Guys, i apologize if i’ve just missed a topic, but i can’t quite find a descriptive enough topic on ospf. Here’s the deal. I’ve read the manual twice and tried to implement it a few times. The only thing lacking in the manual is two areas, which is required(i think) in my goal. Anyways, just a few questions to hopefully get me started on the right track in my understanding.
In an area, do you have to have separate private ip subnets for each link between routers or can you have the same subnet across the area. Please bear with me, i might me missing something obvious.
Ex. Router A,B,C in area 1. Instead of A-B(10.0.1.0/24),A-C(10.0.2.0/24, B-C(10.0.3.0/24). Could you do: A(10.0.1.1)-B(10.0.1.2)-C(10.0.1.3)
In a situation with two areas…How does one area access the other area when they are on different ip subnets? I’ve seen where the ospf dynamically provides the routes out. I might just be missing a configuration attribute where i don’t see the route back down the chain.
I understand the router-id takes the highest available ip address, but what part does this actually play in the whole ospf environment.
1)if you use OSPF whitch is for big local autonomous sstems (AS) then you might want to use diferent subnets, also you can set more than two or routers in same subnet. you will not see any diference - OSPF will work for you if you configure OSPF correctly, if you create one big subnet, then you are using bridging and ruin all the routed network.
2)in one area there should not be more than like 40 - 50 routersso, 30 rotuers in one area - you do not have to worry about.
connection between areas are done by area border routers. that is router that is in two areas.
do not forget that backbone area should be continous
to identify what is not working if anything goes wrong
so on the area border routers, i need to redistribute default route? also, with the backbone area does that mean that i can assign all these little subnets to each area but then have one subnet for the backbone area across the network? Or is that the wrong approach?
You have to decide how much routing information needs to be visible inside the non-backbone areas for them to achieve optimal routing decisions. If they have only one gateway towards the backbone anyway, then yes, only injecting a default route into the non-backbone area can be a good solution. Furthermore, if you do not import external routing information into the non-backbone area that you need to forward to the backbone then you can make it a stub area to reduce the size of the link state database and cut down on the number of link state updates throughout your network.
The backbone area itself can be made up of any number of networks, it does not need to be only one. This is true for non-backbone areas as well.
