EIGRP routing

[hillolnath]

Scenario:see attached diagram

I

have two catalyst switches at location A and location B connected via 100 Mbps ethernet ( directly terminating in switch) through service provider.I have also MPLS link from both location A and location B.

I require routing to be done at location A switch in such a way that my preferred path to location C should be through MPLS link and preferred path to location B only through 100 Mbps ethernet link.

However on failure of either MPLS link or Ethernet, i should be able to reach both location B and C from location A.

How do i define the EIGRP routing and where. what would be the exact commands.

Would appreciate your help on best possible way to do this.

[Richard Burts]

Hillol

You do not provide enough information for us to be able to give you specific commands to configure EIGRP to meet your requirements. But I offer these suggestions as a way to get you started.

- It is not clear from your diagram whether EIGRP route summarization will be an issue. I suspect that you might be better to turn it off.

- there must be at least one VLAN (and therefore subnet) connecting CatalystA and CatalystB.

- there must be a network statement in EIGRP on both switches that includes this subnet. This ensures that CatalystA and CatalystB will become EIGRP neighbors.

- there should be network statements in EIGRP on each Catalyst for the subnets where end stations are connected that need network connectivity.

- CatalystA must have a network statement in EIGRP that includes the subnet that connects CatalystA to the router (and MPLS) to which it is connected.

- CatalystB must have a network statement in EIGRP that includes the subnet of at least one of the routers (and MPLS) to which it is connected (and including both routers would be better).

- CatalystA should learn the subnets that reside at CatalystB both from CatalystB directly and also from the MPLS. You must assure that the metric learned from CatalystB for these subnets is more favorable than the metric learned from MPLS. It probably will be, but if it is not I would suggest configuring an EIGRP offset list to increase the metric learned from MPLS for these subnets.

- CatalystB should learn the subnets that reside at CatalystA both from CatalystA directly and also from the MPLS. You must assure that the metric learned from CatalystA for these subnets is more favorable than the metric learned from MPLS. It probably will be, but if it is not I would suggest configuring an EIGRP offset list to increase the metric learned from MPLS for these subnets.

- CatalystA should learn the subnets that reside at locationC both from CatalystB directly and also from the MPLS. You must assure that the metric learned from CatalystB for these subnets is higher than the metric learned from MPLS. It may or may not be that way naturally. I would suggest configuring an EIGRP offset list to increase the metric learned from MPLS for these subnets.

- CatalystB should learn the subnets that reside at locationC both from CatalystA directly and also from the MPLS. You must assure that the metric learned from CatalystA for these subnets is higher than the metric learned from MPLS. It may or may not be that way naturally. I would suggest configuring an EIGRP offset list to increase the metric learned from MPLS for these subnets.

If you configure EIGRP according to these suggestions I believe that you will satisfy your requirements.

HTH

Rick