Moving the Ball on a lot of Small OSPF Points

Analysis
Dec 8, 20083 mins

I started a thread of posts on OSPF config issues a few posts ago. This week, I’m going to load up and wrap up this one example, and maybe take us a little deeper on OSPF after that. But first, today, I’ll get to answering the OSPF Router ID question, the neighborship question, and then discuss the issue of what the routers advertise – so on the next post, I can discuss what LSAs flow.

First, to wrap up the Router ID issue, here’s what each router chooses. (To see the full config, go here.)

  • R1: 10.1.1.1, because the loopback is shut down, leaving Fa0/0 as the only interface that has an IP address (and the only one that is presumably up/up.)
  • R2: 10.1.1.222, because it’s the only interface.
  • R3: 3.3.3.3, because of the two working loopbacks, 3.3.3.3 is the numerically higher IP address.

The main gotchas here were the fact that the interface has to be up at the time OSPF is initialized, and that it’s the highest IP address, rather than the highest interface number. Assuming that interfaces work unless they are shutdown, the above lists the correct OSPF RIDs.

As you might have guessed, I wasn’t going to make up a problem that asked about neighbor relationships without purposefully breaking at least one. 😉 Looking at the list of restrictions from my last post, with some quick subnetting math, you can see that R2 will have a problem because its IP address on Fa0/0 is in a different subnet as compared to R1 and R3, as follows:

R1 and R3: Subnet 10.1.1.0/26, range of addresses 10.1.1.1    – 10.1.1.62

R2:             Subnet 10.1.1.192, range of addresses 10.1.1.193 – 10.1.1.254

So, R2 won’t become neighbors with either R1 nor R3.

Looking at the list of restrictions on becoming neighbors, R1 and R3 don’t have any problems. So, at the end of the process, R1 and R3 will become neighbors on the LAN, and R2 will not.

Next, let’s take a look at a new part of the problem – namely, what subnets will the routers try to advertise? Knowing that information will be helpful in then figuring out what LSAs to expect to end up the three routers’ OSPF databases.

When OSPF is enabled on an interface, OSPF (usually) does two things:

1) Performs neighbor discovery by sending Hellos

2) Advertises about the subnet connected to the interface

I’m going to concentrate on number 2 for now. So, the question is, on which interfaces do R1, R2, and R3 enable OSPF? Well, for those of you who enjoy having a little time to answer for yourself, I’ll wait to post the literal answer until Tuesday (via just a response to this post).

Additionally, start chewing on the following questions, which I’ll write about in my next post later in the week:

What LSAs (and types) will each router know?

What routes will each have in their routing tables?