Assessing LEDBAT's Delay Impact

The major goal of the LEDBAT congestion control mechanism is to limit the amount of queuing delay that it causes. Using simulations, we show two aspects of the mechanism that may sometimes make LEDBAT's delay impact quite significant: first, it updates its "base delay" every 10 minutes, but may measure its own induced delay in doing so, even in the presence of significant transmission jitter. This may lead to a continuous growth of the fixed delay that a LEDBAT flow produces if it runs for long enough. Second, it requires competing traffic to be long-lived and greedy for it to reduce its rate accordingly, however, such traffic is in general not latency-critical.

[1]  J Gettys,et al.  Bufferbloat: Dark Buffers in the Internet , 2011, IEEE Internet Computing.

[2]  D. Wischik Short messages , 2008, Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences.

[3]  Dario Rossi,et al.  LEDBAT: The New BitTorrent Congestion Control Protocol , 2010, 2010 Proceedings of 19th International Conference on Computer Communications and Networks.

[4]  Rolf Winter,et al.  Out of my way - evaluating Low Extra Delay Background Transport in an ADSL access network , 2010, 2010 22nd International Teletraffic Congress (lTC 22).

[5]  Michael Welzl,et al.  Less-than-Best-Effort Service: A Survey of End-to-End Approaches , 2013, IEEE Communications Surveys & Tutorials.

[6]  Amuda James Abu,et al.  Impact of Delay Variability on LEDBAT Performance , 2011, 2011 IEEE International Conference on Advanced Information Networking and Applications.

[7]  Pei Cao,et al.  NS-2 TCP-Linux: an NS-2 TCP implementation with congestion control algorithms from Linux , 2006 .

[8]  Janardhan R. Iyengar,et al.  Low Extra Delay Background Transport (LEDBAT) , 2012, RFC.

[9]  Dario Rossi,et al.  The Quest for LEDBAT Fairness , 2010, 2010 IEEE Global Telecommunications Conference GLOBECOM 2010.