[Babel-users] Question about AHCP

[Henning Rogge]

Am Sonntag 29 November 2009 15:40:23 schrieb Juliusz Chroboczek:
> > (If the network would have a mesh-wide multicast implementation, you
> > could just use it and forget about the forwarders, right ?)
> 
> Bootstrapping issues.  The main reason AHCP implements its own multicast
> routing rather than relying on the IP layer is that at the point when
> you run AHCP, the network has not bootstrapped yet.
IPv6 should be up and running at this point, because it's stateless.

> I find it simpler and more reliable to make AHCP a self-contained
> protocol, hence the need for AHCP forwarders.
I understand this.

> > My problem with AHCP for IPv4 is that it's a little bit difficult to
> > manage in a mesh with a "distributed" administration. Do you ever
> > planned to add the possibility that AHCP-servers could somehow "trade"
> > free IP space so you don't need the totally strict IP separation of
> > servers. Or maybe to distribute the leases to other AHCP-servers so
> > you have some redundancy ?
> 
> Yes, that's prefix delegation.  Discussed on this list a few weeks ago.
> 
> The trouble with relying too much on prefix delegation is that it
> introduces a single point of failure,
Why ?

> so I personally prefer to have
> multiple completely independent AHCP servers.  Note that you should not
> need more than 2 or 3 AHCP servers in your mesh, so the amount of
> administration should be quite reasonable.
I think it would be nice if AHCP could support a "distributed/stateless" IPv4 
mode too.

The problem in IPv4 with the typical "roll up a random IP and do duplicate 
detection" IPv6 strategy is that the chance of collision is insane high 
(birthday paradoxon). But maybe there is some way to reduce this chance.

A good first step would be not to choose any IP from the available IP subnet 
if there is a route present which does cover only a part of the subnet. This 
way you can evade the obvious active IPs and reduce the chance that you 
collide with an IP already in use.
Another interesting point would be an external interface in AHCP where you 
could tell the protocoll "this IP is a collision, do something about it" or 
"this is a list of IPs you should NOT use". This way you could adapt "ipv4-
stateless AHCP" to situations where a different instance in the mesh node has 
"insider knowledge" about the used IPs.

Henning Rogge
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.alioth.debian.org/pipermail/babel-users/attachments/20091129/0f9c119b/attachment.pgp>