摘要
对OBS网络核心节点中所采取的多种改善QoS的策略进行了研究并提出了一种反馈证实机制以减少数据突发的丢失.其原理是在BCP传输建立端到端通路的过程中,如果在某核心节点处不能成功预约波长信道,该节点将在数据突发离开边缘节点之前返送NACK证实信息,取消边缘节点的该数据突发调度输出计划,并将其重新缓冲排队以备再次调度发送.这样可以尽量为数据突发提供可靠的波长路径,减少数据突发的丢失.
A new mechanism of feedback acknowledge is proposed for avoiding loss of data bursts. It is the principle that if failed to establishing end-to-end wavelength path during setting up, the fault core node will send back a acknowledge information （NACK） to the original edge node, once the original edge node receives NACK information before the data burst to be scheduled has not be output,the plan of scheduling the data burst for output will be cancelled and that data burst will be buffered again to be ready for scheduling next time. In such way,a more reliable wavelength path and a less loss may be provided for data burst.
出处
《光子学报》
EI
CAS
CSCD
北大核心
2005年第11期1706-1709,共4页
Acta Photonica Sinica
