Re: [re-ECN] Charter Question

[John Leslie <[e-mail]>]

Ron Bonica <[e-mail]> wrote:
> On 5/10/2010 2:25 PM, Woundy, Richard wrote:
>> [Ron Bonica wrote:]
>> 
>>> This is troubling. Before we invest in tool development, we really
>>> should know something about the intended use of the tool.
>> 
>> I think I gave an example on this list last Friday.
>> 
>> See <http://www.ietf.org/mail-archive/web/re-ecn/current/msg00690.html>.
> 
> I think that you gave a very fine example. Unfortunately, according to
> John, it is out of scope.

   There may be some misunderstanding here...

] [Ron Bonica wrote:]
]] If I do have this right, who will those routers use this information?
]
] I believe we consider that question out of scope;

   I do consider the quetion of what any existing router does with the
information out of scope.

   But we have a pre-WG document which I believe _will_ see some work
after the WG is formed:

http://tools.ietf.org/html/draft-tschofenig-conex-ps-02

describing the problem we set out to address.

   I can understand that someone working for a router manufacturer may
be concerned how this will affect router design.

   But there appears to be a consensus (whatever that may mean in the
absence of a WG) that the answer is, "Not at all." We're thinking in
terms of middlebox operations, and asking no more of existing routers
than to perform ECN marking if it's convenient.

   (BTW, draft-tschofenig-conex-ps basically says that ISPs want to do
the right thing, but in the absence of congestion exposure they're
forced to use less appropriate metrics to "estimate" congestion.)

> Rather than putting it out of scope, I think that it should be the WGs
> first order of business.

   I agree a Problem Statement deserves to be an early milestone.
What do you think is needed that's not in draft-tschofenig-conex-ps?

--
John Leslie <[e-mail]>