Re: types of traffic in tcp congest control

[Joe Touch <[e-mail]>]

J Wu wrote:
> On Sun, 12 May 2002, Joe Touch wrote:
> 
> 
>>
>>J Wu wrote:
>>
>>>Hi Li,
>>
>>...
>>
>>>As my understanding of S.Low's paper that the mice stands for small scale
>>>transmition with delay sensitive and the elephant stands for large scale
>>>transmition. The small scale transmitions will not cause any congestion,
>>>but the large scale transmitions will.
>>
>>That is an assumption, one which other results, notably of bandwidth
>>limited highly-shared core links (e.g., the US-UK link) refute. They
>>indicate that mice can indeed cause congestion amongst themselves, even
>>in the absence of elephants.
>>
>>Joe
> 
> 
> Hi Joe,
> Well, if only the mice can still congest the link, which shows that the
> network resources are under severe shortage. The only way to solve under
> this situation is hardware solution. But I think even under this situation
> restrict the elephants will also do benefit to alleviate the stage of
> congestion.

Only if there _are_ elephants. In the transatlantic case, a huge number 
of mice can swamp the connection - yes, a severe shortage. One solution 
is to "RED" the SYN packets of new connections, or limit how many 
connections can be established.

> And also, it seems that most network congestions are take place at access
> network which is not the core network.

See above; this congestion, by definition, takes place inside the core, 
not at the access.

> Another question maybe far from this topic: As the invent of new core
> transmission techniques like DWDM, is congestion control still needed in
> core network?

Overprovisioning, also somewhat by definition, precludes the need for 
congestion control. Whether WDM technologies will provide sufficient 
overprovisioning, or just will amplify the disparity between the 
all-optical high-speed and electronic lowspeed paths is still an open 
question.

Joe