[Babel-users] Babel at WBM and "hoping over the same channel is evil"

[Henning Rogge]

On Donnerstag 23 April 2009 15:55:08 Juliusz Chroboczek wrote:
> >> and if you want to understand why it's possible in Babel but not in OLSR
> >> to do approximations of their metric, check the Babel draft, section
> >> 2.5.2.
> >
> > I'm pretty sure it's possible with a proactive protocol like OLSR (of
> > course it would require an extension in the TCs to transport the
> > channels of the links).
>
> Babel is proactive too.  The difference is that OLSR is link-state,
> while Babel is distance-vector.
>
> The issue is that a metric that takes deversity into account looses
> a property known as isotonicity.  In a link-state protocol, you cannot
> use a non-isotonic metric without persistent routing loops.  Babel can
> use a non-isotonic metric just fine, although in that case it is no
> longer deterministic.
The MIC paper I have contains a sollution for the "non-isotonic" parts of MIC.

> (In practice, this means that channel assignment will depend on the
> order in which your routers boot.)
>
> There are solutions to the isotonicity problem, but they require using
> multiple routing tables.  If you implement that in OLSR, I'm looking
> forward to stealing your code ;-)
You just have to set the "source interface" for the different routes. Should 
not be difficult.

Henning
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.alioth.debian.org/pipermail/babel-users/attachments/20090423/f69b8d46/attachment.pgp>