【摘要】：隨著移動(dòng)通信技術(shù)的不斷發(fā)展,移動(dòng)通信業(yè)務(wù)量和網(wǎng)絡(luò)流量呈現(xiàn)爆炸式增長(zhǎng),傳統(tǒng)的蜂窩式接入網(wǎng)架構(gòu)已經(jīng)逐漸失去優(yōu)勢(shì)。因此,能適應(yīng)未來新環(huán)境且高性能低成本的集中式接入網(wǎng)的概念和架構(gòu)應(yīng)運(yùn)而生。中科院計(jì)算所提出的超級(jí)基站網(wǎng)絡(luò)架構(gòu)就屬于集中式接入網(wǎng)架構(gòu),通過引入基站虛擬化技術(shù)為基站處理資源的共享和靈活分配創(chuàng)造了良好的基礎(chǔ),從而減少建設(shè)和維護(hù)基站的開銷。因此,本文基于基站虛擬化中協(xié)議處理實(shí)體虛擬化展開研究,具體內(nèi)容如下:(1)針對(duì)基于KVM協(xié)議處理資源虛擬化平臺(tái)不便于對(duì)運(yùn)行在其中的協(xié)議處理實(shí)體管理的問題,本文通過研究上層管控工具Virt-manager和OpenStack對(duì)虛擬化平臺(tái)進(jìn)行管控,并采用具有不同特征的測(cè)試工具針對(duì)虛擬化平臺(tái)損耗與負(fù)載兩項(xiàng)關(guān)鍵指標(biāo)進(jìn)行測(cè)評(píng)。結(jié)果表明KVM虛擬化技術(shù)在一定程度上可以滿足超級(jí)基站對(duì)協(xié)議處理資源平臺(tái)虛擬化的性能要求。(2)針對(duì)傳統(tǒng)測(cè)試工具難以全面包含協(xié)議棧的特征,本文基于非面向連接協(xié)議的算法設(shè)計(jì)了一種能夠較全面體現(xiàn)出協(xié)議棧特征的測(cè)試程序。該程序由用戶側(cè)的基準(zhǔn)程序和基站側(cè)的基站協(xié)議棧軟件兩部分構(gòu)成,綜合考慮了實(shí)際協(xié)議棧中所需維護(hù)的多項(xiàng)信息,并對(duì)基站接收上行數(shù)據(jù)進(jìn)行模擬,最終在平臺(tái)中對(duì)程序運(yùn)行效果進(jìn)行驗(yàn)證。(3)針對(duì)協(xié)議處理實(shí)體在虛擬化平臺(tái)動(dòng)態(tài)遷移中耗時(shí)長(zhǎng)且丟包的問題,本文基于KVM的預(yù)拷貝算法,同時(shí)結(jié)合通信業(yè)務(wù)的特點(diǎn),建立了一種遷移成本模型。模型主要考慮了頻繁變化的臟頁量,通過在系統(tǒng)中添加共享存儲(chǔ)節(jié)點(diǎn)的方法以減少這種類型的臟頁量傳輸。在超級(jí)基站環(huán)境中的測(cè)試結(jié)果表明,新的架構(gòu)可在一定程度上減少遷移耗時(shí)與丟包。
[Abstract]:With the continuous development of mobile communication technology, mobile communication traffic and network traffic show explosive growth, the traditional cellular access network architecture has gradually lost its advantages. Therefore, the concept and architecture of centralized access network which can adapt to the new environment and high performance and low cost emerges as the times require. The super base station network architecture proposed by the Chinese Academy of Sciences belongs to the centralized access network architecture, which creates a good foundation for the sharing and flexible allocation of base station processing resources by introducing base station virtualization technology, thus reducing the cost of building and maintaining base stations. Therefore, this paper studies the virtualization of protocol processing entity in base station virtualization, the specific contents are as follows: (1) in view of the problem that the virtualization platform of resource processing based on KVM protocol is not convenient to manage the protocol processing entity running in it, this paper studies the upper control tools Virt-manager and OpenStack to control the virtualization platform. The test tools with different characteristics are used to evaluate the loss and load of virtualization platform. The results show that KVM virtualization technology can meet the performance requirements of super base station for protocol processing resource platform virtualization to a certain extent. (2) aiming at the fact that traditional testing tools are difficult to include protocol stack in an all-round way, this paper designs a test program based on non-connection protocol algorithm, which can fully reflect the characteristics of protocol stack. The program is composed of two parts: the reference program on the user side and the base station protocol stack software on the base station side. Considering many kinds of information needed to be maintained in the actual protocol stack, and simulating the uplink data received by the base station, the running effect of the program is verified in the platform. (3) aiming at the problem that the protocol processing entity takes a long time and loses packets in the dynamic migration of the virtualization platform, this paper is based on the pre-copy algorithm of KVM. At the same time, according to the characteristics of communication service, a migration cost model is established. The model mainly considers the frequently changing number of dirty pages, and reduces this type of dirty page transmission by adding shared storage nodes to the system. The test results in super base station environment show that the new architecture can reduce migration time and packet loss to a certain extent.
【學(xué)位授予單位】：蘭州理工大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TN929.5

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 孫茜;田霖;周一青;石晶林;;基于NFV與SDN的未來接入網(wǎng)虛擬化關(guān)鍵技術(shù)[J];信息通信技術(shù);2016年01期

2 伍陽;;基于Docker的虛擬化技術(shù)研究[J];信息技術(shù);2016年01期

3 費(fèi)薇;李淑英;衷宜;徐建;;基于預(yù)拷貝的虛擬機(jī)內(nèi)存狀態(tài)動(dòng)態(tài)遷移優(yōu)化[J];南京理工大學(xué)學(xué)報(bào);2015年05期

4 王森;朱常鵬;韓博;;一種基于qemu的動(dòng)態(tài)遷移模型[J];計(jì)算機(jī)科學(xué);2015年S1期

5 陳清金;陳存香;張巖;;Docker技術(shù)實(shí)現(xiàn)分析[J];信息通信技術(shù);2015年02期

6 CHEN Shanzhi;SUN Shaohui;WANG Yingmin;XIAO Guojun;Rakesh Tamrakar;;A Comprehensive Survey of TDD-Based Mobile Communications Systems from TD-SCDMA 3G to TDLTE-Advanced 4G and 5G directions[J];中國通信;2015年02期

7 雷秋燕;張治中;程方;胡昊南;;基于C-RAN的5G無線接入網(wǎng)架構(gòu)[J];電信科學(xué);2015年01期

8 楊懋;楊旭;李勇;金德鵬;蘇厲;曾烈光;;基于虛擬化的軟件定義無線接入網(wǎng)結(jié)構(gòu)[J];清華大學(xué)學(xué)報(bào)(自然科學(xué)版);2014年04期

9 姚華超;王振宇;;基于KVM-QEMU與Libvirt的虛擬化資源池構(gòu)建[J];計(jì)算機(jī)與現(xiàn)代化;2013年07期

10 田霖;翟國偉;黃亮;周一青;石晶林;;基于集中式接入網(wǎng)架構(gòu)的異構(gòu)無線網(wǎng)絡(luò)資源管理技術(shù)研究[J];電信科學(xué);2013年06期

相關(guān)碩士學(xué)位論文 前7條

1 盧瑩;5G基站虛擬化基站資源管理關(guān)鍵技術(shù)研究[D];華北電力大學(xué)(北京);2016年

2 鐘林;基于NFS的云存儲(chǔ)網(wǎng)關(guān)關(guān)鍵技術(shù)研究及系統(tǒng)實(shí)現(xiàn)[D];華南理工大學(xué);2015年

3 張苗;面向分布式計(jì)算的網(wǎng)絡(luò)編程框架設(shè)計(jì)與實(shí)現(xiàn)[D];電子科技大學(xué);2015年

4 林春;基于KVM設(shè)備虛擬化技術(shù)的研究[D];西安電子科技大學(xué);2014年

5 李媛;Xen虛擬機(jī)動(dòng)態(tài)遷移機(jī)制優(yōu)化策略的研究與設(shè)計(jì)[D];北京郵電大學(xué);2014年

6 任劍;無線基站虛擬化研究及實(shí)現(xiàn)[D];中國科學(xué)院大學(xué)(工程管理與信息技術(shù)學(xué)院);2013年

7 賀佰元;面向電信業(yè)務(wù)的Xen虛擬機(jī)遷移的研究及優(yōu)化[D];北京郵電大學(xué);2013年

，

本文編號(hào)：2502025