[Iccrg] Re: #2/2 draft-mayutan-ledbat-congestionarchitecture-00.txt

[Bob Briscoe]

Mayutan,

At 15:37 21/03/2010, Mayutan A. wrote:
>Hi Bob,
>
>Thanks once again for your feedback. See inline.
>
>On Sun, Mar 21, 2010 at 1:03 AM, Bob Briscoe 
><<mailto:rbriscoe at jungle.bt.co.uk>rbriscoe at jungle.bt.co.uk> wrote:
>Mayutan, Xiaoming, KK,
>
>Mail #2 of 2: Thoughts on the decomposition itself
>
>Bandwidth-estimation module
>---------------------------
>I propose it's better to think of this as a 
>Window-overshoot-mitigation module.
>
>i) Bandwidth is the wrong concept. A cc is looking for how much data 
>it can risk setting off in flight when it's got insufficient 
>information to determine a precise answer. That's a window concept 
>[unit: byte], not a bandwidth concept [unit: b/s]. Take QS. The 
>router thinks in terms of spare b/s. But the transport has to 
>multiply that by RTT to get the initial window it should use.
>
>
>Currently, what I have in mind is the estimation of actual bandwidth 
>to assist the flow-control module and leave it to the flow-control 
>module to convert this to a relevant window size. In theory, this 
>module could indicate window size too. Thanks for the alternate name 
>suggestion.

I'm not concerned about just the name. The semantics of window is 
what I think you need here, not bandwidth. This is important to get right.

[A similar example: Stas Shalunov pointed out to me once (I think 
repeating someone else) that if MTU had been defined in units of 
time, not bytes, packets would have grown naturally as link 
technology increased and data networking would have evolved very 
differently. Effectively the MTU is the liaison concept between IETF 
and IEEE. Similarly, the interface to your top module is important to 
get in the right units.]


>ii) It's not so much about estimating the right window. It's more 
>about ensuring the window is not too wrong, while trying to 
>otherwise be as greedy as possible.
>
>
>By estimation, I meant the above. Thanks for phrasing it well. I 
>should probably use a different word here or the name suggested by you.

Cheers



Bob


________________________________________________________________
Bob Briscoe,                                BT Innovate & Design 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://oakham.cs.ucl.ac.uk/pipermail/iccrg/attachments/20100321/8880de75/attachment.html