降低點對點網路電視系統中瞬間壅塞之批次加入方法研究

A Batch Join Scheme for Flash Crowd Reduction in IPTV Systems

林棨山(Chi-Shan Lin)

指導教授 ： 鍾添曜

元智大學/資訊學院/資訊工程學系/碩士(2011年)

https://doi.org/10.6838/YZU.2011.00329

摘要

點對點影音串流技術對於網際網路電視廣播技術來說已經是一個很有效率而且也廣受歡迎的方法。但是目前點對點網路電視系統總是面臨著瞬間壅塞效應這個巨大的挑戰。瞬間壅塞效應意味著在很短的時間內，成千上萬個使用者同時加入一個廣受歡迎的頻道。當這個問題發生的時候，大量使用者在加入一個特定頻道時，將面臨到巨大的延遲，甚至加入請求很可能直接被丟棄。過去，很少有研究針對這個議題進行深入探討並提出有效的解決辦法。在這篇論文中，我們訂定出幾個會引起瞬間壅塞效應的關鍵要素：加入請求的壅塞，網路控制訊息的管理瓶頸以及點對點系統拓樸維護的管理瓶頸。基於我們的分析中，我們提出一個利用批次處理的方式取代現有系統中依序處理使用者請求的方法。我們的批次處理方法同時也解決現有系統只會利用活動中的節點作為服務提供者的缺點。我們會同時利用活動中以及新加入的節點共同建立一個虛擬子樹。實驗數據中顯示出我們的批次加入方法可以很有效的降低瞬間壅塞所造成的影響，並且很明顯的可以看出，無論是節點失敗率或使用者重新發送加入請求的次數，都大量的被降低。

關鍵字

網路電視 ； 即時影音串流 ； 點對點 ； 獎勵機制 ； 頻寬感知 ； 地域感知

並列摘要

Peer-to-peer (P2P) streaming is an effective and popular approach for large scale television multicasting over Internet. However, it is always a major challenge when thousands of peers join a popular P2P IPTV channel in a short time, so called the problem of flash crowd. When the problem occurs, a large number of users suffer from large latency in joining a channel and even get blocked. In the past, rare studies are focused on the flash crowd issue. In this thesis, we identify several key factors that may cause the flash crowd problem: request congestion, control message, and maintenance overhead. Based on our analysis, we present a batch join scheme instead of processing new users one by one. Our batch join process also solves the problem of the current join processes that only use existing active peers to serve new users. We generate a virtual sub tee based on both current active peers and a batch of new peers. Simulation results demonstrate that our batch join scheme significantly mitigate the flash crowd situation. Moreover, both of the peer blocking rate and re-join times are significant reduced.

並列關鍵字

contribution ； location aware ； bandwidth ； PPLive ； live streaming ； p2p

參考文獻

[23] Tein-Yaw Chung, Yang-Hui Chang, Kun-Hung Chen, and Yung-Mu Chen, "High Performance Overlay File Distribution With Resource Discovery and Service Scheduling," Journal of Information Science and Engineering. Vol.25, No.3, May, 2009, pp.861-875.

[4] X. Zhang, J. Liu, B. Li and T. O. Yum, “Cool-streaming/donet: A data-driven overlay network for efficient live media streaming,” Proc. IEEE INFOCOM, vol. 3, pp. 2102–2111, March, 2005.

[5] S. Banerjee, B. Bhattacharjee, and C. Kommareddy, “Scalable application layer multicast,” Proc. of ACM SIGCOMM, pp. 205-217, August, 2002.

[8] V. N. Padmanabhan, H. J. Wang, P. A. Chou, and K. Sripanid-kulchai, “Distributed streaming media content using cooperative networking,” Proc. ACM NOSSDAV, pp. 177–186, May 2002.

[10] N. Magharei, R. Rejaie, and Y. Guo, “Mesh or Multiple-Tree: A Comparative Study of Live P2P Streaming Approaches,” Proceedings of IEEE INFOCOM, pp. 1424-1432, May 2007.