Research of Two Optical Data Center Network Frameworks with High Throughput and Low Latency

XIONG Li-ting; ZHANG Shao-biao; JIE Xu-han

doi:10.13756/j.gtxyj.2020.06.005

您当前的位置：

首页 >

文章列表页 >

Research of Two Optical Data Center Network Frameworks with High Throughput and Low Latency

- Research of Two Optical Data Center Network Frameworks with High Throughput and Low Latency
- Study on Optical communication Issue 6, Pages: 17-20(2020)
- 作者机构：
  
  华东交通大学理工学院
- 作者简介：
- 基金信息：
- DOI：10.13756/j.gtxyj.2020.06.005
  CLC： TP308;TP393.02
- Published：2020，
扫描看全文
[1]熊丽婷,张绍彪,揭吁菡.两种高吞吐量低延迟光数据中心网络架构研究[J].光通信研究,2020(06):17-20+76.

XIONG Li-ting, ZHANG Shao-biao, JIE Xu-han. Research of Two Optical Data Center Network Frameworks with High Throughput and Low Latency. [J]. (6):17-20(2020)
[1]熊丽婷,张绍彪,揭吁菡.两种高吞吐量低延迟光数据中心网络架构研究[J].光通信研究,2020(06):17-20+76. DOI： 10.13756/j.gtxyj.2020.06.005.

XIONG Li-ting, ZHANG Shao-biao, JIE Xu-han. Research of Two Optical Data Center Network Frameworks with High Throughput and Low Latency. [J]. (6):17-20(2020) DOI： 10.13756/j.gtxyj.2020.06.005.

摘要

针对电数据中心网络（DCN）延迟高和吞吐量低等问题

文章提出了两种无源光DCN架构

一种针对小型DCN

一种针对大型DCN。两种架构均采用了阵列波导光栅路由器（AWGR）、耦合器和解复用器等无源光器件

并针对每种规模的架构给出了波长分配和分组传输方法。仿真结果表明

与电DCN（如Fat-tree和扁平蝶（FBFLY）网络）相比

所提方法具有更低的延迟和更高的吞吐量。而与其他光互连架构（如数据中心光开关（DOS）和Petabit）相比

其性能大致相当

分组延迟低于18μs

但成本更低。

Abstract

Aiming at the problem of high latency and low throughput in electric Data Center Network（DCN）

two kinds of passive optical DCN architecture are proposed

one for small DCN and the other for large DCN. Using passive optical devices such as Arrayed Waveguide Grating Router（AWGR）

coupler and demultiplexer

wavelength assignment and packet transmission algorithms are proposed for each scale architecture. The results show that the proposed method has lower latency and higher throughput compared with electric DCN

such as Fat-tree and Flat Butterfly（FBFLY） networks. It is also shown that the performances of proposed optical DCN architectures are similar to those of other optical interconnection architectures（such as Datacenter Optical Switch（DOS） and Petabit）. The packet delay is less than 18 μs with lower cost.

关键词

数据中心网络阵列波导光栅路由器波长分配延迟吞吐量

Keywords

DCNAWGRwavelength assignmentlatencythroughput

references

更多指标>

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Design of Data Flow Scheduling Algorithm Under SDN Architecture

A Novel Subcarrier Allocation Scheme in OFDM-PON

Research on a fan-out splitting-supporting optical multicast scheduling algorithm

Study on OCDM-based optical code-label routing information and parallel processing

Research on T-MPLS based routing and wavelength assignment

Related Author

No data

Related Institution

Fuzhou Institute of Technology

Guomai Information College,Fujian University of Technology

Key Lab of Specialty Fiber Optics and Optical Access Networks,Shanghai University

Key Lab of Optical Fiber Communication Technology,Chongqing University of Posts and Telecommunications

College of Information Science and Engineering,Yanshan University

⁰