臺灣博碩士論文加值系統

隨著網路頻寬的快速增長，人們越來越流行透過網路線上觀看影片。為了應付多變的網路狀況，影音串流服務的供應商如Youtube和Netflix，皆透過DASH技術來將影音內容傳輸給他們的使用者們。DASH Server會根據DASH Client的頻寬來動態的調整所播放影音的bitrate。為了瞭解在地下鐵系統中的網路頻寬變化，本研究實作了一款行動APP，用來測量行動裝置在搭乘台北捷運的過程中，可使用的網路頻寬頻寬，並將此網路頻寬變化資料進行分析。本研究發現，當乘客在隧道內可使用的網路頻寬比在捷運站台中可使用的網路頻寬大幅下降，對於DASH影音串流技術的解析度切換，這種劇烈的網路頻寬變化時，常會導致播放暫停的狀況。為了解決這個問題，本研究設計了一套M-Lowp演算法，此M-Lowp演算法可以有效預測捷運行駛過程中的網路頻寬變化，並動態的調整影音解析度，使得影音在地下鐵系統中能順暢的播放，藉此來提升使用者體驗質量。體驗質量(Quality of Experience, QoE)是一種用戶主觀意見，在本篇研究中，本研究透過下列的QoE指標，包括影片凍結次數(freeze counts)、影片凍結時間(freeze duration)、平均影音質量(average bitrate)、解析度切換次數(switch counts)與低解析度video segment的數量(number of low quality video segments)。在本研究中，我們將測量到的實際網路頻寬變化分別使用不同的DASH影音串流演算法進行測試，包含Buffer level演算法與Rate adaption演算法。本研究並訂定了一套雷達圖表示法，根據本研究的實驗結果，本研究提出的M-Lowp演算法表現的比先前的演算法，在上述的QoE指摽評比下，都有更好的表現。

With the rapid increase of network bandwidth, it becomes popular for people to watch video over the Internet. To cope with dynamic and heterogeneous network condition, video-service providers, e.g., YouTube and Netflix, use DASH streaming technology (Dynamic Adaptive Streaming over HTTP) to serve contents to their users. A DASH server dynamically adjusts rate of streaming to a client according to the client’s available bandwidth. In this talk, we will present our studies on DASH streaming to passengers in an underground rapid transit. People taking mass rapid transit, such as Taipei MRT, are used to watch streaming video during their trips. We start our study with developing an APP to measure network bandwidth available to a mobile device on an MRT trip and compiled our measurements in Taipei MRT into an archive. We observed that bandwidth available during a passenger’s cart enters and leaves a station is usually much higher than that during the cart travels in the tunnel. It is a challenge for streaming algorithm to adapt to dramatic changes in the amount of bandwidth. We then use the data set to benchmark different streaming algorithms including rate adaptation and buffer control. We also designed a streaming algorithm M-Lows which predicts network bandwidth in underground MRT and dynamically adjust video resolution to optimize quality of experience of watching video for users taking Taipei MRT. The quality of experience (QoE) is an index of users’ subjective opinions. In this study, the following QoE indices are used: video freeze times, video freeze duration, average video playback rate, number of resolution-switching events, and number of low resolution video segments. From our experimental results, it is shown that our proposed M-Lows scheduling performs better than the previously published algorithms.

第一章 簡介 1
1.1 研究動機 1
1.2 論文架構 6
第二章 相關研究 7
2.1 Dynamic Adaptive Streaming over HTTP (DASH) 7
2.2 DASH Scheduling Problem Formulation 10
2.3 Buffer Level Algorithm 12
2.3 Rate Adaption Algorithm 14
2.4 Low-To-High(L2H)/L2H Buffer(L2HB) Algorithm 16
2.5 M-Low Algorithm 19
第三章 研究問題的定義 21
3.1 定義M-Low Optimal 22
第四章 本研究提出的演算法 24
4.1演算法架構與運作流程 24
4.2演算法中所使用的符號 25
4.3 Lower envelope(LE)演算法 27
4.4 M-Low Optimal(M-Lowo)演算法 33
4.5 M-Low Prediction (M-Lowp) algorithm 43
4.5.1 網路頻寬測量與建檔 43
4.5.2 網路頻寬測量APP 45
4.6 頻寬預測與調整機制 47
4.6.1 Cold start prevention機制 50
4.6.2 Freeze prevention機制 51
第五章 實驗結果 53
5.1 QoE Ability Chart 55
5.2 On-off simulation 56
5.3 Real trace simulation 62
第六章 結論與未來展望 67
參考文獻 69

[1]I. Cisco, “Cisco visual networking index: Global mobile data traffic forecast update 2015-2020 white paper,” Available: http://www.cisco.com/c/en/us/solutions/collateral/service-provider/visual- networking-index-vni/mobile-white-paper-c11-520862.html
[2]Youtube : https://www.youtube.com
[3]Netflix : https://www.netflix.com
[4]O. Oyman, S. Singh, "Quality of experience for HTTP adaptive streaming services", IEEE Commun. Mag., vol. 50, no. 4, pp. 20-27, Apr. 2012.
[5]Thomas Stockhammer, Dynamic adaptive streaming over HTTP --: standards and design principles, Proceedings of the second annual ACM conference on Multimedia systems, February 23-25, 2011, San Jose, CA, USA [doi>10.1145/1943552.1943572]
[6]Taipei metro : https://www.metro.taipei/cp.aspx?n=91974F2B13D997F1
[7]A. Zambelli, “Iis smooth streaming technical overview,” Microsoft Corporation, vol. 3, 2009.
[8]M. Levkov, “Video encoding and transcoding recommendations for http dynamic streaming on the adobe® flash® platform,” White Paper, Adobe Systems Inc, 2010.
[9]R. Pantos, Apple Inc, “Http live streaming,” IETF draft, Available: https://tools.ietf.org/html/draft-pantos-http-live-streaming-19 (last access: July 2016).
[10]R. Fielding, J. Gettys, J. Mogul, H. Frystyk, L. Masinter, P. Leach, T. Berners-Lee, Hypertext transfer protocol - HTTP/1.1, June 1999.
[11]G. Pierre , M. van Steen, Globule: a collaborative content delivery network, IEEE Communications Magazine, v.44 n.8, p.127-133, August 2006 [doi>10.1109/MCOM.2006.1678120]
[12]T. Stockhammer. “Dynamic Adaptive Streaming over HTTP - standards and design principles,” ACM International Multimedia Conference, In Proceedings of the 2nd annual ACM conference on Multimedia systems, pp. 133-144, 2011.
[13]C. Müller, S. Lederer and C. Timmerer, “An Evaluation of Dynamic Adaptive Streaming over HTTP in Vehicular Environments,” ACM International Multimedia Conference, In Proceedings of the 4th Workshop on Mobile Video, pp. 37-42, 2012.
[14]C. Liu, I. Bouazizi and M. Gabbouj, “Rate Adaptation for Adaptive HTTP Streaming,” ACM International Multimedia Conference, In Proceedings of the second annual ACM conference on Multimedia systems, pp. 169-174, 2011.
[15]Y. Liu, W. Chung, M. Lee, K. Wang, S. Chang, C. Wu, R. Chang and J. Ho. “Scheduling of Optimal DASH Streaming,” ACM International Conference Proceeding Series, In Proceedings of the International Conference on Internet of things and Cloud Computing, Article No. 43, 2016.
[16]Android : https://www.android.com/
[17]Chunghwa Telecom : https://www.cht.com.tw/
[18]Tableau : https://www.tableau.com/