HCBS混合光交换网络研究设计

【作者】 张瑾；

【导师】 徐世中；

【作者基本信息】 电子科技大学 ， 通信与信息系统， 2010， 硕士

【摘要】 随着科技的进步,宽带通信网络有了巨大的发展。但企业及个人用户对计算机网络应用的不断增多以及点到点多媒体流和基于IP的实时业务的快速增长使得网络通信量猛增,现有网络已经不堪重负,基于波分复用(WDM)的光网络正在成为主要的下层传送网络。WDM光网络包括基于“电路交换”模式的光电路交换(OCS)网络和基于“分组交换”模式的光突发交换(OBS)网络及光分组交换(OPS)网络。OCS适用于传送持续时间长、延时要求高的业务流,OPS/OBS能较好的支持突发性业务,但OPS还无法大面积商用,OBS则存在丢包率高、延时较大等缺陷。鉴于此,研究人员提出了将多种光交换技术结合的混合光交换网络。第一章主要介绍现有的混合光交换网络,主从型混合光交换网络、平行型混合光交换网络以及集成型混合光交换网络。分别从网络结构、节点功能、转发流程等方面对三者进行分析,并指出了各自的问题。第二章提出了一种新的OCS/OBS混合光交换组网方案——混合光电路突发交换(Hybrid optical Circuit and Burst Switching, HCBS)。该方案首先根据节点对之间的数据业务需求,建立由OCS光路构成的虚拓扑(OCS层)。然后用链路上的剩余的波长资源(信道)构建OBS网络(层)。当网络中出现故障,导致一条或多条OCS光路中断时,可以用已构建的OBS层,将受影响的数据分组组装成突发包发送,尝试恢复数据传送。在网络正常工作时,数据业务均通过OCS层发送。当出现突发业务导致OCS光路过载时,可以通过OBS发送超出OCS层传送容量的部分数据。HCBS由OBS层来提供OCS层光路的保护和应对流量的突发性,提高了资源的利用率。为了验证HCBS的可行性和性能,在第三章中作者借助网络仿真软件搭建了仿真平台,从丢包率、延时、多失效对抗能力三个方面将HCBS、带保护光路的OCS、单一OBS网络进行了对比。仿真结果表明,在不同的网络故障场景下,HCBS均能够较好地恢复数据业务的传送;在网络正常工作时,也能较好地应对突发业务。第四章对HCBS做了进一步的优化设计,借助多拓扑路由(MTR)的思想从OBS层快速故障恢复和OBS层流量工程两个方面对HCBS做了改进,对提出的方案进行了仿真验证。第五章介绍了HCBS仿真平台。第六章对全文做了总结。更多还原

【Abstract】 As the development of science and technology, broadband communication networks have got gigantic development. But Government, finance community, telecoms as well as individual consumer’s desire for computer network applications as well as Point-to-Point multi-media stream and IP-based real time service’s striving are triggering dramatic growth of network traffic. Under this circumstance, traditional electric networks are no longer likely to bear this burden. Under this circumstance WDM based optical networks is becoming the main underlying transmission network. WDM based optical networks consist of optical circuit switching network(OCS) and the other two, OPS and OBS, whose full name are optical packet switching and optical burst switching. OCS networks are commonly used to carry flows who has long duration time and require low-latency while OPS and OBS are more suitable for bursty flows. But OPS can’t be put into large-scale commercial use right now, while OBS have performance defects such as high burst loss rate, latency. In view of this, researchers have come up with the idea of combining different switching techniques to form hybrid optical networks.The first chapter give an brief introduction of the current hybrid optical networks, namely master-slave, parallel and integrated hybrid optical networks. We analyse their network architecture, node function, packet forwarding process and pointed out each’s disadvantage. In the end, we made a comparison of the three and give a conclusion.We presents a new OCS/OBS hybrid optical switching network (HCBS) in chapter two, The OCS layer of the proposed network is composed of light paths which are established according to the traffic demands between all the source-destination nodes. And the remaining wavelength channels on each physical link are used to build the OBS layer. When one or more light path break down due to network failure(s), the affected data packets could be assembled into bursts and transported through the OBS layer. In normal operating scenario, OCS layer carries most of the traffic. If bursty traffic overwhelms the transmission capacity of OCS layer, the excess part of traffic could be sent through OBS. So the main idea of HCBS is to use OBS layer as a protection layer of OCS and a layer to deal with network flow’s burst. In order to validate the feasibility and efficiency of HCBS, we build up a software platform based on OPNET, which is a commonly used tool for addressing network planning,engineering, R&D. During the validation process, we compare HCBS with proteced OCS network, single layer OBS network from packet lost rate, latency and ability to combat mulitple failures. Simulation results show that, compared with the pure OCS network with restoration and pure OBS network, the proposed hybrid optical switching network could achieve lower packet loss rate under different network failure scenarios and could accommodate bursty data streams more efficiently.In chapter four, we further devleop HCBS from two aspects. By using the ieda of mulitple topology routing, OBS layer fast recovery scheme and OBS layer traffic engineering scheme are put forward and specified. In chapter five, the two proposed schemes are simulated on the platform we former build and analysis of the result are given. Chapter six introduce our simulation platform and the last chapter is the conclusion of this dissertation.更多还原