P2P(Peer-to-peer對等網際網路技術)近年來已經廣泛的被使用在即時影像串流分享、通訊應用與檔案分享上，也因為全球各地網路設備的進步，因此關注各式重大比賽與遊戲實況的人越來越多了，人們為了提升轉播即時比賽與遊戲實況的效能，P2P也逐漸的成為在網際網路傳輸上的一個重要技術。
    在本論文中，將透過賽局理論的觀點基於社群關係，提出一個新式的分享機制Novel Bilateral Incentive Mechanism (NBIM)，激勵使用者自發性的貢獻自身擁有的資源，同時也能有效抑制惡意節點，由於點對點傳輸軟體上的資源猶如公有資源，所以人人都可以使用，以至於大多數的人都不太願意分享自身資源，另一方面，由於點對點傳輸軟體上大多採取帳號制或匿名制的方式來確認身份，所有人都可以分享檔案資源，因此，造成了有心人士藉此分享不明的檔案來欺騙其他使用者以獲取大量的系統資源，因此本論文將基於社群關係設計一個新式的P2P分享機制，促使使用者透過長期的合作，促進節點分享，也同時透過社群關係中節點間的關係抑 free-riders 與惡意節點，最後透過賽局理論來分析在本機制中不同情況下的使用者的策略選擇與其相對應的報酬所得，期望本機制可以有效的抑制P2P系統中的 free-riders 與惡意節點，且達到更公平的分配網路資源給所有的使用者的目的。
    透過模擬結果與數據分析，我們證明了NBIM確實讓本系統可以有效的促進使用者貢獻自身的資源，抑制本系統中的free-riders與malicious nodes可用頻寬以提升系統效能，NBIM 平均抑制了free-riders 擁有可用頻寬的31% 與 malicious node擁有可用頻寬的41%，節點在P2P分享網路中可用頻寬的減少將導致其行為能力受限，不僅解決了 BT 與 Gnutella 這類 P2P 網路中因 free-riders 而導致的效能低落的問題也處理掉系統的散佈惡意資源的malicious nodes 。最後由於本系統分配資源與激勵機制的結合，增加了 friendly nodes 更能藉此分享自身的可以分享的最大頻寬來提升系統效能，讓其他節點可以在本系統中享有最好的系統效能與安全分享的環境。

P2P file-sharing system has been developed rapidly over the past years. P2P file-sharing mechanisms such as Gnutella, BitTorrent and Private Tracker have been used popularly. However the free-rider and malicious node problem hinders the efficient utilization of P2P networks. Hence this paper makes improvements on the original BitTorrent by proposing a novel bilateral incentive mechanism (NBIM) to restrain free riders and malicious nodes simultaneously. 
	The impact of proposed mechanism on free-riders and malicious nodes is analyzed using Peersim. The simulated results prove that the proposed mechanism can restrain average 31 percent behavioral capability of free riders and average 41 percent behavioral capability of malicious nodes to improve the performance of the P2P network.

目錄 
致謝	I 
摘要	II 
Abstract	III 
目錄	IV 
圖目錄	V 
表目錄	VI
CHAPTER 1  INTRODUCTION	1
1.1 Background	1
1.2 Objectives	2
1.3 Thesis Organization	3
CHAPTER 2  RELATED WORK	4
2.1 Overview of P2P Architectures	4
2.1.1 Centralized Architectures	4
2.1.2 Decentralized Architecture	6
2.1.3 Hybrid Decentralized Architectures	8
2.2 Social Network	10
2.3 Evolutionary Game Theory	11
2.4 Previous Study	11
CHAPTER 3  METHODOLOGY	13
3.1 Social Relation	15
3.2 Counter Mechanism	15
3.3 Reward and Punishment Mechanism	19
3.3.1 Bandwidth Contribution Mechanism	20
3.3.2 Evaluation System	22
3.4 Incentive Mechanism Analysis	24
3.4.1 Evolutionary Game Theory	24
3.4.2 System Fairness and Recidivist Punishment	26
CHAPTER 4 SIMULATION RESULTS AND PERFORMANCE ANALYSIS	28
4.1 Simulation Environment	28
4.2 P2P Simulator	30
4.2.1 PeerSim	30
4.3 Simulation Results and Performance Analysis	31
4.3.1 Simulation Analysis of Single Behavior	32
4.3.2 Simulation Analysis of Multiple-Behavior	44
CHAPTER 5 CONCLUSION AND FUTURE	45
REFERENCE	46

圖目錄
Figure 2-1: Napster Architecture	5
Figure 2-3: Gnutella architecture	6
Figure 2-4: BitTorrent Architecture	8
Figure. 3-1. System architecture of NBIM	14
Figure. 3-2. System flow diagram of Bandwidth Contribution mechanism in terms of User	21
Figure. 3-3. System flow diagram of Evaluation system in terms of User	23
Figure. 3-4. System flow diagram of NBIM	23
Figure. 3-5. Curve of arctangent function	26
Figure. 3-6. Comparison graph	27
Figure. 4-1. The node classified result of condition 1	32
Figure. 4-2. The node classified result of condition 3	34
Figure. 4-3. The node classified result of condition 3	36
Figure. 4-4. The node classified result of condition 4	38
Figure. 4-5. The node classified result of condition 5	41
Figure. 4-6. The node classified result of condition 6	44

表目錄

Table 3-1: Definition of node type	19
Table 3-2: Payoff table of the example	25
Table 4-1: The first simulation scenario	29
Table 4-2: The second simulation scenario	29
Table 4-3: Simulation parameters	30
Table 4-4: Average bandwidth in condition 1 comparison table	33
Table 4-5: Average bandwidth in condition 2 comparison table	35
Table 4-6: Average bandwidth in condition 3 comparison table	37
Table 4-7: Average bandwidth in condition 4 comparison table	39
Table 4-8: Average bandwidth in condition 5 comparison table	42
Table 4-9: Average bandwidth of multiple behavior comparison table	45

[1] COHEN, Bram. Incentives build robustness in BitTorrent. In: Workshop on Economics of Peer-to-Peer systems. 2003. p. 68-72
[2] RIPEANU, Matei. Peer-to-peer architecture case study: Gnutella network. In:Peer-to-Peer Computing, 2001. Proceedings. First International Conference on. IEEE, 2001. p. 99-100.
[3] Peer-to-Peer Technologies and Protocols, http://ntrg.cs.tcd.ie/undergrad/4ba2.02/p2p/.
[4] STEINER, Moritz; EN-NAJJARY, Taoufik; BIERSACK, Ernst W. A global view of kad. In: Proceedings of the 7th ACM SIGCOMM conference on Internet measurement. ACM, 2007. p. 117-122.
[5] FAN, Bin; LUI, John; CHIU, Dah-Ming. The design trade-offs of BitTorrent-like file sharing protocols. IEEE/ACM Transactions on Networking (TON), 2009, 17.2: 365-376.
[6] ADAR, Eytan; HUBERMAN, Bernardo A. Free riding on Gnutella. First monday, 2000, 5.10.
[7] STOICA, Ion, et al. Chord: A scalable peer-to-peer lookup service for internet applications. ACM SIGCOMM Computer Communication Review, 2001, 31.4: 149-160.
[8] ROWSTRON, Antony; DRUSCHEL, Peter. Pastry: Scalable, decentralized object location, and routing for large-scale peer-to-peer systems. In:Middleware 2001. Springer Berlin Heidelberg, 2001. p. 329-350.
[9] RATNASAMY, Sylvia, et al. A scalable content-addressable network. ACM, 2001.
[10] HO, Ai; MAIGA, Abdou; AÏMEUR, Esma. Privacy protection issues in social networking sites. In: Computer Systems and Applications, 2009. AICCSA 2009. IEEE/ACS International Conference on. IEEE, 2009. p. 271-278.
[11] WEIBULL, Jörgen W. Evolutionary game theory. MIT press, 1997.
[12] WU, Tin-Yu, et al. Incentive mechanism for P2P file sharing based on social network and game theory. Journal of Network and Computer Applications, 2014, 41: 47-55.
[13] ZHANG, Yunchang; CHEN, Shanshan; YANG, Geng. SFTrust: A double trust metric based trust model in unstructured P2P system. In: Parallel & Distributed Processing, 2009. IPDPS 2009. IEEE International Symposium on. IEEE, 2009. p. 1-7.
[14] ANDERSEN, Reid, et al. Trust-based recommendation systems: an axiomatic approach. In: Proceedings of the 17th international conference on World Wide Web. ACM, 2008. p. 199-208.
[15] QI, Jian-Jun; LI, Zeng-Zhi. Managing trust for secure active networks. In: Multi-Agent Systems and Applications IV. Springer Berlin Heidelberg, 2005. p. 628-631.
[16] MANI, Mehdi; NGYUEN, Anh-Minh; CRESPI, Noel. What's up: P2P spontaneous social networking. In: Pervasive Computing and Communications, 2009. PerCom 2009. IEEE International Conference on. IEEE, 2009. p. 1-2.
[17] CENTENO, Roberto; BILLHARDT, Holger; HERMOSO, RamóN. Persuading agents to act in the right way: An incentive-based approach. Engineering Applications of Artificial Intelligence, 2013, 26.1: 198-210.
[18] ISLAM, Khondkar R. A cost-effective distributed architecture for content delivery and exchange over emerging wireless technologies. 2013. PhD Thesis. George Mason University.
[19] SALAH, Hisham; ELTOWEISSY, Mohamed. Towards a personalized trust management system. In: Innovations in Information Technology (IIT), 2012 International Conference on. IEEE, 2012. p. 373-378.
[20] ZHU, Yanlin; YANG, Yuliang. Reseach on LT-based communication between peers in BT system. In: Consumer Electronics, Communications and Networks (CECNet), 2012 2nd International Conference on. IEEE, 2012. p. 1980-1983.
[21] QIU, Dongyu; SRIKANT, Rayadurgam. Modeling and performance analysis of BitTorrent-like peer-to-peer networks. In: ACM SIGCOMM computer communication review. ACM, 2004. p. 367-378.
[22] QIN, Fenglin; LIU, Ju; ZHENG, Lina. Improving robustness and fairness on BitTorrent-Like P2P systems. In: Communications and Networking in China, 2009. ChinaCOM 2009. Fourth International Conference on. IEEE, 2009. p. 1-5.