I'm trying to understand the benefit of multiple areas in an OSPF implementation, if there is one for my needs. I manage the network for 10 buildings, with roughly 15 vLANS per building. Right now I have one area with all the routers in that area. It has been suggested to break-up the single area into three different areas (labeled a,b,c) but fail to understand the benefit. I tried to diagram the setup in ASCII, we'll see how it turns out.
R R --a - backbone
| \ / |
| DC |
R1 R2 --b
/ \ / \ / \ / \ / \ / \
R R R R R R R --c - Each router at this level has two uplinks, one to R1 and one to R2.
Is there any benefit to this design?
I wonder if you are familiar with this write-up
So the answer is "it depends on how do you plan your OSPF areas"
How OSPF areas are divided is sort of a religious-type conversation. I would suggest basing your area design on the following:
1) Natural network hierarchy - make sure 0 is at the center of the hierarchy (this may not be the top of the hierarchy)
2) Isolating areas for summarization and ASBR functionality
3) NSSA type areas for "leaf" networks
4) Do you need any filtering? if so, consider how your areas are laid out.
There are probably plenty of other reasons to break an area into an area. I generally just kind of 'see' where the areas make the most sense.