[Iccrg] Re: #1/2 draft-mayutan-ledbat-congestionarchitecture-00.txt

[Mayutan A.]

Hello Bob,

Thanks a lot for your feedback. See inline for comments.

On Sun, Mar 21, 2010 at 12:48 AM, Bob Briscoe <rbriscoe at jungle.bt.co.uk>wrote:

> Mayutan, Xiaoming, KK,
>
> It seems a good idea to try to decompose the sub-functions of CC. I suspect
> this is best kept to the CC design process though (primarily ICCRG). Rather
> than an organising principle for standardisation as well (as suggested in
> the draft).
>

Our primary focus was to attempt such a decoupling in the LEDBAT WG for a
scavenger transport protocol, before attempting a more general one in the
ICCRG.

>
> So I'll split my thoughts into 2 mails:
> #1 (this one) Applicability of the decomposition: design, or standards too?
> #2 (next) thoughts on the decomposition itself
>
> Sub-questions to help determine whether we should decompose CC standards:
>
> Q1. Are we likely to need lots of different scavenger CCs, or are we more
> likely to be refining and improving a single track of progressively more
> generic/graceful/cool ones?
> A1. I say we probably need just one track
>

I would agree to having just one track too, but based on a decoupled
architecture (not completely decoupled though). This would make it easier to
propose extensions. For e.g, if someone comes up with a better delay based
congestion detection tool, or a more efficient available bandwidth
estimation tool, we could just standardize the mechanism. The mechanism
could then be used not only by a scavenger protocol, but also by other
congestion control protocols. To phrase it differently, once the LEDBAT WG
standardizes a one-way delay based scavenger protocol, other CC protocols
may want to use the one-way delay based mechanism to detect congestion, but
their reaction to detected congestion might vary from that of a scavenger CC
protocol.

>
> Q2. Does experience show that the sub-functions are cleanly separated, to
> the extent that they could be plugged together in different ways?
> A2. I think the skill of designing a good congestion control is about how
> well the parts complement each other. In fact, one theme of the talk I'm
> planning for ICCRG this week is that we need to be more careful to think
> about the cross-effects of some flows in slow-start on others in congestion
> avoidance (& vv) when designing CC.
>

I agree that the various modules need to complement each other. This draft
is an attempt to study how and to what extent can we decouple the various
components.

So rather than showing Fig 1 as if there is one common interface between the
> paris of modules, it's more important to show that the interface depends on
> which type of outer module is used. For instance "CONGESTION_DETECTED" could
> be a binary congestion event, or a delay variable, or a QS bit-rate.
>

Currently, what I had in mind was a binary indicator
(congested/not-congested) or a multilevel indicator.  We need to give more
thought to whether we require multiple interfaces between the modules or
could we come up with a method to map the various indicators to a common
"CONGESTION_DETECTED" indicator.

>
>
> Here's an example of where we've already got more modularisation than we
> should have in the TCP standards (in your I-D's terminology, the flow
> control module behaviour should depend on the which congestion detection
> module we choose):
>
> If my window is 30x yours, and I get 40 ECN marks within one RTT and you
> get 1, I should reduce more than once within the RTT. Whereas if I get 40
> drops in one RTT, it's more correct to only reduce once.
>
> Reasoning: If congestion detection is by ECN not drop, it implies the
> congestion must be from AQM not tail drop. Whereas 40 drops can mean a whole
> load of synchronised tail drop, 40 ECN marks can't mean synchronised drop.
> So I should take 40 ECN marks more seriously.
>
> I would take this example as a pro for a decoupled architecture. The
congestion-detection module could take this into consideration while mapping
the
obtained indication to the  "CONGESTION_DETECTED"  variable. The flow
control module could be designed based on just the value returned by the
"CONGESTION_DETECTED" variable and need not have to differentiate between a
drop, an ecn marking or increase in delay.


>
> Finally, another thought on applicability: This I-D isn't only applicable
> to scavenger CCs. It's clearly applicable to all CC design work (again
> primarily ICCRG).
>
> That's not to say it can't help LEDBAT in future. LEDBAT is on the
> experimental track, so it would be fine if we learn something from this
> decomposition that can help improve LEDBAT between experimental and Proposed
> Standard stages.
>

Yes, we are planning to present it at the LEDBAT and make a short
presentation at the ICCRG too. We feel that the LEDBAT WG might be a good
starting point, and that the approach could be extended to all kinds of
congestion control protocols.

Mayutan

>
> Bob
>
>
> At 19:40 03/03/2010, Mayutan A. wrote:
>
>  We have submitted a new draft to the LEDBAT WG and are planning to make a
>> short presentation at the LEDBAT WG and the ICCRG.
>> We are looking forward to all your comments.
>> <
>> http://tools.ietf.org/html/draft-mayutan-ledbat-congestionarchitecture-00
>> >
>> http://tools.ietf.org/html/draft-mayutan-ledbat-congestionarchitecture-00
>>
>>
>> Feedback is welcome,
>> Mayutan Arumaithurai
>>
>
>
>
>
> ________________________________________________________________
> Bob Briscoe,                                BT Innovate & Design
>



-- 
Mayutan Arumaithurai
0049-551-2712647 (Home number)
0049-176-20322049 (mobile number)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://oakham.cs.ucl.ac.uk/pipermail/iccrg/attachments/20100321/6a8d9f42/attachment.html