Re: [tsvwg] [saag] Comments on draft-ietf-tsvwg-transport-encrypt-08.txt

["Martin Thomson" <[e-mail]>]

On Wed, Nov 6, 2019, at 20:58, Colin Perkins wrote:
> It certainly wasn’t my intent with this draft to say that all these 
> practices should be kept; rather to stimulate the discussion about what 
> should be preserved, what it’s necessary to find alternatives for; etc. 
> If this doesn’t come across clearly in the draft. we should fix it.

Thanks Colin,

I didn't start with that expectation and the introduction does a fine job of setting this out.  However the draft makes some very strong statements in the conclusion and - in a few places - the body that leave very little room for interpretation.

I'm happy to help identify specific cases, but it seemed to be fairly consistent throughout.  For the bulk of the document, perhaps the right thing to do is take another pass with a critical eye.  I don't know what to do about the conclusions section though.

I just want to say that while I appreciate the effort taken to document techniques, I'm not entirely convinced that it is the only way forward.  Discussion of existing practice can distract from the underlying needs.

To take the example of capacity planning, I had a great conversation about 18 months ago with Lee Howard about what that might look like in the presence of more opaque data flows between endpoints.  I don't think that we reached a specific conclusion, but it was refreshing to talk about the actual problem and not get bogged down in the detail of specific techniques.  Though the discussion touched on many existing and potential methods, our focus never left the important question of what the goals were and how there might be some shared interest in achieving that goal.

Being able to focus on the underlying requirements like that can be hard, but my sense is that this is where we'll need to reach before anything meaningful will happen.  I have to confess, despite enduring the spin bit debate, I find the fact that I have still trouble articulating exactly what requirements that mechanism addresses to be concerning.