【摘要】：下一代互聯網中，隨著寬帶和web2.0的發(fā)展，IPTV、VoD、高清視頻等流媒體應用逐漸成為寬帶應用的主流。流媒體應用所具有的高帶寬、高訪問量和高服務質量要求對以盡力而為為核心的互聯網提出了巨大的挑戰(zhàn)。然而內容分發(fā)網絡CDN(Content Delivery Network)由于資源采用靜態(tài)部署方式，會導致資源部署不足或資源浪費的問題。而利用P2P(Peer to Peer)進行流媒體分發(fā)，由于P2P網絡節(jié)點的來去自如，會造成服務質量的不穩(wěn)定。 云計算作為一種新型的網絡化計算模式，以按需付費為商業(yè)模式，可為用戶提供幾乎無限可擴展的共享虛擬化資源，這種彈性計算的特性為解決上述CDN和P2P所面臨的問題提供了機會。構建在云平臺之上的流媒體分發(fā)系統(tǒng)，能夠解決傳統(tǒng)分發(fā)存在的問題，同時可顯著降低流媒體應用提供商的運營成本，因而受到廣泛的關注。 構建基于云平臺的流媒體分發(fā)系統(tǒng)，需要面對來自各個地理區(qū)域大量高度動態(tài)的需求，為保證服務質量需要就近采用各區(qū)域中多個云服務提供商的數據中心，即需要基于分布式云數據中心架構(以下簡稱分布式云)來構建專門面向流媒體應用的分發(fā)網絡。本文針對典型的多信道和多用戶區(qū)域的應用場景，對基于分布式云的分發(fā)網絡的建立和內容分發(fā)算法進行研究，考慮了分發(fā)拓撲、各云數據中心計費方式、用戶位置和請求速率等因素，通過將其建模為有向斯坦納樹，給出的內容分發(fā)算法能為用戶提供服務質量滿足的流媒體服務，并降低分發(fā)的開銷和代價。算法經過在matlab和yalmip線性規(guī)劃工具箱的運行環(huán)境下進行驗證和性能分析，結果顯示在不同的拓撲下，給出的啟發(fā)式算法均能以較低的時間復雜度獲得近優(yōu)解，可方便地應用于商業(yè)的流媒體內容分發(fā)系統(tǒng)。
[Abstract]:With the development of broadband and web2.0, streaming media applications such as IPTV,VoD, HD video have gradually become the mainstream of broadband applications in the next generation Internet. Streaming media applications with high bandwidth, high access and high quality of service requirements pose a great challenge to the best-effort Internet. However, due to the static deployment of resources in content distribution network (CDN (Content Delivery Network), it will lead to the problem of insufficient deployment of resources or waste of resources. Because P2P network nodes come and go freely, the quality of service will be unstable when P2P (Peer to Peer) is used to distribute streaming media. Cloud computing as a new network computing model, pay-on-demand as a business model, can provide users with virtually unlimited scalability of shared virtualization resources. This kind of elastic computing provides an opportunity to solve the problems faced by CDN and P2P. Streaming media distribution system based on cloud platform can solve the problems existing in traditional distribution and significantly reduce the operating costs of streaming media application providers. To build a streaming media distribution system based on cloud platform, we need to face a large number of highly dynamic needs from various geographic regions, and to ensure the quality of service, we need to adopt the data center of multiple cloud service providers in each region in order to ensure the quality of service. That is, the distributed cloud data center architecture (hereafter referred to as distributed cloud) is needed to build a distribution network for streaming media applications. In this paper, for the typical multi-channel and multi-user application scenario, the establishment of distributed cloud-based distribution network and content distribution algorithm are studied, and the distribution topology and the billing methods of cloud data centers are considered. User location and request rate are modeled as directed Steiner tree. The proposed content distribution algorithm can provide streaming media service satisfying quality of service to users and reduce the overhead and cost of distribution. The algorithm is verified and analyzed under the running environment of matlab and Yalmip Linear programming Toolbox. The results show that under different topologies, the heuristic algorithm can get the near-solution with low time complexity. Can be easily applied to commercial streaming media content distribution system.
【學位授予單位】：鄭州大學
【學位級別】：碩士
【學位授予年份】：2014
【分類號】：TP393.02

【參考文獻】

相關期刊論文 前10條

1 吳吉義;平玲娣;潘雪增;李卓;;云計算:從概念到平臺[J];電信科學;2009年12期

2 孫健;賈曉菁;;Google云計算平臺的技術架構及對其成本的影響研究[J];電信科學;2010年01期

3 吳俊;徐溟;;公有云服務計費模式比較研究[J];電信科學;2012年01期

4 郭曉霞;張乃光;丁森華;;基于有線網的媒體定向分發(fā)技術實現初探[J];廣播與電視技術;2009年09期

5 許四平;;SaaS軟件即服務模型研究[J];硅谷;2009年04期

6 劉菲;張波;;淺談中小企業(yè)私有云計算解決方案[J];硅谷;2010年14期

7 臧運港;陳光喜;;基于P2P和CDN的混合流媒體分發(fā)模型及分析[J];計算機應用研究;2010年06期

8 王洪亮;;云計算專題(2) 云計算的起源與定義[J];科技浪潮;2010年03期

9 魏婷;劉煉;;關于P2P對等網絡研究的淺析[J];科技信息;2010年22期

10 方巍;文學志;潘吳斌;薛勝軍;;云計算:概念、技術及應用研究綜述[J];南京信息工程大學學報(自然科學版);2012年04期

，

本文編號：2434699