Re: [arch-d] Call for Comment: <draft-trammell-wire-image-04> (The Wire Image of a Network Protocol)

[Christian Huitema <[e-mail]>]

On Sep 15, 2018, at 9:55 AM, Randy Bush <[e-mail]> wrote:

>> Ok. The motivation for this draft is indeed he increasing deployment
>> and coverage of encryption down the stack, which we take as a given. A
>> few sentences to make this context clear could be useful.
> 
> i kind of liked just saying that strong encryption is becoming
> ubiquitous, is here to stay, and the ietf thinks that is a good thing.
> this has consequences for applications and middleboxes that have grown
> used to being able to see the traffic in cleartext.
> 
>> The whole point of this line of work is to define a solution space for
>> the (technical) problems that arise when “strong encryption is here to
>> stay”
> 
> for some of the consequences, there is no "solution."  this may not be a
> bug.

We discussed that a lot when reviewing Kathleen's draft. There is a grab bag of stuff that have been put under the "network management" umbrella, from monitoring whether a given path is still working to being able to insert or replace ads. There is no doubt that some of that is legit and useful. The question then is where to place the line between "yes that's useful" and "forget about it". And then, how to best accommodate the useful part when most of the packet is encrypted.

-- Christian Huitema