Re: [Snac] draft-lemon-stub-networks-05 for your consideration...

[Juliusz Chroboczek <[e-mail]>]

Hi Ted,

Thanks for the document.

I like Sections 3.1 through 3.4, at first sight they seem useful and easy
enough to implement.  I have never implemented a DNS-SD proxy, and hence
have no opinion on Sections 3.5 and 3.6.  Section 6 raises the right
issues, but I need more time to think about the mechanisms you propose.

On the other hand, and as I've already mentioned, I strongly dislike
Section 4.  Ted, would you be willing to consider splitting out Section
4 into a separate document?  By moving out all the complexity and
brittleness of NAT64 into a separate document, it would greatly increase
the shelf life of the base document.

One thing that concerns me is what happens when a stub network with
multiple routers boots.  Both routers will generate a ULA, then the router
with the numerically smaller ULA will deprecate its ULA.  (I have the same
concern about prefixes obtained from DHCPv6-PD in Section 6.1)

The current draft does not consistently distinguish between the infra and
the stub link: it often says things like "The stub router sends a router
advertisement" without specifying which interface is being used.  It's
easy enough to work out from the context what you mean, but it does take
some effort and makes the document more difficult to read.

I've also noticed a couple of instances of "this" without a clear
antecedent.  For example, Section 3.1.3 starts with "This is important"
without specifying what "this" is.  Again, the implied antecedent is easy
enough to work out from context, but I found myself stumbling and going
back in the document at that point.

-- Juliusz