Scalable scheduling architectures for high-performance crossbar-based switches
Scalable scheduling architectures for high-performance crossbar-based switches
复制标题
用于基于交叉开关的高性能交换机的可扩展调度架构
DOI：
10.1109/hpsr.2004.1303441
复制
发表时间：
2004
期刊：
影响因子：
--
通讯作者：
C. Tsui
中科院分区：
文献类型：
--
作者：
Jing Liu;M. Hamdi;Qingsheng Hu;C. Tsui
本文提出了一种新的可扩展的调度结构，高性能的交叉开关为基础的虚拟输出队列（VOQ）计划。在传统的交换架构中，调度器是由一个单一的集中式调度设备实现的，相比之下，建议的调度架构连接几个小的调度设备串联和仲裁算法并行执行。由此，每个单个调度设备的输入通过考虑它们的本地传输请求以及全局输出可用性信息两者来建立到一组输出的连接。这种架构的优点在于它能够用几个小的调度设备来实现大的调度器（> 64），以及它能够实现高性能调度。我们首先介绍了一个分布式并行轮询调度算法（DPRR）的建议架构。通过对各种可接纳业务的仿真结果分析，表明DPRR的性能远优于集中式调度器上常用的其他轮循调度算法的性能。我们还证明了在Bernoulli i.i.d.统一的流量，DPRR实现100%的吞吐量。此外，我们引入了一个分布式并行循环调度算法与内存（DPRRM）作为DPRR的改进版本，使其稳定的任何允许的流量。
This paper presents a novel scalable scheduling architecture for high-performance crossbar-based switches with virtual output queuing (VOQ) scheme. In contrast to traditional switching architectures where the scheduler is implemented by one single centralized scheduling device, the proposed scheduling architecture connects several small scheduling devices in series and the arbitration algorithm is executed in parallel. Thereby the inputs of each single scheduling device establish connections to a group of outputs, by considering both their local transmission requests as well as global output availability information. The advantage of this architecture lies in its ability to implement large schedulers (> 64) with several small scheduling devices as well as in its capability to achieve high-performance scheduling. We first introduce a distributed parallel round robin scheduling algorithm (DPRR) for the proposed architecture. Through the analysis of simulation results on various admissible traffics, it is shown that the performance of DPRR is much better than the performance of other round robin scheduling algorithms commonly used on centralized schedulers. We also prove that under Bernoulli i.i.d. uniform traffic, DPRR achieves 100% throughput. Moreover, we introduce a distributed parallel round robin scheduling algorithm with memory (DPRRM) as an improved version of DPRR to make it stable under any admissible traffic.