12 - OSPF Routing - Implementing OSPF over NBMA 1

Lets go over and cover the network types:

Hub and Spoke (Star)
Most cost effective (cheapest) but major draw backs are reduancy and LATENCY is a big hit, ecspecially on VoIP and delay, sensitive traffic!

Partial MESH
Only the major Sites have meshing and less important sites do not

Full MESG
Very expensive, but with newer technologies such as MPLS or PIVPN not such a big deal,
however we are not focusing on the technology at this point :oP

SO the MODES:
Well we have already covered and implemented 2 of the 5 modes;

Additional notes:
Here it acts just like it was an ethernet segment, but over the WAN instead, we have
multicast messages, we elect a DR and BDR ... BUT we have the neighbours set statically,

Also the DR and BDR ... it has to be R4, as both R2 and R3 do not have full connectivity and because of this will form a FULL relationship with R4 (however if this was a full mesh
then it would not matter who was DR and BDR as everyone would form a FULL relationship and this would be just like a ethernet segment)

But with partial mesh or hub and spoke you need to get involved and tweak it alittle

So, R4 will exchange routes with neighbours on 224.0.0.6 and then send the msgs back out on 224.0.0.5

But this depends on your network topology! If we take the example below a single subnet does not make sense here .... so you could  ...

Here we have chosen a different mode of OSPF for the 2 companies below who have just a PVC running between them, you can mix and match the modes!

The goal is to get these guys exchanging routes and forming neighbours over our NBMA cloud, so the point-2-multipoint is just another tool to make it happen :0)

Makes connections work exactly like point to point (multicast to 224.0.0.5) ...no DR and BDR
but its over NBMA,
ONLY DRAWBACK,. you need a unique subnet per neighbour that you are using over the cloud