Load Early Detection (LED): A Congestion Control Algorithm Based on Routers' Traffic Load

Efficient bandwidth allocation and low delays remain important goals, especially in highspeed networks. Existing end-to-end congestion control schemes (such as TCP+AQM/RED) have significant limitations to achieve these goals. In this paper, we present a new and simple congestion control scheme, called Load Early Detection (LED), that achieves high link efficiency and low persistent queue length. The major novelty of LED is that it gauges the congestion level by measuring the traffic load instead of the queue length. We show analytically that this feature of LED enables it to detect congestion earlier and react faster than other schemes. To gain insight into the behavior of LED and compare it with RED, we analyze a simple fluid model, and study the relation between stability and throughput, especially for low delays. We evaluate the performance of LED using extensive ns2 simulations over a wide range of network scenarios. Our simulation results show that LED achieves significantly higher link efficiency than RED (up to 83%), REM (up to 141%), and PI controller (up to 88%), especially in the presence of low delays.

[1]  IEEE Network Magazine , 2004, IEEE Network.

[2]  David D. Clark,et al.  Explicit allocation of best-effort packet delivery service , 1998, TNET.

[3]  Sally Floyd,et al.  Promoting the use of end-to-end congestion control in the Internet , 1999, TNET.

[4]  Steven H. Low,et al.  REM: active queue management , 2001, IEEE Netw..

[5]  Vishal Misra,et al.  Fluid-based analysis of a network of AQM routers supporting TCP flows with an application to RED , 2000, SIGCOMM 2000.

[6]  K. K. Ramakrishnan,et al.  A Proposal to add Explicit Congestion Notification (ECN) to IP , 1999, RFC.

[7]  Arjan Durresi,et al.  Guidelines for optimizing multilevel ECN using fluid-flow-based TCP model , 2002, SPIE ITCom.

[8]  Sally Floyd,et al.  TCP and explicit congestion notification , 1994, CCRV.

[9]  Walter Willinger,et al.  On the Self-Similar Nature of Ethernet Traffic ( extended version ) , 1995 .

[10]  Sally Floyd,et al.  Measuring the evolution of transport protocols in the internet , 2005, CCRV.

[11]  Wu-chang Fengy,et al.  BLUE: A New Class of Active Queue Management Algorithms , 1999 .

[12]  Donald F. Towsley,et al.  On designing improved controllers for AQM routers supporting TCP flows , 2001, Proceedings IEEE INFOCOM 2001. Conference on Computer Communications. Twentieth Annual Joint Conference of the IEEE Computer and Communications Society (Cat. No.01CH37213).

[13]  Raj Jain,et al.  Ordered BECN: Why we need a timestamp or sequence number in the RM Cell , 1994 .

[14]  Donald F. Towsley,et al.  A control theoretic analysis of RED , 2001, Proceedings IEEE INFOCOM 2001. Conference on Computer Communications. Twentieth Annual Joint Conference of the IEEE Computer and Communications Society (Cat. No.01CH37213).

[15]  Sally Floyd,et al.  Quick-Start for TCP and IP , 2007, RFC.

[16]  Walter Willinger,et al.  On the self-similar nature of Ethernet traffic , 1993, SIGCOMM '93.

[17]  Hitay Özbay,et al.  Introduction to Feedback Control Theory , 1999 .

[18]  Sally Floyd,et al.  HighSpeed TCP for Large Congestion Windows , 2003, RFC.

[19]  Sally Floyd,et al.  Adap-tive RED: An algorithm for increasing the robustness of RED , 2001 .

[20]  O. Yaniv.  IEEE Control Systems Magazine , 2001 .

[21]  Leonard Barolli,et al.  Congestion Control Using Multilevel Explicit Congestion Notification , 2007 .

[22]  Fernando Paganini,et al.  Internet congestion control , 2002 .

[23]  Raj Jain,et al.  The ERICA switch algorithm for ABR traffic management in ATM networks , 1998, TNET.

[24]  QUTdN QeO,et al.  Random early detection gateways for congestion avoidance , 1993, TNET.

[25]  Arjan Durresi,et al.  Congestion control using multilevel explicit congestion notification in satellite networks , 2001, Proceedings Tenth International Conference on Computer Communications and Networks (Cat. No.01EX495).