[manet] Re: BABEL question

"Templin \(US\), Fred L" <[email protected]> Mon, 21 Jul 2025 15:22:41 +0000
Newsgroups gmane.ietf.manet
Message-ID <BN0P110MB1420D2301876DB46B828D320A35DA@BN0P110MB1420.NAMP110.PROD.OUTLOOK.COM>
HI Hennig,

> -----Original Message-----
> From: Henning Rogge <[email protected]>
> Sent: Sunday, July 20, 2025 11:44 PM
> To: Templin (US), Fred L <[email protected]>
> Cc: Juliusz Chroboczek <[email protected]>; [email protected] List <[email protected]>
> Subject: [EXTERNAL] Re: [manet] Re: BABEL question
> 
> > > > I think that is the only alternative for disconnected BABEL clouds that
> > > > have not recently been in contact with an Internet gateway that
> > > > advertises a non-LL prefix - or is there some other IPv6 prefix that can
> > > > be used in the disconnected case?
> > >
> > > fc00::/7
> >
> > ULAs in their standard form are topology-oriented addresses; nodes in isolated
> > MANETs that may partition and/or overlap with other MANETs require some
> > form of topology-independent address:
> 
> Not really, as long as you don't route the whole prefix into the
> interface. Babel (and most other adhoc routing protocols) will happily
> provide all the host routes going into the interface, allowing
> partitioning the network without issues.

The point is that RFC4193, Section 3.1 defines a format for ULAs that includes a
40-bit "Global ID" followed by a 16-bit "Subnet ID" field in the /64 prefix. These 
values are to be assigned by infrastructure reference points, and not randomly
assigned by some isolated MANET router seeking to configure a unique /128.
If MANET routers did that, then there would be opportunity for conflict with 
a properly-administered ULA prefix if the MANET ever came in contact with the
Internet.

ULAs are therefore "topology-dependent" addresses when what is really
needed for MANET is some form of "topology-independent" address. The
topology-independent address should have no subnet prefix structure in
the leading N bits and should provide maximal bits for a node to self-configure
statistically unique values in the trailing N-128 bits in order to form a truly
unique /128. There would then never be any possibility for overlap with a
topology-dependent prefix such as one received in an RA PIO, while there
would be a well-known IPv6 prefix value to differentiate these topology
independent "MANET Local Addresses" from topology-dependent ULAs.

Fred


_______________________________________________
manet mailing list -- [email protected]
To unsubscribe send an email to [email protected]