[Iccrg] Initial value of dwnd

[[e-mail] (denis.collange)]

Hello,

We have a question concerning the Compound TCP draft, along the lines of
point 4 in
http://oakham.cs.ucl.ac.uk/pipermail/iccrg/2008-March/000468.html ("4.
What happens to dwnd during slow-start, and when is slow-start exited?")

The draft states that "dwnd is initialized to zero while the connection
is in slow start" (p. 6), and then "Delay-based component only kicks in
when cwnd is larger than some threshold, currently set to 38 packets
assuming 1500 byte MTU" (p.11).

However, starting from zero, the recursion (3) on p. 7 gives a negative
value, so that dwnd would always be set to 0. In order for dwmd to be
positive the initial value needs to be such that: 
	alpha*dwnd(t)^k - 1 > 0
(if dwnd is restricted to integer values a larger initial value might be
needed depending on how the rounding is done).

Clearly an initial value of 38 does satisfy this requirement but it is
not clear whether this is the initial value that should be assigned to
dwnd after slow-start or when the cwnd is larger then the specified
threshold (38 in the draft). 

We think that the draft should clearly state the initial value for dwnd
every time the delay component is activated.  It might also be
interesting to comment on why the window update function is not the same
as in the previously published papers about Compound TCP. 


Thanks,

Denis and Alberto
 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://oakham.cs.ucl.ac.uk/pipermail/iccrg/attachments/20081114/045524f5/attachment.html
>From [e-mail]  Fri Nov 14 19:15:39 2008
From: [e-mail] (Bob Briscoe)
Date: Fri Nov 14 19:07:20 2008
Subject: Heresy in Minneapolis (was: Re: [Iccrg] Heresy recapped)
In-Reply-To: <[e-mail]>
References: <[e-mail]>
	<[e-mail]>
	<[e-mail]>
	<[e-mail]>
	<[e-mail]>
	<[e-mail]>
	<[e-mail]>
Message-ID: <[e-mail]>

Matt & I have been having a side conversation about this.

In summary, we agree on what we don't want, but there's less 
consensus on the path ahead.

I'd like to suggest that those interested in what the IETF needs to 
do about relaxing TCP-friendliness make sure they're around in ICCRG. 
I've bcc'd a few folks who I suspect will be interested but might not 
naturally look in on ICCRG.

I haven't put this to the r-g chairs, but I imagine a discussion will 
start in response to Matt's talk, and might need some time to adjorn 
to a bar afterwards (there's one more session before the end of the 
day afterwards tho). Perhaps we'll get together a truly ad hoc Bar BoF :)

I know Matt is also talking on this in TSVWG & TSVAREA, but I imagine 
ICCRG ought to be where any initial activity migrates to (& I think 
Matt agrees).

My interest is that I believe TCP friendliness has become a 
self-imposed barrier to innovation. What's the point of having the 
e2e principle if you stop yourself and everyone else using the 
freedom it gives on some dodgy grounds you can't really justify?

See "Problem Statement: Transport Protocols Don't Have To Do Fairness "
<draft-briscoe-tsvwg-relax-fairness-01.txt>



Bob


At 03:23 16/04/2008, Matt Mathis wrote:
>On Tue, 15 Apr 2008, S. Keshav wrote:
>
>>Matt,
>>         The paradigm holds sway only in the minds of academia. I 
>> think (based almost purely on cynicism), that in the real world 
>> TCP friendliness never had a chance. There is a long history of 
>> TCP accelerators, multi-connection applications, UDP-blasters, 
>> packet classifiers-and-delayers, and who knows what else that have 
>> never cared about it. So, why do we need to phase it out?
>
>I would tend to agree.  However, isn't this list supposed to 
>represent academia?
>
>>Its already a done deal.
>
>Not in TCPM, TSVWG, etc, where dogmatic attachment to TCP-friendly 
>is probably hurting the IETF.  This is where we need to change minds 
>and some deeply entrenched documents.
>
>I think the ADs are probably listening - do they have any commemts?
>
>Thanks,
>--MM--
>
>_______________________________________________
>Iccrg mailing list
>[e-mail]
>http://oakham.cs.ucl.ac.uk/mailman/listinfo/iccrg

____________________________________________________________________________
Bob Briscoe, <[e-mail]>      Networks Research Centre, BT Research
B54/77 Adastral Park,Martlesham Heath,Ipswich,IP5 3RE,UK.    [phone]