摘要
为了解决光分组在核心节点处的竞争问题,提出一种基于有界延迟的光缓存队列模型。在缓存容量有限的基础上,采用有界延迟机制,当分组到达时计算它的延迟,若延迟没有超过预设值,就将它放入缓冲区,否则丢弃分组。仿真结果表明,该模型能够降低光分组缓存时延,有效避免光分组在交换输出端的竞争,提高光缓存器性能,展示出不同于传统光缓存队列的一些特点。
To solve the contention problem of optical packets in core node, a novel queuing model of optical buffer based on the bounded delay is presented. The bounded delay mechanism on the basis of limited capacity is adopted, when a packet arrives, its delay is calculated, if the delay does not exceed the preset value, it is placed into the buffer, or the packet is dropped. The simulation results show that the model can reduce buffer delay of optical packets, avoid contention of optical packets in switching output ports effectively, and improve the performance of optical buffer. Some features of the model that differ from traditional optical buffer queue are demonstrated.
出处
《计算机工程与设计》
CSCD
北大核心
2014年第5期1506-1509,1520,共5页
Computer Engineering and Design
基金
国家自然科学基金项目(61262074
61162008
61163059)
广西可信软件重点实验室开放课题基金项目(kx201101)
广西高校优秀人才计划基金项目(桂教人201065)
广西自然科学回国基金项目(2012GXNSFCA053009)
关键词
全光网
光缓存器
竞争解决
有界延迟
网络仿真
all-optical network
optical buffer
contention resolution
bounded delay
network simulation
