Multiple Area OSPF with 200+ branch offices.

[nurinalmas]

What is the best way to define OSPF areas for this OSPF environment?

One Main Central site.

One Disaster Recovery site.

200+ branch office sites connected via point-to-point leased lines to both the Main Central & Disaster Recovery sites. (ie. we have 200+ links to the Main Central site and another 200+ links to the Disaster Recovery site).

Early thanks for all responses.

[marcelo_sucholicki]

what topology do you have? is this star?

[nurinalmas]

Yes. Its a star WAN topology with leased circuits to every branch offices. The main problem is this:

Disaster Recovery Site do not have an aggregation router accepting the 200+ redundant connections from the branch offices. So I end up having a pair of routers for each link. Now, I have 200+ routers at the Disaster recovery site, connecting to the Core Switch 6509.

The Main Central Site is also accepting 200+ primary connections (leased circuits) from the Branch offices, but there is a pair of 7513 Aggregation routers.

Each branch office is a small LAN with 2 access routers, one router linking to the Main Site, and another redundant router linking to the Disaster Recovery site.

Considering the large number of routers and sites, I have a hard time designing the OSPF areas.

[JORGE RODRIGUEZ]

Hi, there are quite many things to look into. Backbone Considerations, area considerations. Types of topologies, mesh or hub and spoke.

IP addressing, Virtual links for backups/failover.

Generally for a large IP internetwork you would want to look into implementing a hierarchical design.

Here are couple of links with some reference.

OSFP design guidelines

http://www.cisco.com/univercd/cc/td/doc/cisintwk/idg4/nd2003.htm#16749

http://www.cisco.com/warp/public/732/Tech/routing/ospf.shtml

Rgds

Jorge

[rais · Accepted solution]

I think the following URL gives a very nice solution for your problem.

http://forum.cisco.com/eforum/servlet/NetProf?page=netprof&CommCmd=MB%3Fcmd%3Dpass_through%26location%3Doutline%40%5E1%40%40.ee8f3a5/6#selected_message

...running simpler protocol like RIP/EIGRP between remotes and your edge, then redistributing into OSPF.

You can also go the stub way and define each remote-network a unique stub-area, but that would mean that unwanted data may also head your way.

Thanks.

[nurinalmas]

I agree with you totally on all the considerations.

However, I have inherited the network at its current state. The physical infrastructure and the IP addressing scheme are there and I cannot change it.

Its a matter of optimizing the current network.