Re: [re-ECN] Charter Question

[John Leslie <[e-mail]>]

[e-mail] <[e-mail]> wrote:
> 
> On Mon, May 10, 2010 at 01:43:53PM -0400, John Leslie wrote:
>> 
>> Recall, the proposed charter says:
>> ] 
>> ] The mechanism to be developed by the CONEX WG will enable the sender
>> ] to also relay the congestion information back into the IP layer,
>> ] such that the total level of congestion is visible to all IP devices
>> ] along the path.
>> 
>> Thus, at any point along the path -- most likely near the sender
>> or near the receiver -- it will be practical to aggregate the expected
>> congestion for some flows and see whether this is consistent with the
>> contractual arrangement.
>> 
>> What to do if it isn't is clearly out-of-scope.
> 
> er... you have already proposed a use for the congestion information,
> i.e. aggregate flows... and then, its no longer the original
> sender/recevier,

   My bad... I meant "flow(s)". One can aggregate congestion-marking of
packets for a single flow.

> is the node that has chosen to mediate/aggregate and the source/target (or
> other middle-box that has chosen to aggregate/deaggregate based on local
> policy)

   I don't understand that question.

   As far as "policy", I didn't envision anything like the "policies"
of BGP routing. I envisioned contractual arrangements between ISP and
customer about how much congestion the customer will "cause" -- where
I believe the average customer doesn't intend to cause much, but
other customers might be willing to pay for a higher grade of service.

>> This -- and only this -- is "the intended use of the tool".
> 
> me thinks this vastly exceeds Bobs description - its moved past 
> a protocol change; "relay congesstion information back into the IP
> layer" to a specific implementation choice.

   Neither Bob nor I speak for the WG that may form.

   (Bob can speak to how much violence I've done to his design; I
don't think I've done much at all...)

   And, of course, one _must_ move past the protocol definition to
reach an "intended use".

>> Being who we are, we inevitably speculate about other uses such
>> a tool will enable. Any good tool enables uses beyond the one it
>> was designed for.
> 
> amen - but lets get the tool working before we try and use it for
> a given purpose.

   Agreed!

--
John Leslie <[e-mail]>