Cisco Support Community
Showing results for 
Search instead for 
Did you mean: 

Welcome to Cisco Support Community. We would love to have your feedback.

For an introduction to the new site, click here. If you'd prefer to explore, try our test area to get started. And see here for current known issues.

New Member

OSPF - connecting multiple remote sites via LES

What is good general practice if for example you have 15 remote locations connecting to a central site. IE: hub and spoke.

Each spoke of the topology is a lan extension service.

The routing protocol is OSPF

would it be better to have each LES circuit configured as a P-P link - unique subnet /30 bit mask? or single subnet for all routers eg: /24bit mask.

A s'L3' switch running OSPF is currently being used to aggregate the LES circuits - a lot of work is being done throughout the network including the Core being revised -

New Member

Re: OSPF - connecting multiple remote sites via LES

Good practice would be for layer 3 to match what's truly happening at the lower layers. If each circuit to the 15 locations is a single point-to-point link, then layer 3 should reflect that with the multiple 30-bit subnets. If those circuits are aggregated into a point-to-multipoint mesh, then layer 3 should reflect that, too with the single larger subnet. If you're concerned about propagating all those "WAN networks," you can always summarize them. Don't confuse the actual addressing scheme with how the routers will pass the routes. Routes can be rolled up or separated out, and will be redistributed and such. The actual networks underlying those routes won't change (much).

New Member

Re: OSPF - connecting multiple remote sites via LES

Thanks for your guidence -

What you say makes sense. In addition to your remarks, one thing that confused me was the fact that at both ends the LES circuits are connected via ethernet (as you would expect), this by defalt is a broadcast medimum and the devices set up relationships in relation to this - for example DR and BDR. There is probably nothing wrong with this or does it create overhead that normally isn't there on a normal Point to point link? Another thing I noticed at the location were I have come across this is that all of attached ports on the aggregate switch has become DR for that particular link IE: it is DR x 10 for all 10 seperate subnets!

What are your thoughts on this?