1. Home
  2. Cisco Systems
  3. How do you connect two OSPF autonomous systems?

How do you connect two OSPF autonomous systems?

You have two entirely separate OSPF systems. You wish to share routing information between the two systems, but you don't want to join the systems into one big OSPF system (you still want two area 0's).

Since OSPF does not use an AS number, as does EIGRP, how do you redistribute routes between the two systems? Is it as simple as using an ASBR and redistribute routes as you would any other non-OSPF routing protocol? In that case, wouldn't configuring two interfaces of an ASBR to join two different areas, of two different OSPF systems just screw up the protocol and confuse it?

Thanx Jim

Reply to
jimbo
Loading thread data ...

Right. Going to the first part of your reply, I did assume I would need two separate OSPF instances running, via the use of two different process IDs (and ofcourse joining the correct interface to the corresponding system). But it wasn't clear to me that simply by doing that, OSPF knows you are connecting the router to two different ASs. The books I have been studying emphasise that the process ID is NOT an autonomous system number. Yet, in this configuration, you EFFECTIVELY have used the process ID values to tell OSPF that the router is connecting to two different autonomous systems.

I suppose that since you could run two instances of OSPF with all interfaces in the same autonomous system, a process ID is indeed NOT an AS number, but you can use it as one. Does that make sense?

Thanx for the reply! J

Reply to
jimbo

Jim,

why don't you just configure two different OSPF processes, one for each AS? Make sure to put the correct interfaces into the corresponding OSPF process. Both OSPF processes are then ASBR for their AS respectively. You can then redistribute between the two OSPF processes (redistribute ospf...) using filtering along the way. As far as I understand, the main motivation for this kind of config is the possibility to apply filtering between both sides.

As far as I know, cisco doesn't exactly recommend that, but it is a valid and working configuration.

Otherwise, if you configure an OSPF-Process with interfaces in both of you "area 0" (the same process), you get one united "area 0" - as far as I understood thats what yo do not want.

Regards, Axel

jimbo wrote:

Reply to
Axel Gärtner

My recommendation would be to either bite the bullet and merge the two OSPF systems (assuming the only thing stopping you is the work involved) -or- treat the merge as two independent autonomous systems and use an inter-autonomous routing protocol (BGP) to control redistribution of routes between the two systems. The former optimizes route selection but requires complete trust and consistent implementation across both AS while the latter recognizes that route metrics and the like may differ between AS and supports a wide range of controls for route redistribution and usage.

If time is of the essence (whether time to implement or sensitivity to down time), go with BGP.

Reply to
Vincent C Jones

Understood. Thanx for taking the time . . .

Reply to
jimbo

Well yes, it has the same effect. Since the process-ID IS NOT an AS number, its scope is local to the router you configure it. Config would be

router ospf 1 network x.x.x.x

Reply to
Axel Gärtner

Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.