Hi Bernie, 

Thank you for the review. 

Please see inline.

Cheers,
Med

> -----Message d'origine-----
> De : Bernie Volz (volz) [mailto:[email protected]]
> Envoyé : jeudi 12 janvier 2017 14:49
> À : [email protected]
> Cc : [email protected]; [email protected]; Terry Manderson
> Objet : Review of draft-ietf-softwire-multicast-prefix-option-11
> 
> Hi:
> 
> A few comments on this draft.
> 
> Section 2:
> 
> The dslite-multicast draft uses uPrefix64 instead of U_PREFIX64?
> Consistency between the various softwire documents might be useful?

[Med] Works for me. 

> 
> Nit: SSM_PREFIX64 has an extra closing parenthesis (after the RFC4607
> reference).

[Med] Fixed.

> 
> Section 3:
> 
> I am not sure that the *-length fields are correctly documented? Can they
> really be 0-128. I understand 0 if there is none, and I can see 1-96
> potentially, but does more than 96 make sense? It seems that the other
> documents referenced typically say to use the prefix (96 bits – perhaps
> zero extended if needed to make 96 bits) and then add 32-bits IPv4 for
> multicast (unicast is a bit more complex when prefix is not 96). But none
> of this other text seems to imply that the prefix length can be > 96?

[Med] As mentioned in Section 5, only 96 is allowed. I already fixed Section 3 
accordingly. 

> 
> Also, while 2001::/96 could be sent as 2001::/16 in terms of the option
> encoding (as bits 17 to 96 are 0), the two are vastly different in terms
> of what they might mean (if the prefix is always mapped to /96 it may not
> matter). I just wonder if this needs to be clarified some? I believe the
> first encoding (2001::/96 is always intended as that communicate the true
> prefix length).
> 
> Is the reference to “Section 6 of RFC7051” correct? Was this intended to
> be RFC 7227 (Section 6 there is titled “Avoid Conditional Formatting”).

[Med] Section 6 of RFC7051 is correct. It documents the use of Well-known DNS 
name for the unicast-only case. I updated the text to make this explicit. 

NEW:
   Such side effect conflicts with the recommendation to support the
   Well-Known DNS Name heuristic discovery-based method for unicast-only
   environments (Section 6 of [RFC7051]).

> 
> I think it would be useful to clearly state that multiple instances of
> this option are allowed. Yes, if you carefully read the text (in Section 4
> & 5) this is evident, but it is useful to state this explicitly when
> defining the option!

[Med] Happily. I added to NEW text to section 3:

"Multiple instances of OPTION_V6_PREFIX64 may be returned to a DHCPv6 client."

> 
> Section 5:
> 
> Nit: There are two “Pv4 mulitcast” instances; are these supposed to be
> “IPv4 multicast”?

[Med] Fixed.

> 
> Note here that the text for ASM and SSM says to insert last 32-bits. This
> would imply that a prefix length > 96 does not work. (See comments above
> for section 3.)

[Med] Yes, only 96 is allowed. See above.

> 
> And, there is talk here about scope selection and adding the same
> reference as in Section 4 (to RFC7346) might be useful? Client
> implementers may not read the server section, so duplicating the reference
> here would be useful.

[Med] Works for me. Done!

> 
> - Bernie
> 
> 
> 

_______________________________________________
Softwires mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/softwires

Reply via email to