[Olsr-users] Flooding IPv6 Prefix with olsrd2

[Juliusz Chroboczek]

> We discussed HNCP in a small group here. At the end we agreed that HNCP
> should be used on user nodes which are comparable to CPE devices [1].
> In the first place we discarded to use HNCP in mesh because it looks
> like overkill for our special problem.

Please distinguish between Homenet (which defines a particular network
architecture) and the HNCP protocol, which is able to do much more than
just the Homenet arch.  Both hnetd and shncpd implement a number of
features that are not in Homenet (but perfectly legal according to the
HNCP protocol).

Shncpd is just 4000 lines of code and compiles to just 48 kB of text on
AMD64, and this includes a DHCPv4 server and an RA server which could
be compiled out if necessary.

> - How many messages are generated in a mesh with 500 nodes where the
> infrastructure is not static?

Way too much, granted.

  https://tools.ietf.org/html/draft-jin-homenet-dncp-experience-00

> As far as I know every node knows the status of all other nodes.

Yes, that's the big fail of HNCP.  It makes me really angry.

(What's worse, the prefix assignment algorithm doesn't even need that
information -- but the IETF has historically been dominated by link-state
people.)

> Whenever the status of one node changes (e.g. new neighbour) all other
> nodes are informed.

Yes.  This can be worked around somewhat by adding hysteresis.

> - Do we need all this? We only need prefix and address assignment.

Please look at shncpd, and let me know which bits need to be made optional.

> - For every local link a /64 net is choosen in Homenet.

Shncpd supports assigning just a single address to an interface -- see the
shncpd manual page and the "-M" flag.  If you need a different address
assignment policy, this can easily be added.

-- Juliusz