[Babel-users] draft-ietf-babel-rfc6126bis-02

[David Schinazi]

Juliusz,

Thanks for making these edits, they look great.
The only real issue is the handling of NextHop and RouterID
with unknown mandatory sub-TLVs, which was discussed on another thread.

I also think Appendix C (Considerations for protocol extensions)
should be changed now that we have mandatory sub-TLVs.
I'll try to contribute text.

David


> On May 24, 2017, at 10:22, Juliusz Chroboczek <jch at irif.fr> wrote:
> 
> Dear all,
> 
> I've just published a new version of the Babel protocol specification:
> 
>  https://tools.ietf.org/html/draft-ietf-babel-rfc6126bis-02
> 
> This version containts some fairly major changes, the most notable being
> the addition of mandatory bits to the extension subprotocol.  There have
> also been some fairly technical changes to the procedures for sending of
> requests, which should not invalidate any existing implementations.
> 
> The mandatory bit makes the protocol more easily extensible by making it
> possible to explicitly encode the fact that an extension is not backwards
> compatible.  It has greatly simplified the packet format of Matthieu
> Boutier's source-specific extension [1], and is used by Gwendoline
> Chouasne's TOS-specific extension [2].
> 
> [1] https://github.com/boutier/babeld/tree/dev
> [2] https://github.com/Gwendocg/babeldToS
> 
> Both babeld and sbabeld have support for mandatory bits in their
> "mandatory" branches.  I'll wait a few days to see if there are any flaws
> in this proposal, then merge into trunk.  Please consider implementing
> mandatory bits if you have an implementation of Babel.
> 
> The backwards compatibility of this change is reasonably strong, but
> somewhat weaker than what we at Babel Towers have been doing previously.
> More exactly:
> 
>  - new implementations of Babel will interoperate with old
>    implementations as long as the former don't use any extensions that
>    the latter don't understand;
>  - new implementations of Babel that use the new extensions (new-style
>    source-specific routing, TOS-specific routing) will not interoperate
>    with old implementations, and might even create routing loops.
> 
> We will refrain from deploying the new extensions until all implementations
> have acquired support for mandatory bits.
> 
> Please read.  Please think it over.  Please comment.
> 
> -- Juliusz
> 
> _______________________________________________
> Babel-users mailing list
> Babel-users at lists.alioth.debian.org
> http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/babel-users