傳統上TCP的流量控制是window based，即利用滑動視窗(sliding window)來做流量控制，然而window based方法是藉著偵測網路上封包遺失、壅塞發生及傳輸逾時來被動地調整TCP之傳送速度。在網路速度日益增加的今日，被動式調速對長距離大RTT(Round Trip Time)且具高頻寬的長寬網路(Long Fat Networks, LFNs)卻顯得反應時間太長，TCP因而難以迅速反應達到應有的頻寬。
另一方面，随著目前無線網路的日益成長，原本在有線網路上所提供之服務與應用也漸移植至無線環境。受限於無線環境之雜訊及干擾，封包遺失在無線環境中屬於常有現象；另外在無線網路中跨基地台的交遞(handoff)也會造成封包遺失及傳輸逾時。因此在無線網路上傳統TCP藉著封包遺失及傳輸逾時來調整傳送速度的策略則顯得無法獲得目前無線網路應有頻寬。
我們提出一個新的傳輸控制協定：Active Rate Anchoring TCP (ARCH-TCP)，藉由packet-pair量測以及補償機制使ARCH-TCP得以迅速且精確地測得目前網路之可用頻寬。我們使用J-Sim來模擬並驗證我們的想法。此傳輸控制協定同時可用於有線及無線網路之中：在有線網路中，可以解決LFNs所帶來反應時間過長的問題、測得可用頻寬供ARCH-TCP充份運用並且不影響一般TCP之運作；在無線網路中，則可減輕網路在封包遺失及handoff時對流量降低產生的影響。

Conventional TCP is window based, which exploits the sliding window mechanism to conduct the flow control. It increases the sending window additively and decreases the sending window multiplicatively in response to successful transmission and, packet loss/timeout events respectively. While the mechanism works quite well in normal networks, TCP can hardly reach the ideal bandwidth utilization in long fat networks (LFNs) due to long delay and bursts of packet losses. Besides, as wireless and mobile computing has become popular today, packet loss in such networks may occur due to noise, interference and handoff across different domains. TCP could not react to different situations effectively as it sees all packet losses as an indication of network congestion.
In this thesis, we proposed a new transmission control mechanism called Active Rate Anchoring TCP (ARCH-TCP). In ARCH-TCP, we explicitly integrate bandwidth measurement into TCP to solve the aforementioned problem. Specifically, we exploit packet-pair measurement to quickly estimate bandwidth share and then RTT variation is observed to compensate measurement error. We built the model in J-Sim network simulator to evaluate the effectiveness of our proposal. We found that ARCH-TCP can react to network conditions quickly and precisely in both wired and wireless networks and both in the normal networks and LFNs.

Chapter 1 Introduction 1
Chapter 2 Related Works 3
2.1 Basic Concept of TCP 3
2.1.1 End to End 3
2.1.2 Connection-Oriented 3
2.1.3 Reliability 3
2.1.4 Slow Start 5
2.1.5 Congestion Avoidance 5
2.1.6 Fast Retransmit 6
2.1.7 Fast Recovery 6
2.2 TCP and Its Variants for High-Speed Network Environments 6
2.2.1 TCP Reno 6
2.2.2 TCP Vegas 7
2.2.3 TCP Westwood and Westwood+ 7
2.2.4 HighSpeed TCP 8
2.2.5 FAST TCP 8
2.3 Simulation Scenarios 9
2.3.1 TCP Simulation 1: Reno’s Response in Short RTT 9
2.3.2 TCP Simulation 2: Reno’s Response in Long RTT 11
2.3.3 TCP Simulation 3: Modified Reno’s Response in Long RTT 12
2.3.4 TCP Simulation 4: Vegas’ Response in Long RTT 13
2.3.5 TCP Simulation 5: Reno’s and Vegas’ Responses in Wireless Handoff 14

Chapter 3 Design Strategies and Implementation of ARCH-TCP 16
3.1 Integrating Packet-Pair Measurement 17
3.1.1 Bottleneck Effect of Packet-Pair in FIFO Queueing Network 17
3.1.2 Congestion Window Adjustment in Sender 18
3.1.3 Maxburst Mechanism 18
3.1.4 Basic Operation of Packet-Pair Measurement 18
3.2 Fairness and Mismatch of Packet-Pair Measurement 19
3.3 Compensation and Counterpoises 21
3.3.1 Cumulative Rate Average 21
3.3.2 Minimum RTT Difference Compensation 22
3.3.3 Average RTT Compensation 23
3.3.4 Running Average on Measured Bandwidth and RTT 23
3.4 ARCH-TCP Operations 24
3.4.1 Slow Start Phase 24
3.4.2 ARCH Normal Phase 24
3.4.3 Packet Loss and Timeout 25
3.5 Benefit and Advantage of ARCH-TCP 25

Chapter 4 Simulation Results and Analysis 26
4.1 Simulation Environment Setup 26
4.2 Simulation Results 27
4.2.1 ARCH’s response in long RTT 27
4.2.2 ARCH’s responses in wireless handoff 29
4.2.3 Coexistence of ARCH-TCP and TCP Reno 31
4.2.4 Coexistence of Multiple ARCH Flows in Long RTT 33
4.2.5 Coexistence of Multiple ARCH Flows in Different RTTs 36

Chapter 5 Conclusions and Future Work 39
5.1 Conclusions 39
5.2 Future Work 40

References 41

[1] Sally Floyd and Kevin Fall, “Promoting the Use of End-to-End Congestion Control in the Internet,” IEEE/ACM Transactions on Networking, Vol. 7, Issue 4, pp. 458–472, Aug. 1999.
[2] D. M. Chiu and R. Jain, “Analysis of the Increase and Decrease Algorithms for Congestion Avoidance in Computer Networks,” Computer Networks and ISDN Systems, Vol. 17, Issue 1, pp. 1–14, 1989.
[3] M. Allman, V. Paxson, and W. Stevens, “TCP Congestion Control,” RFC 2581, April 1999.
[4] C. Zhang and V. Tsaoussidis, “TCP Smoothness and Window Adjustment Strategy,” IEEE Transactions on Multimedia, Vol. 8, no. 3, pp. 600–609. June 2006.
[5] V. Jacobson, R. Braden, and D. Borman, “TCP Extensions for High Performance,” RFC 1323, May 1992.
[6] Jay Aikat and et al., “Variability in TCP Round-trip Times,” in Proc. of the 2003 ACM SIGCOMM conference on Internet measurement. 2003, pp. 279–284.
[7] D. Katabi, M. Handley, and C. Rohrs, “Congestion Control for High Bandwidth- Delay Product Networks,” in Proc. ACM SIGCOMM, Pittsburgh, PA, Aug. 2002.
[8] B. S. Bakshi and et al., “Improving Performance of TCP over Wireless Networks,” in International Conference on Distributed Computing Systems. May 1997, pp. 365–373.
[9] D. S. Eom and et al., “Improving TCP Handoff Performance in Mobile IP Based Networks,” Computer Communications, Vol. 25, Issue 7, pp. 635–646, April 2002.
[10] Y. R. Yang and et al., “Two Problems of TCP AIMD Congestion Control,” in Proc. of ICNP 2000.
[11] V. Jacobson, “Congestion Avoidance and Control,” in Proc. ACM SIGCOMM '88, Aug. 1988, pp. 314-329.
[12] W. Stevens, “TCP Slow Start, Congestion Avoidance, Fast Retransmit, and Fast Recovery Algorithms,” RFC 2001, Jan. 1997.
[13] L. S. Brakmo and L. L. Peterson, “TCP Vegas: End to End Congestion Avoidance on a Global Internet,” IEEE Journal on selected areas in communications, Vol. 13, Issue 8, pp.1465–1480, Oct. 1995.
[14] G. Hasegawa, K. Kurata, and M. Murata, “Analysis and Improvement of Fairness between TCP Reno and Vegas for Deployment of TCP Vegas to the Internet,” in Proc. 2000 International Conference on Network Protocols, Nov. 2000, pp. 177–186.
[15] M. Gerla and et al., “TCP Westwood: Congestion Window Control Using Bandwidth Estimation,” in Proc. of IEEE Globecom, Vol. 3, pp. 1698–1702, Nov. 2001.
[16] S. Mascolo and et al., “Performance Evaluation of Westwood+ TCP Congestion Control,” Performance Evaluation, Vol. 55, Issue 1–2, pp. 93–111, Jan. 2004.
[17] S. Floyd, “HighSpeed TCP for Large Congestion Windows,” RFC3649, Dec. 2003.
[18] D. Phillips and J. Hu, “Analytic Models for Highspeed TCP Fairness Analysis,” in ICON, Vol. 2, pp. 720–730, Nov. 2004.
[19] T. Hatano, H. Shigeno, and K. Okada, “TCP-friendly Congestion Control for Highspeed Network,” in SAINT'07, Jan. 2007.
[20] D. X. Wei and et al., “FAST TCP: Motivation, Architecture, Algorithms, Performance,” IEEE/ACM Transactions on Networking, Vol. 14, Issue 6, pp. 1246–1259, Dec. 2006.
[21] F. Paganini and et al., “A New TCP/AQM for Stable Operation in Fast Networks,” in Proc. IEEE INFOCOM, Vol. 1, pp. 96–105, 2003.
[22] C. Jin and et al., “Method and Apparatus for Network Congestion Control Using Queue Control and One-Way Delay Measurements,” United States Patent 20060050640.
[23] The J-Sim Project, http://www.j-sim.org
[24] Srinivasan Keshav, “A Control-Theoretic Approach to Flow Control,” in Proc. ACM SIGCOMM'91, pp.3–15, Sep. 1991.
[25] C. Dovrolis, P. Ramanathan, and D. Moore, “What Do Packet Dispersion Techniques Measure?” in Proc. of IEEE INFOCOM, Vol. 2, pp. 905–914, 2001.
[26] Hsiu-Hung Lai, “A Study of Rate-based TCP Mechanisms,” Master thesis, Department of Electrical Engineering, National Sun Yat-Sen University, 2006.
[27] J. Aweya and et al., “Enhancing Network Performance with TCP Rate Control,” in Proc. IEEE Globecom, Vol. 3, pp. 1712–1718, 2000.
[28] J. Semke, J. Mahdavi, and M. Mathis, “Automatic TCP Buffer Tuning,” in Proc. ACM SIGCOMM'98, pp. 315–323, Sep. 1998.
[29] K. Fall and S. Floyd, “Simulation-based Comparisons of Tahoe, Reno, and SACK TCP,” ACM Computer Communication Review, Vol. 26, Issue 3, pp. 5–21, July 1996
[30] M. Mathis and et al., “TCP Selective Acknowledgment Options,” RFC 2018, Oct. 1996.
[31] S. Floyd and T. Henderson, “The NewReno Modification to TCP's Fast Recovery Algorithm,” RFC 2582, April 1999.