本文選題：網(wǎng)絡虛擬化 + 虛擬鏈路　； 參考：《國防科學技術(shù)大學》2014年碩士論文

【摘要】：網(wǎng)絡虛擬化技術(shù)是通過在相同的物理網(wǎng)絡基礎設施上進行資源的切分,構(gòu)建多個并存虛擬網(wǎng)絡的基礎支撐技術(shù)。不同的虛擬網(wǎng)絡內(nèi)運行的流量互不干擾,從而可以在不影響現(xiàn)有網(wǎng)絡工作的情況下,用來實驗新的網(wǎng)絡體系結(jié)構(gòu);不同的虛擬網(wǎng)絡,也可以運行異構(gòu)的協(xié)議棧,為不同的應用提供不同的服務保障。因而,網(wǎng)絡虛擬化技術(shù)是當前學術(shù)研究的熱點,可以解決物理網(wǎng)絡資源的利用率低下、新型的體系結(jié)構(gòu)無法進行充分的實驗、許多新增業(yè)務的不同需求難以同時滿足等問題。但是由于虛擬網(wǎng)絡的鏈路依托于物理網(wǎng)絡鏈路,存在拓撲緊耦合問題,使得虛擬網(wǎng)絡的拓撲無法獨立于物理網(wǎng)絡,按照用戶的需求動態(tài)創(chuàng)建。本文為了克服上述問題,結(jié)合網(wǎng)絡虛擬化技術(shù)和現(xiàn)有的交換機制,提出了基于多跳虛擬鏈路的快速交換機制。該機制在虛擬網(wǎng)絡中使用標簽交換的設計理念,在多跳虛擬鏈路入口結(jié)點處打上到多跳虛擬鏈路出口結(jié)點的“標記”,中間結(jié)點不處理網(wǎng)絡層協(xié)議,進行多跳虛擬鏈路報文的快速交換。為提高虛擬網(wǎng)絡可靠性,本文在結(jié)合現(xiàn)有故障修復方案的基礎上,提出了基于彈性標簽的多跳虛擬鏈路快速交換機制。本文的主要研究成果如下:1)提出了基于多跳虛擬鏈路的網(wǎng)絡虛擬化模型。該模型基于構(gòu)建多跳虛擬鏈路的需求,對現(xiàn)有網(wǎng)絡虛擬化模型進行了改進,將現(xiàn)有的網(wǎng)絡虛擬化模型的數(shù)據(jù)平面縱向切分為轉(zhuǎn)發(fā)引擎數(shù)據(jù)平面和多跳虛擬鏈路數(shù)據(jù)平面,轉(zhuǎn)發(fā)引擎平面共享多跳虛擬鏈路平面,實現(xiàn)虛擬網(wǎng)絡接口報文,在多跳虛擬鏈路平面中的快速交換。2)提出了轉(zhuǎn)發(fā)引擎數(shù)據(jù)平面和多跳虛擬鏈路數(shù)據(jù)平面間的虛擬接口映射機制。該機制可以靈活完成多跳虛擬鏈路數(shù)據(jù)平面的物理接口到轉(zhuǎn)發(fā)引擎數(shù)據(jù)平面的虛擬接口的映射,在發(fā)送報文給轉(zhuǎn)發(fā)引擎數(shù)據(jù)平面時,將物理接口映射為虛擬接口,在收到轉(zhuǎn)發(fā)引擎數(shù)據(jù)平面的報文時,將虛擬接口映射為物理接口。3)提出了基于彈性標簽的多跳虛擬鏈路抗毀機制。該機制主要為虛擬網(wǎng)絡中的報文提供備份路徑轉(zhuǎn)發(fā),在鏈路或者節(jié)點故障時,能夠?qū)笪倪M行本地處理或者回溯到上游節(jié)點進行處理,使數(shù)據(jù)報文具有“彈性”,提高了虛擬網(wǎng)絡的可靠性。4)使用click開源軟件,實現(xiàn)了基于彈性標簽的多跳虛擬鏈路交換機制的原型系統(tǒng),對多跳虛擬鏈路交換機制和基于彈性標簽的多跳虛擬鏈路的轉(zhuǎn)發(fā)性能進行了測試,通過對比實驗結(jié)果,分析了轉(zhuǎn)發(fā)性能的影響因素。
[Abstract]:Network virtualization technology is a basic supporting technology to construct multiple virtual networks by splitting resources on the same physical network infrastructure. The traffic running in different virtual networks does not interfere with each other, so that they can be used to experiment with new network architectures without affecting the work of existing networks, and different virtual networks can also run heterogeneous protocol stacks. Provide different services for different applications. Therefore, network virtualization technology is a hot topic in current academic research. It can solve the problems of low utilization of physical network resources, inadequate experiments of new architecture, and the difficulty of meeting the different requirements of many new services. But because the link of virtual network depends on the link of physical network, there exists the problem of topological tight coupling, which makes the topology of virtual network unable to be independent of physical network, which is created dynamically according to the user's demand. In order to overcome the above problems, a fast switching mechanism based on multi-hop virtual link is proposed by combining the network virtualization technology and the existing switching mechanism. In this scheme, label switching is used in the virtual network to mark the exit node of the multi-hop virtual link at the entrance node of the multi-hop virtual link, and the middle node does not deal with the network layer protocol. Fast exchange of multi-hop virtual link packets. In order to improve the reliability of virtual network, this paper proposes a fast switching mechanism of multi-hop virtual link based on elastic label based on the existing fault repair schemes. The main research results of this paper are as follows: 1) A network virtualization model based on multi-hop virtual link is proposed. Based on the requirement of constructing multi-hop virtual link, this model improves the existing network virtualization model and divides the data plane of the existing network virtualization model into forwarding engine data plane and multi-hop virtual link data plane vertically. The forwarding engine plane shares the multi-hop virtual link plane to realize the virtual network interface packets. The fast switching between the forwarding engine data plane and the multi-hop virtual link data plane is proposed. The virtual interface mapping mechanism between the forwarding engine data plane and the multi-hop virtual link data plane is proposed. This mechanism can flexibly map the physical interface of the multi-hop virtual link data plane to the virtual interface of the forwarding engine data plane, and map the physical interface to the virtual interface when sending packets to the forwarding engine data plane. When receiving the packets from the data plane of the forwarding engine, the virtual interface is mapped to the physical interface. 3) A resilient label based multi-hop virtual link survivability mechanism is proposed. This mechanism mainly provides backup path forwarding for packets in virtual network. In the event of link or node failure, the mechanism can process packets locally or back to upstream nodes, making data packets "flexible". Using click open source software, the prototype system of multi-hop virtual link switching mechanism based on elastic label is implemented. The multi-hop virtual link switching mechanism and the forwarding performance of multi-hop virtual link based on elastic label are tested. By comparing the experimental results, the factors affecting forwarding performance are analyzed.
【學位授予單位】：國防科學技術(shù)大學
【學位級別】：碩士
【學位授予年份】：2014
【分類號】：TP393.0

【參考文獻】

相關(guān)期刊論文 前1條

1 陳利兵;;BFD技術(shù)在IP承載網(wǎng)中的應用[J];現(xiàn)代電信科技;2008年01期

，

本文編號：2083520