| 基于帧填补的MMM Clos网络按序分组交换算法 |
| |
| 引用本文： | 高雅, 邱智亮, 张茂森, 黎军. 基于帧填补的MMM Clos网络按序分组交换算法[J]. 电子与信息学报, 2012, 34(11): 2715-2720. doi: 10.3724/SP.J.1146.2012.00617 |
| |
| 作者姓名： | 高雅 邱智亮 张茂森 黎军 |
| |
| 作者单位： | 1. 西安电子科技大学综合业务网理论及关键技术国家重点实验室 西安 710071
2. 空间微波技术国家重点实验室 西安 710100 |
| |
| 基金项目： | 长江学者和创新团队发展计划(IRT0852);国家863计划项目(2008AA01A332,2009AA01A335);核高基项目(2009ZX01033-003-001)资助课题 |
| |
| 摘 要： | 与其它类型的Clos网络相比，各级带缓存的MMM(Memory-Memory-Memory)Clos网络使得交换网络的配置时间最小化，但是MMM交换中间级缓存的存在会引起输出端口的信元乱序。该文提出了一种满帧填补扩展算法(EPF)，采用逐帧转发的方式来避免MMM Clos网络中的乱序问题。新算法在输入级和输出级采用固定周期轮转方式，中间级采用最早信元优先输出调度，具有复杂度低，可分布式控制，以及不需要缓存加速的特点。分析和仿真结果表明该算法是稳定的，即输入输出为可允许业务时，算法可达100%吞吐率。
|
| 关 键 词： | 三级Clos网络 分组交换 负载均衡 分组保序 |
| 收稿时间： | 2012-05-21 |
| 修稿时间： | 2012-09-03 |
Padded-frame Based in-sequence Dispatching Scheme for Memory-Memory-Memory(MMM) Clos-network |
| |
| Gao Ya, Qiu Zhi-Liang, Zhang Mao-Sen, Li Jun. Padded-frame Based in-sequence Dispatching Scheme for Memory-Memory-Memory (MMM) Clos-network[J]. Journal of Electronics & Information Technology, 2012, 34(11): 2715-2720. doi: 10.3724/SP.J.1146.2012.00617 |
| |
| Authors: | Gao Ya Qiu Zhi-liang Zhang Mao-sen Li Jun |
| |
| Affiliation: | Gao Ya① Qiu Zhi-liang① Zhang Mao-sen① Li Jun② ①(State Key Laboratory of Integrated Service Networks,Xidian University,Xi’an 710071,China) ②(Science and Technology on Space Microwave Laboratory,Xi’an 710100,China) |
| |
| Abstract: | Compared with other types of Clos-netwok switch, the Memory-Memory-Memory (MMM) switch minimizes the time required to configure a Clos-network switch. However the use of buffers in the middle stage of an MMM switch can potentially cause the forwarding of cells to the outputs in out-of-sequence order. A novel scheduling algorithm, called Extended Padded Frame (EPF) scheme, is proposed, which provides in-sequence service for MMM Clos-network by adopting a frame-based approach. The new solution adopts a predetermined cyclic shift configuration at the first and third stages and the oldest-cell-first selection for arbitrations at the second stage. The algorithm is decentralized and requires no speedup. Analysis and simulation results show that the new algorithm is stable, that is to say, EPF can achieve 100% throughput under any admissible traffic patterns. |
| |
| Keywords: | Three-stage Clos networks Packet switching Load balancing Maintaining packets order |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
| 点击此处可从《电子与信息学报》浏览原始摘要信息 |
|
点击此处可从《电子与信息学报》下载全文 |
