一种低延迟的可扩展调度算法

提出了一种基于阈值的分布式迭代算法。与现有算法不同的是，该算法针对可扩展网络交换调度结构的特点，为处于最高优先级的调度器安排了2次迭代，第1次迭代用阈值方法找出一些较长的VOQ（虚拟输出队列），并在最高优先级时隙之前的一个时隙完成，以缩短信号的处理时间。仿真结果表明，该算法与现有算法相比，在大流量的uniform流量模式下，延时性能和吞吐率获得了明显的提高。同时，该算法的硬件代价小，有效地实现了性能和复杂度的良好折中。

A Low Latency Fair Scalable Scheduling Architecture

After reviewing the iterative algorithms and self-adaptive threshold algorithms,we represent a new algorithm based on the fair scalable scheduling architecture: iterative self-adaptive threshold FSSA.By employing this new algorithm,the performance of the system improves greatly at a relatively low cost.This research focuses on the efficiency of the scheduling cycle with highest superiority to achieve high performance at low cost.The result shows that compared with FSSA,the new algorithm achieves better performance on average delay and throughput,especially under uniform traffic model.