[Bloat] Bloat on Layer 2 Was: ECN & AQM Hall of Fame?

[Jim Gettys]

On 01/31/2011 12:35 PM, Dave Täht wrote:
> "Richard Scheffenegger"<rscheff at gmx.at>  writes:
>
>> BTW, I found this legacy document, where the authors boldly claim that
>> more buffers are always better for 802.11 networks, to circumvent
>> costly TCP congestion control decisions....
>> http://csl.snu.ac.kr/~ecpark/papers/TCP_WLAN_TMC08.pdf
>
>
> Citing Section 3.2
>
> "Effect of the Maximum Congestion Window Size on Fairness and
> Utilization"
>
>   "Based on the observation of asymmetric behavior of TCP congestion
>   control shown in Figs. 2 and 4, we can infer that the unfairness
>   problem can be alleviated by preventing packet loss from occurring. We
>   can avoid packet loss due to buffer overflow by either making the
>   buffer size, B, sufficiently large or by restricting the maximum
>   congestion window size, Wmax . In this section, we study the effect of
>   Wmax on fairness and aggregate throughput. We set B = 50 packets and
>   Wmax = 10..80 packets."
>
> I would love it if they could re-run their simulation setting "B"
> according to the buffer sizes for wireless devices we are now seeing in
> the field, which are in the 128..1500 packet range (not counting
> retries!), under poor radio conditions.
>
Note that lossy wireless networks behave much better than "clean" 
variable bandwidth wireless networks; packet drops due to such losses at 
least cause TCP to back off sometimes....
			- Jim