. This algorithm integrates buffer pre-dispensation in fixed points of time with buffer competition between the points of time, and makes huffer dispensation have the character of rationality and forecasting. Simulation results show that the ratio of high priority packets dropped is brought down obviously, and the new algorithm pays the attention to the transmission of low priority packets, and doesn't import extra delay.
出处
《计算机工程》
CAS
CSCD
北大核心
2006年第10期116-118,共3页
Computer Engineering
关键词
优先绂
缓存管理
缓存竞争
流量预测
算法
Priority
Buffer management
Buffer competition
Traffic forecasting
Algorithm