论文部分内容阅读
分组调度在异步光分组交换(OPS)节点中起着重要作用,一种智能的调度算法能减小分组的丢失率,增加节点的吞吐量,从而提高服务质量。文章提出了一种基于队列长度和时延约束的解决队列中分组“饿死”和时延问题的权重调度算法。算法通过计算队列长度和分组在队列中等待时间确定调度的权重值,以减小分组在核心节点处的资源竞争和等待时延,调度器计算每个输入端口的N×K个虚拟输出队列的权重值并调度最大权重值队列到可用的输出信道上。分析和仿真结果表明所提的权重调度算法具有高吞吐量和低分组丢失率特点。
Packet scheduling plays an important role in asynchronous optical packet switching (OPS) nodes. An intelligent scheduling algorithm can reduce the packet loss rate, increase the throughput of nodes and improve the service quality. This paper presents a weight scheduling algorithm based on queue length and delay constraint to solve the problem of packet “starvation” and delay in the queue. The algorithm determines the scheduling weight value by calculating the length of the queue and the waiting time of the packet in the queue so as to reduce the resource competition and the waiting delay of the packet at the core node. The scheduler calculates the N × K virtual output queues of each input port Weights and schedules the queue with the largest weight to the available output channels. The analysis and simulation results show that the proposed weight scheduling algorithm has high throughput and low packet loss rate.