公開分享網路上提供影音串流服務之頻寬管理問題研究

A Study of Bandwidth Management Problems for Video Streaming Services on Public-Shared Networks

張宏義(Hong-Yi Chang)

指導教授 ： 黃能富

國立清華大學/電機資訊學院/資訊工程學系/博士(2010年)

摘要

近來這幾年透過用戶端來架構一個網路系統的觀念被提出來，當使用者分享所擁有的頻寬讓大家使用，此使用者就有機會在Internet任何角落使用其他使用者所分享的頻寬存取網路。我們定義此觀念為“Public-shared Network”，至目前為止FON是這世界上唯一符合“Public-shared Network”的觀念。我們提出了一個嶄新的觀念透過“Public-shared Network”所分享的頻寬來建置一個有效率的影音串流傳遞系統。然而Internet的可用頻寬經常變動，用戶端可能會因頻寬的縮減而導致無法正常的接收影音串流，為了讓用戶端能在不同的頻寬變動得狀況下正常接收影音串流，因此我們整合了可調式編碼的技術(SVC；Scalable Video Coding)至我們所設計的影音串流傳遞系統中，但在論文中我們證明使用最少的分享頻寬來建置一個滿足不同頻寬用戶端的SVC影音串流傳遞系統為NP-Hard的問題，因此我們試著使用split/merge的方法來建置一個滿足不同頻寬用戶端具有可擴充、強健及可靠的影音串流傳遞系統。在論文中我們提出兩個最佳化的演算法能使得使用最少的分享頻寬來建置一個滿足不同頻寬用戶端，在論文中我們也設計了頻寬資源回收再利用的機制增加用戶端接收影音串流的延遲及減少網路中的設備負擔及流量本地化的機制減少頻寬的耗用。除此之外，我們將頻寬管理問題擴展至雲端及利用矩陣來簡化運算，我們提出使用最少的雲端伺服器及SVC技術來建置一個有效率及具擴展性的雲端影音串流傳遞系統稱為Cloud Video Streaming Platform (CVSP)。本論文所提出的影音串流傳遞系統已被實作出來並印證論文中所提出來的所有觀點。

關鍵字

即時影音串流 ； 公開分享網路 ； 可調式編碼

並列摘要

Recently, the idea of constructing the networking system by users themselves has been presented. With the concept of sharing user’s own bandwidth to the public, users have the opportunities to access the Internet anywhere via the shared bandwidth. We call this concept the “Public-shared Network”. At present, FON is the only public-shared network in the world. We propose a novel concept of using the sharable bandwidth of public-shared network, like FON network, to construct an efficient video delivery system. However, as the available Internet bandwidth changed frequently and dynamically, it is possible that clients are unable to receive the video streaming due to lack of bandwidth. For clients with different bandwidth to receive the video streaming successfully and to use the Internet bandwidth more efficiently, the scalable video coding is considered to construct the video streaming delivering system. The NP-Hard problem then arises about as to constructing architecture to serve the video streaming requests by using a minimum amount of sharable bandwidth. Then we try to use the split-merge method to construct a scalable, robust, and high availability video delivery system as complete binary tree structures. Two optimal problems are defined to arrange the public-shared bandwidth so that all clients are served while minimizing system resource usage. This dissertation also designs the resource management scheme for resource recycling and reuse, which improves the streaming continuity experienced by clients and reduces the overall system load experienced by devices, and the traffic-localized mechanism is proposed to saving Internet link bandwidth. In addition, we have extended the problem of bandwidth management to the cloud and the matrix operation method is proposed to simplify the notation and operation. We propose the cloud video delivery system called Cloud Video Streaming Platform (CVSP) that uses public cloud servers, to construct an efficient and scalable video streaming delivery platform with Scalable Video Coding (SVC) technology. The proposed system is also implemented to illustrate the feasibility of the whole concept.

並列關鍵字

Live Video Streaming ； Public-shared Network ； Scalable Video Coding

參考文獻

[1] Taehyun Kim, M.H. Ammar, "A comparison of heterogeneous video multicast schemes: Layered encoding or stream replication," IEEE Transactions on Multimedia, Vol. 7, Issue 6, pp.1123-1130, 2005.

[2] Wu, Y. T. Hou, and Y.-Q. Zhang, "Scalable video coding and transport over broad-band wireless networks," Proceedings of the IEEE, vol. 89, pp. 6-20, 2001

[3] Qian Zhang, Quji Guo, Qiang Ni, Wenwu Zhu, Ya-Qin Zhang, "Sender-adaptive and receiver-driven layered multicast for scalable video over the Internet," IEEE Trans. on Circuits and Systems for Video Technology, Vol. 15, Issue 4, pp.482 – 495, 2005.

[11] Zhang, S. Jaminand, and L. Zhang, “Host Multicast: A Framework for Delivering Multicast to End Users,” IEEE INFOCOM, 23-27 June 2002, New York, USA, pp.1366-1375.

[16] Nen-Fu Huang, Hong-Yi Chang, Yuan-Wei Lin, and Kuo-Shiang Hsu, “A Novel Bandwidth Management Scheme for Video Streaming Service on Public-Shared Network,” IEEE ICC 2008, Beijing, China, May 2008.