OSPF design

[info]

Hi Guys,

I'm currently researching the possibility of the implementation of OSPF in our core network. First, let me share a topology. 

It basically concerns traffic from the subnets connected to firewall cluster 1 that is going to firewall cluster 2. That traffic passes through the core router. Currently, the core router is configured using static routes. Although that's working fine, we don't want to have this administrative overhead anymore and are therefore considering the implementation of OSPF.

Since OSPF uses the concept of areas, the information and discussion I haven't been able to find, is, what would be a good approach to the network design of this OSPF implementation.

For example:

-> Should we go with one "big" area 0 or should we put the router in area 0, the firewall cluster 1 in area 1 and the firewall cluster 2 in area 2? This part of the network (basically our core) can be considered stable, we will expect low convergence there.  The main advantage of multiple areas is that convergence is limited to a single area.

-> Since the firewalls have the default gateway pointed at the edge cluster, a stub area cannot be used. 

I'd love to hear your takes on this.

[Reza Sharifi]

The easiest deployment for this scenario is to put all the devices in one area (0) since you only have a max of 8 devices if the core is redundant.

HTH

[Inderjeet Singh]

You can think of multiple areas if you do not require to send all routes from cluster 1 to cluster2 .. or you need to send summary addresses from cluster1 to cluster2 or you need some sort of filtering.. Else, everythng can be in one area 0  to keep configuration simple...