[Babel-users] [Henning Rogge] Re: Ideas about Babel

[Juliusz Chroboczek]

> typically it's not necessary for OLSR to have a good knowledge about
> nodes far away, you just need a rough view of the big picture, because
> your routing algorithm will only calculate the next hop.

This is indeed the philosophy of EIGRP, DSDV, AODV and Babel.  These are
distance-vector protocols that include mechanisms that ensure that even
when nodes have a partial view of the network, routing loops do not happen.

This is not the case of SPF-based protocols, which under some circumstances
will create temporary routing loops.  OSPF includes two mechanisms to
ensure that these routing loops disappear after a bounded time:

  1. LSA flooding is reliable in OSPF;
  2. SPF is only computed within a single area, and distance-vector is used
     for routing between areas.

> resynchronization in OLSR is done through regular floods of TCs and
> timeout values.

I understand that.  Unfortunately, in the presence of packet loss, this is
not enough, and large OLSR networks have found the need to reduce disable
the MPR mechanism, and to reduce the TC interval to ridiculously low values
in order to ensure that resynchronisation happens in a timely manner.

Since OLSR was designed for dense networks, it should come as no surprise
that it doesn't work well in sparse networks.

                                        Juliusz