臺灣博碩士論文加值系統

　　本論文中，我們提出一個無線TCP連線下以量測遺失封包為基礎的適應性視窗控制法，稱為TCP-WO（waiting and observing）。TCP-WO是對於無線TCP連線下一個既簡單又有效的視窗（congestion window）調整的壅塞控制方法，能有效提升吞吐量。TCP-WO的基本概念是利用監測網路封包的遺失評估網路狀態進而調適視窗的大小。為了實現這個概念，當首次逾時（timeout）或是重複確認（duplicated ACKs）的事件發生後，TCP-WO 開始監測逾時與重複確認發生次數，監測區間的時間長度為一個RTT（round trip time）。於監測區間中TCP-WO統計逾時與重複確認次數。根據封包的遺失測量，來評估網路連線的狀態；並修改乘式遞減（multiplicative decrease）演算法來調整視窗，改善原來TCP-Reno以固定比例調整視窗的方法。

　　我們將描述所使用的網路模型、TCP-WO的設計、數學模型分析以及如何利用NS-2模擬TCP-WO的動態來評估TCP-WO在無線網路環境下之吞吐量、對可用頻寬的適應程度與穩定度、公平性與友善性的效能。模擬結果顯示於具有壅塞的網路與當無線通道不佳時，例如有10％的位元錯誤率（bit error rate）的情況下，TCP-WO相對於TCP-Westwood、TCP-Jersey、TCP-Veno與TCP-Reno的吞吐量分別改善了103％、25％、22%與96％。同樣於多個TCP 連線下，模擬結果同樣證明對於TCP-Reno來說，TCP-WO同樣能保持公平性與友善性。

　In this thesis, we propose a novel approach to adaptively control congestion windows of wireless TCP connections based on loss measurement, called TCP-WO (waiting and observing). TCP-WO is a simple and effective scheme to adjust the congestion window (cwnd) of a wireless TCP connection and to improve the throughput performance. The key idea of TCP-WO is a window adaptation based on the measurement of packet loss. To realize the idea, TCP-WO monitors the timeout and duplicated ACKs in an observing duration, and the length of observing duration is an RTT (round trip time) when the first timeout or duplicated ACK is encountered. In this duration TCP-WO counts the number of timeouts and duplicated ACKs. According to the measurements, TCP-WO, a variant of TCP-Reno, adjusts the multiplicative decrease algorithm based on the estimation of network status, rather than a constant decrement.

　In this thesis, network modeling, the operations of TCP-WO, mathematical analysis and simulation results derived from a NS-2-based network environment are presented. Simulation results show that in a congestion network with 10 % bit error rate (BER) of Bad state for wireless channel, TCP-WO achieves relative throughput improvements of 103%, 25%, 22%, and 96% in comparison with TCP-Westwood, TCP-Jersey, TCP-Veno, and TCP-Reno, respectively. In the case of multiple TCP flows, simulation results also demonstrate that TCP-WO maintains a fair and friendly behavior with respect to TCP-Reno.

Abstract II
Figure List VII
Table List VIII
Chapter 1. Introduction 1
Chapter 2. Related Works of Wireless TCP 4
2.1. Congestion Control in a Wireless Connecti 5
2.2. Congestion Windowing Methods 6
2.3. Summary 11
Chapter 3. TCP-WO: An Adaptive Congestion Windowing Approach to Wireless TCP Based on Loss Measurement 13
3.1. Connection Quality Estimation of End-to-End Connection on Wired and Wireless Network 14
3.2. Congestion Control Based on Loss Measurement 20
3.2.1. Slow-Start Algorithm 20
3.2.2. Additive Increase Algorithm 20
3.2.3. Multiplicative Decrease Algorithm 21
3.2.4. Modifications in TCP-WO for Loss Measurement 22
3.3. Adjustment of Congestion Window in TCP-WO 23
3.4. Adjustment of Slow-Start Threshold in TCP-WO 24
3.5. Operation of TCP-WO 28
3.5.1. Overview 29
3.5.2. First Step: Monitor 31
3.5.3. Second Step: Determination of Observing Duration 33
3.5.4. Third Step: Starting Observing Duration 34
3.5.5. Fourth Step: Observing Duration 35
3.5.6. Fifth Step: Adjustment 36
Chapter 4. Throughput Evaluation of TCP-WO 39
4.1. Simulation Environment 40
4.2. Independent Packet Loss 41
4.3. Burst Error 44
4.3.1. Burst Error Model 45
4.3.2. Fading 46
4.3.3. Blackout 54
Chapter 5. Response, Fairness, and Friendliness Evaluation of TCP-WO 58
5.1. Simulation Environment 58
5.2. Response of Bandwidth Adaptation 59
5.3. Fairness of TCP-WO 63
5.4. Friendliness of TCP-WO 64
Chapter 6. Conclusion and Future Works 70
6.1. Conclusion 70
6.2. Future Works 71
References 72

[1] V. Jacobson, ”Congestion avoidance and control,” ACM Comput. Commun. Rev., vol. 18, pp. 314–329, Aug. 1988.
[2] V. Jacobson, Modified TCP Congestion Avoidance Algorithm, Apr. 30, 1990. end2end-interest mailing list.
[3] J. Hoe, “Improving the Start-Up Behavior of a Congestion Control Scheme for TCP”, in Proc. of SIGCOMM ’96 Symposium on Communications Architectures and Protocols, Aug. 1996.
[4] K. J. Astriiom and B. Wittenmark, Computer controlled systems, Prentice Hall, Englewood Cliffs, NJ, 1997.
[5] A. A. Abouzeid, S. Roy and M. Azizoglu, “Stochastic modeling of TCP over lossy link”, in Proc. of INFOCOM’2000, vol. 3, Tel Aviv, Mar. 2000, pp. 1724-1733.
[6] S. Mascolo, M.Y. Sanadidi, C. Casetti, M. Gerla, and R. Wang, ”TCP Westwood: End-to-End Congestion Control for Wired/Wireless Networks,” Wireless Networks J., vol. 8, pp. 467-479, 2002.
[7] C. P. Fu, and S. C. Liew, ”TCP Veno: TCP Enhancement for Transmission Over Wireless Access Networks, ” IEEE J. on Select. Areas in Commun., vol. 21,no. 2, Feb. 2003.
[8] A. Capone,L Frattaand, and F. Martignon, “Bandwidth Estimation Schemes for TCP over Wireless Networks,” IEEE Trans. on Mobile Computing, vol. 3, no. 2, Apr.-Jun. 2004.
[9] K. Xu, Y. Tian, and N. Ansari, “TCP-Jersey for Wireless IP Communications,“ IEEE J. on Select. Areas in Commun., vol. 22, no. 4, May 2004.
[10] A. Bakre and B. R. Badrinath, “I-TCP: Indirect TCP for mobile hosts,” in Proc. ICDCS 95, May 1995, pp. 136–143.
[11] K. Brown and S. Singh, “M-TCP: TCP for mobile cellular networks,” ACM Comput. Commun. Rev., vol. 27, no. 5, pp. 19–43, 1997.
[12] K. Wang and S. K. Tripathi, “Mobile-end transport protocol: An alternative to TCP/IP over wireless links,” in proc. of IEEE INFOCOM’98, vol. 3, San Francisco, Mar.1998, pp. 1046-1053.
[13] H. Balakrishnan, S. Seshan, and R. H. Katz, “Improving reliable transport and handoff performance in cellular wireless networks,” ACM/Baltzer Wireless Networks J., vol. 1, no. 4, pp. 469–481, Dec.1995.
[14] S. Keshav and S. Morgan, “SMART retransmission: Performance with overload and random losses,” in Proc. of IEEE INFOCOM’97, vol. 3, Kobe, Apr. 1997, pp. 1131-1138.
[15] K. Ratnam and I. Matta, “WTCP: An efficient mechanism for improving TCP performance over wireless links,” in Proc. of Int Symp. Comput. Commun., 1998, pp. 74–78.
[16] H. Balakrishnan and R. H. Katz, “Explicit loss notification and wireless web performance,” in Proc. of IEEE GLOBECOM Internet Mini-Conf.,Sydney, Australia, Nov. 1998.
[17] S. Floyd, “TCP and explicit congestion notification,” ACM Comput. Commu. Rev., Vol. 24, pp. 10-23, Oct. 1994.
[18] L. S. Brakmo, S. W. O’Malley, and L. L. Peterson, “TCP Vegas: New techniques for congestion detection and avoidance,” in Proc. SIGCOMM’94, Oct. 1994, pp. 24–35.
[19] J. S. Ahn, P. B. Danzig, Z. Liu, and E. Yan, “Evaluation with TCP Vegas: Emulation and Experiment,” ACM Comput. Commun. Rev., vol. 25, no. 4, pp. 185–195, Aug. 1995.
[20] J. Mo, R. J. La, V. Anantharam, and J.JeanWalrand, “Analysis and Comparison of TCP Reno and Vegas,” in Proc. of INFOCOM’99, vol. 3, New York, Mar.1999, pp. 1556-1563.
[21] U. Hengartner, J. Bolliger, and T. Gross, “TCP Vegas Revisited,” in Proc. of INFOCOM’2000, vol. 3, Tel Aviv, Mar. 2000, pp. 1546-1555.
[22] C. P. Fu, L. -C. Chung, and S. C. Liew, “Performance Degradation of TCP Vegas in Asymmetric Networks and its Remedies,“ in Proc. ICC’2001, vol. 10, 2001, PP. 3229-3236.
[23] D. Chiu and R. Jain, “Analysis of the increase/decrease algorithms for congestion avoidance in computer networks,” J. Comput. Networks, vol.17, no. 1, pp. 1–14, Jun. 1989.
[24] R. Jain, D. Chiu, and W. Hawe, “A quantitative measure of fairness and discrimination for resource allocation in shared computer systems,” DEC, Res. Rep.TR-301, 1984.
[25] NS-2 network simulator, Version 2, LBL, http://www-mash.cs.berkeley.edu/ns.
[26] REAL 5.0 Overview, http://www.cs.cornell.edu/skeshav/real/overview.html.
[27] TCP Westwood modules for NS-2, http://www.telematics.polito.it/casetti/tcp-westwood.
[28] TCP Jersey modules for NS-2, http://web.njit.edu/anl/download.html.