傳輸控制協定(TCP)被視為網際網路上最重要的傳輸層通訊協定，尤其是可靠的傳輸、流量控制和壅塞控制，然而TCP針對所有相互競爭資料流的公平分配頻寬的議題，卻沒有適當的處理辦法。隨著全球資訊網和互動性應用的流行，傳播於網際網路上的短流(short-lived flows)數量亦隨之成長，由於傳統TCP的不足，使得短流無法獲得可用頻寬的公平分享而必須忍受較長的延遲以及較低的流量。因此在網際網路上找出可以適應各種新流量類別問題的有效答案是必須的。
為能以更公平的方式分享瓶頸頻寬，本文提出兩個跨層級和不計狀態的佇列管理辦法，分別是Drop Maximum (DM)和Early Drop Maximum (EDM)，其根本的想法只是找出並丟棄使用超出平均配額資料流的封包，以及保持佇列使用率於較低水平。將TCP傳送端的congestion window大小存放於IP封包的options欄位，那麼運作於路由器的建議方法便可依照封包內大小來決定是否要被丟棄，例如發生雍塞，那麼佇列中擁有最大congestion window的封包將被丟棄，並使傳送端降低傳送率並釋出部分占用的頻寬給其他競爭的資料流，這樣的做法讓整個系統以快速和公平分配頻寬的方式趨於平衡。而不計狀態的方式更令建議的方法擁有實作和擴充性上的許多好處。
我們利用大量的模擬來驗證建議方法的可用性及有效性，在同質的情境上，Drop Maximum這個簡單的封包丟棄的方法，的確表現得比另外兩個不計狀態的佇列管理方法Drop Tail和Random Early Drop出色。然而在異質的情境上， Random Early Drop由於有額外的緩衝區控制機制而取得優於僅有封包丟棄機制的其他方法的表現，為突破這個困境，本文另外提出Early Drop Maximum方法。模擬結果顯示出這個方法的確在許多方面優於既有的不計狀態技術包括Drop Tail, Drop Maximum和Random Early Drop，例如可用頻寬的公平分配及短流回應時間的縮短。

Transmission Control Protocol (TCP) has been recognized as the most important transport-layer protocol for the Internet. It is distinguished by its reliable transmission, flow control, and congestion control. However, the issue of fair bandwidth-sharing among competing flows was not properly addressed in TCP. As web-based applications and interactive applications grow more popular, the number of short-lived flows conveyed on the Internet continues to rise. With conventional TCP, short-lived flows will be unable to obtain a fair share of available bandwidth. As a result, short-lived flows will suffer from longer delays and a lower service rate. It is essential for the Internet to come up with an effective solution to this problem in order to accommodate the new traffic patterns.
With a more equitable sharing of bottleneck bandwidth as its goal, two cross-layer stateless queue management schemes featuring Drop Maximum (DM) and Early Drop Maximum (EDM) are developed and presented in this dissertation. The fundamental idea is to drop packets from those flows having more than an equal share of bandwidth and retain low level of queue occupancy. The congestion window size of a TCP sender is carried in the options field on each packet. These proposed schemes will be exercised on routers and make its decision on packet dropping according to the congestion windows. In case of link congestion, the queued packet with the largest congestion window will be dropped from the queue. This will lower the sending rate of its sender and release part of the occupied bandwidth for the use of other competing flows. By so doing, the entire system will approach an equilibrium point with a rapid and fair distribution of bandwidth. As a stateless approach, these proposed schemes inherit numerous advantages in implementation and scalability.
Extensive simulations were conducted to verify the feasibility and the effectiveness of the proposed schemes. For the simple proposed packet discard scheme, Drop Maximum outperforms the other two stateless buffer management schemes, i.e. Drop Tail and Random Early Drop, in the scenario of homogeneous flows. However, in heterogeneous flows, Random Early Drop gains superiority to packet discard schemes due to its additional buffer occupancy control mechanism. To overcome the lack of proper buffer occupancy control, Early Drop Maximum is thus proposed. As shown in the simulation results, this proposed scheme outperforms existing stateless techniques, including Drop Tail, Drop Maximum and Random Early Drop, in many respects, such as a fair sharing of available bandwidth and a short response time for short-lived flows.

誌謝 ……………………………………………………………………………i
摘要 ……………………………………………………………………….…..ii
ABSTRACT iv
LIST OF FIGURES vi
LIST OF TABLES ix
Chapter 1 INTRODUCTION 1
1.1 HETEROGENEITY OF INTERNET TRAFFIC 1
1.2 AN EXPERIMENT ON TCP FLOW FAIRNESS 2
1.3 MOTIVATION 5
1.4 ORGANIZATION OF THE DISSERTATION 7
Chapter 2 APPROACHES TO FAIRNESS 8
2.1 END-TO-END APPROACH 10
2.2 HOP-BY-HOP APPROACH 16
2.2.1 Stateless Scheme 16
2.2.2 Stateful Schemes – Per-Flow Queuing 19
2.2.3 Stateful Schemes – Per-Class Queuing 20
2.3 JOINT APPROACH 21
2.4 SUMMARY 22
Chapter 3 A SIMPLE CROSS-LAYER STATELESS PACKET DISCARD SCHEME – DROP MAXIMUM 25
3.1 SOME OBSERVATIONS ON TCP FLOW FAIRNESS 26
3.2 ILLUSTRATION OF DROP MAXIMUM 32
3.3 A SIMPLE ANALYSIS OF PACKET DISCARD SCHEMES 33
3.4 PERFORMANCE EVALUATION 35
3.4.1 Homogeneous Flows – Long-lived Flows Only 37
3.4.2 Heterogeneous Flows – Long-lived Flows and Short-lived Flows …………………………………………………………………..40
3.5 SUMMARY 45
Chapter 4 A NOVEL CROSS-LAYER STATELESS QUEUE MANAGEMENT SCHEME – EARLY DROP MAXIMUM 46
4.1 DESIGN METHODOLOGY 46
4.2 ILLUSTRATION OF EARLY DROP MAXIMUM 54
4.3 IMPLEMENTATION ISSUES 56
4.4 PERFORMANCE EVALUATION 57
4.4.1 Single Bottleneck Link – Homogeneous Flows 59
4.4.2 Single Bottleneck Link – Heterogeneous Flows 66
4.4.3 Multiple Bottlenecks Link 77
4.5 SUMMARY 80
Chapter 5 CONCLUSION AND FUTURE WORKS 82
APPENDIX ………………………………………………………………………….85
REFERENCES 87

[1] C. Barakat, E. Altman, W. Dabbous, On TCP performance in a heterogeneous network: a survey, IEEE Communications Magazine 38 (1) (2000) 40–46.
[2] D. Bertsekas, R. Gallager, Data Networks (2nd ed.), Prentice-Hall, 1992.
[3] B. Braden, D. Clark, J. Crowcroft, B. Davie, S. Deering, Recommendations on queue management and congestion avoidance in the Internet, RFC 2309, April 1998.
[4] C. Brandauer, G. Iannaccone, C. Diot, T. Ziegler, S. Fdida, M. May, Comparison of tail drop and active queue management performance for bulk-data and Web-like Internet traffic, In: Proceedings of IEEE Computers and Communications 2001, July 2001, pp. 122–129.
[5] N. Brownlee, K.C. Claffy, Internet stream size distributions, In: Proceedings of ACM SIGMETRICS 2002, June 2002, pp. 282–283.
[6] D. M. Chiu, R. Jain, Analysis of the increase and decrease algorithms for congestion avoidance in computer networks, Computer Networks and ISDN Systems 17 (1) (1989) 1–14.
[7] M. Christiansen, K. Jeffay, D. Ott, F. D. Smith, Tuning RED for Web traffic, IEEE/ACM Transactions on Networking 9 (3) (2001) 249–264.
[8] K. Claffy, G. Miller, K. Thompson, The nature of the beast: recent traffic measurements from an Internet backbone, In: Proceedings of INET 1998, July 1998, [online] http://www.caida.org/Papers/Inet98/index.html.
[9] A. Clerget, W. Dabbous, TUF: tag-based unified fairness, In: Proceedings of IEEE INFOCOM 2001, April 2001, pp. 498–507.
[10] A. Demers, S. Keshav, S. Shenker, Analysis and simulation of a fair queueing algorithm, In: Proceedings of ACM SIGCOMM 1989, October 1989, pp. 1–12.
[11] S. Floyd, K. Fall, Promoting the use of end-to-end congestion control in the Internet, IEEE/ACM Transactions on Networking 8 (4) (1999) 458–472.
[12] S. Floyd, V. Jacobson, Random early detection gateways for congestion avoidance, IEEE/ACM Transaction on Networking 1 (4) (1993) 397–413.
[13] M. Fomenkov, K. Keys, D. Moore, K. Claffy, Longitudinal study of Internet traffic in 1998-2003, In: Proceedings of the Winter International Symposium on Information and Communication Technologies, WISICT 2004, Cancun, Mexico, January 2004, pp. 1–6.
[14] C. Fraleigh, S. Moon, B. Lyles, C. Cotton, M. Khan, D. Moll, R. Rockell, T. Seely, S. C. Diot, Packet-level measurements from the Sprint IP backbone, IEEE Network 17 (6) (2003) 6–16.
[15] S. Fu, M. Atiquzzaman, SCTP: state of the art in research, products, and technical challenges, IEEE Communications Magazine 42 (4) (2003) 64–76.
[16] P. Gevros, J. Crowcroft, P. Kirstein, S. Bhatti, Congestion control mechanisms and the best effort service model, IEEE Network 15 (3) (2001) 16–26.
[17] L. Guo, I. Matta, The war between mice and elephants, In: Proceedings of IEEE ICNP 2001, November 2001, pp. 180–188.
[18] V. Jacobson, Congestion avoidance and control, ACM SIGCOMM Computer Communication Review 25 (1) (1995) 157–187.
[19] A. Kantawala, J. Turner, Queue management for short-lived TCP flows in backbone routers, In: Proceedings of IEEE GLOBECOM 2002, November 2002, pp. 2380–2384.
[20] D. Katabi, M. Handley, C. Rohrs, Congestion control for high bandwidth-delay product networks, In: Proceedings of ACM SIGCOMM 2002, Pittsburgh, Pennsylvania, USA, 2002, pp. 89-102.
[21] H. Kung, M. Lin, F. Kuo, Dynamic QoS queuing control mechanism for multimedia differentiated services, Information Sciences 176 (2006) 3453-3471.
[22] L. Le, J. Aikat, K. Jeffay, F. D. Smith, The effects of active queue management on Web Performance, In: Proceedings of ACM SIGCOMM 2003, August 2003, pp. 265-276
[23] J. S. Li, C. F. Li, C. P. Lo, Effective fair allocation using smart market label auction with CSLF and CR-CSFQ, The Journal of Systems & Software 79 (2006) 665-673.
[24] H. S. Liew, S. F. Chien, S. W. Tan, A. L. Y. Low, K. N. Choong, Approximate fair bandwidth sharing in high-speed multimedia networks, BT Technology Journal 24 (2006) 85-89.
[25] R. Mahajan, S. Floyd, D. Wetherall, Controlling high-bandwidth flows at the congested router, In: Proceedings of the IEEE ICNP 2001, November 2001, pp. 192-201.
[26] M. Mathis, J. Semke, J. Mahdavi, P. S. Center, T. Ott, The macroscopic behavior of the TCP congestion avoidance algorithm, ACM Computer Communication Review 27 (3) (1997) 67–82.
[27] R. Morris, Scalable TCP congestion control, In: Proceedings of IEEE INFOCOM 2000, March 2000, pp. 1176–1183.
[28] I. Nikolaidis, X. Wu, Lifetime-based TCP service differentiation, Computer Communications 28 (1) (2005) 108–124.
[29] T. J. Ott, T. V. Lakshman, L. Wong, SRED: stabilized RED, In: Proceedings of IEEE INFOCOM 1999, March 1999, pp. 1346–1355.
[30] R. Pan, B. Prabhakar, K. Psounis, CHOKe: a stateless active queue management scheme for approximating fair bandwidth allocation, In: Proceedings of IEEE INFOCOM 2000, March 2000, pp. 942–951.
[31] K. K. Ramakrishnan, S. Floyd, Proposal to add explicit congestion notification (ECN) to IP, RFC 2481, January 1999.
[32] S. Ryu, C. Rump, C. Qiao, Advances in Internet congestion control, IEEE Communications Surveys & Tutorials 5 (1) (2003) 28–39.
[33] S. Shakkottai, T. S. Rappaport, P. C. Karlsson, Cross-layer design for wireless networks, IEEE Communications Magazine 41 (10) (2003) 74–80.
[34] M. Shreedhar, G. Varghese, Efficient fair queuing using deficit round-robin, IEEE/ACM Transactions on Networking 4 (3) (1996) 375–385.
[35] W. R. Stevens, TCP/IP Illustrated (Volume 1), Addision-Wesley, MA, 1994.
[36] W. R. Stevens, TCP slow start, congestion avoidance, fast retransmit, and fast recovery algorithms, RFC2001, January 1997.
[37] R. Stewart, Q. Xie, K. Morneault, C. Sharp, H. Schwarzbauer , T. Taylor, I. Rytina, M. Kalla, L. Zhang, V. Paxson, Stream control transmission protocol, RFC 2960, October 2000.
[38] I. Stoica, S. Shenker, H. Zhang, Core-stateless fair queueing: achieving approximately fair bandwidth allocations in high-speed networks, In: Proceedings of ACM SIGCOMM 1998, August 1998, pp. 118–130.
[39] X. Wu, I. Nikolaidis, On the advantages of lifetime and RTT classification schemes, In: Proceedings of IEEE International Conference on Performance, Computing, and Communications, April 2003, pp. 301–308.
[40] P. P. Xiao, Y. T. Tian, A TCP-friendly congestion control mechanism combined with the routers, In: Proceedings of International Conference on Machine Learning and Cybernetics, June 2007, pp. 3223-3228.
[41] S. Yilmaz, I. Matta, On class-based isolation of UDP, short-lived and long-lived TCP flows, In: Proceedings of 9th International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems, August 2001, pp. 415–422.
[42] H. Y. Zhang, J. Jiang, W. H. Dou, Fair bandwidth allocation mechanism with preference to short flows, Journal of Software 18 (2007) 765-774.
[43] L. Zhu, N. Ansari, G. Cheng, K. Xu, Edge-based active queue management, In: IEE Proceedings of Communications, 2006, pp. 55-60.