臺灣博碩士論文加值系統

目前許多普及的網際網路服務使用傳統的用戶伺服器(client-server)模式，其中包含，電子郵件(E-mail)、檔案傳輸協定(File Transfer Protocol)、隨選視訊(Video on Demand)、網路電視(Internet Protocol Television)及全球資訊網(World Wild Web)。然而，當用戶量劇增時，用戶伺服器模式將面臨可容納度及高成本問題。在所有網際網路服務中，網路電視需要佔用最多的資源，例如，儲存空間、頻寬，而且可能容易遇到大量使用者同時使用的瓶頸。為了降低網路電視系統資源的需求，越來越多網路電視服務已利用同儕式(Peer-to-Peer)架構。不過同儕式網路電視系統在切換頻道或節目時不夠靈活，使用者必需花一段時間等待節目的出現，比起傳統電視，同儕式網路電視的人機介面相對地緩慢。而初始遲滯(Start-up delay)是影嚮使用者收看同儕式網路電視意願的關鍵因素。一般而言，除了免費節目，使用者會抱怨頻道切換時間或起始放映遲滯太長。事實也確是如此，目前同儕式網路電視系統的初始遲滯從數十秒到數分鐘不等，因此初始遲滯有很大的改善空間。
在這篇論文中，我們提出一個嶄新的方法用於降低同儕式網路電視初始遲滯。我們讓系統中每個同儕(peer)記錄伙伴(partner)的傳輸率，週期性地推荐傳輸率符合某些條件的伙伴給追縱者伺服器(tracker server)，並視這些被推荐的伙伴為超級同儕(super-peer)。追縱者伺服器將超級同儕及鄰居同儕(neighbor peer)分配給新進同儕。新進同儕能從超級或鄰居同儕快速地接收影像內容而降低初始遲滯。
我們發展一套模擬軟體並和網路模擬器相互合作用以評估我們所提出的方法。數值結果顯示我們所提出的方法能有效地降低初始遲滯。

Many popular services currently deployed on Internet are based on the traditional client-server model, these include E-mail, FTP, Video on Demand (VoD), Internet Protocol Television (IPTV) and World Wild Web (WWW). However, client-server model faces the scalability and high cost problems when the number of client grows up. Among these services IPTV needs the largest amount of resources such as storage and bandwidth, and it may easily encounter the bottleneck with only small number of users. To reduce IPTV system resource requirement, more and more IPTV services are deployed based on Peer-to-Peer (P2P) architecture. Unfortunately, P2P IPTV system always starts slowly whenever you switch to a new channel or program, and it seems awkward comparing with traditional TV if users have to spend quite a while waiting for the program to show up in P2P IPTV.
Start-up delay is a key factor which affects user’s willing regarding whether to subscribe P2P IPTV. In general, unless the program is free, users may complain if the content switching time or the initial video playback delay is too long. Unfortunately, start-up delay in current P2P IPTV systems range from tens of seconds to few minutes, it still has a lot of room to improve.
In this thesis, we propose a novel scheme to reduce P2P IPTV start-up delay. We make every peer in the system to record its partners’ sending rate and periodically recommend thosae partners whose sending rate satisfies some criteria to tracker server as “super-peers”. Tracker server dispatches both super and neighbor peers to the newly joined peer, which then can rapidly receive video content from either super or neighbor peers and reduce start-up delay.
We develop a simulation program to cooperate with network simulator to evaluate our proposed scheme. The numerical result shows that our proposed scheme is able to reduce the start-up delay significantly.

摘要 I
Abstract II
Acknowledgement IV
Contents V
Table List VIII
Figure List IX
Chapter 1 Introduction 1
1.1 Overview 1
1.2 Motivation and Purpose 3
1.3 Organization of the Thesis 3
Chapter 2 Background 5
2.1 Peer-to-Peer Network 5
2.1.1 Introduction 5
2.1.2 P2P on file-sharing 6
2.1.3 BT revolution 7
2.2 Video Streaming Architecture 9
2.2.1 Centralized server architecture 9
2.2.2 IP layer multicast 10
2.2.3 P2P architecture 11
2.3 Tree-Based Overlay 12
2.3.1 Single tree overlay 12
2.3.2 Multiple tree overlay 13
2.4 Mesh-Based Overlay 14
2.4.1 Mesh-based system architecture 16
2.4.2 Peer adaptation 17
2.5 Mesh-Based Software Components 17
2.5.1 P2P streaming engine 18
2.5.2 Media player 19
2.6 Long Start-up Delay Issue 19
2.7 Related Works 20
Chapter 3 Proposed Approaches 23
3.1 Design Philosophy 23
3.2 Main Scheme 24
3.2.1 Super-peer recommendation mechanism 25
3.2.2 Power start-up mechanism 28
3.2.3 Scheme related issues 30
3.3 Software Viewpoint 31
3.3.1 Tracker server algorithm 31
3.3.2 Peer application algorithm 32
3.4 Summary 33
Chapter 4 Simulation and Numerical Result 34
4.1 Simulation Design 34
4.1.1 Experiment on super-peer recommendation mechanism 35
4.1.2 Experiment on start-up delay 36
4.2 Simulation Program 37
4.3 Numerical Results 39
4.2.1 Super-peer ratio 39
4.2.2 Start-up delay 41
4.4 Summary 43
Chapter 5 Conclusion and Future Works 44
References 46

[1] BitTorrent,http://www.bittorrent.com/.
[2] PPLive,http://www.pplive.com/.
[3] PPStream,http://www.pps.tv/.
[4] Python,http://www.python.org/.
[5] Skype,http://www.skype.com/.
[6] SopCast,http://www.sopcast.com/.
[7] TVants,http://www.tvants.com/.
[8] Banerjee, S., B. Bhattacharjee, et al."Scalable application layer multicast." Computer Communication Review 32(4): 205-217, 2002.
[9] Chu, Y. H., S. G. Rao, et al."A case for end system multicast." Ieee Journal on Selected Areas in Communications 20(8): 1456-1471, 2002.
[10] Deering, S."Multicast Routing in Internetworks and Extended LANs." ACM SIGCOMM, 1988.
[11] Hei, X., C. Liang, et al."A measurement study of a large-scale P2PIPTV system." Ieee Transactions on Multimedia 9(8): 1672-1687, 2007.
[12] Hei, X. J., Y. Liu, et al."IPTV over P2P streaming networks: The mesh-pull approach." Ieee Communications Magazine 46(2): 86-92, 2008.
[13] Jiang, X., Y. Dong, et al."GnuStream: a P2P media streaming system prototype." International Conference on Multimedia and Expo 2: 325-328, 2003.
[14] Li, B. and H. Yin"Peer-to-peer live video streaming on the internet: Issues, existing approaches, and challenges." Ieee Communications Magazine 45(6): 94-99, 2007.
[15] Liu, J. C., S. G. Rao, et al."Opportunities and challenges of peer-to-peer and Internet video broadcast." Proceedings of the Ieee 96(1): 11-24, 2008.
[16] Sentinelli, A., G. Marfia, et al."Will IPTV ride the peer-to-peer stream?" Ieee Communications Magazine 45(6): 86-92, 2007.
[17] Tran, D. A., K. A. Hua, et al."A peer-to-peer architecture for media streaming." Ieee Journal on Selected Areas in Communications 22(1): 121-133, 2004.
[18] Xie, S., B. Li, et al."Coolstreaming: Design, theory, and practice." Ieee Transactions on Multimedia 9(8): 1661-1671, 2007.
[19] Zhang, X., J. Liu, et al."DONet/CoolStreaming: A data-driven overlay network for peer-to-peer live media streaming." Proc. IEEE INFOCOM 3: 2102 - 2111, 2005.