T1 failover configuration

Unanswered Question
Mar 16th, 2007

I cuurently have 2 T1 circuits and would like to have a single host only use 1 of these T1's and all other host on the second T1. In case of a failover i would like to have it failover to the other T1. Is this something that is possible to complete on a pair of Cisco 1841 routers?

I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Loading.
Richard Burts Fri, 03/16/2007 - 12:27

Brian

It should certainly be possible to have a pair of 1841 routers implement the logic that most devices use one of the T1s and a single other device use the other T1, and if the first T1 fails to have traffic fail over to the other router.

First you need to set up routing logic so that one T1 is preferred and the second T1 would be used if the first T1 fails. How to do that depends on whether you are using a dynamic routing protocol or are using static routes. Once you get that set up then you configure Policy Based Routing to select traffic from the specific host and to send it over the second T1. PBR uses route maps to change the routing logic for the selected traffic and uses access lists to identify the traffic. 1841s are certainly capable of doing this.

HTH

Rick

bhansen70 Fri, 03/16/2007 - 13:27

Rick

Thanks for replying to my question. What my initial thought would be is to just use default routes with different metrics based on what T1 i want as my primary. With that said it was just brought to my attention to be a little more complicated than that now. The new proposed solution from my EXECS is to have Circuit 1 dedicated as the primary to a single host but if that circuit fails to use Circuit 2. Then Circuit 2 will be dedicated to all traffic not coming from the single host but if this circuit fails to use Circuit 1. These circuits are point 2 points between HQ and branch office. I am thinking that by using poicy based routing and default route metrics that i can still get this accomplished. Does this sound correct?

The routers will be running EIGRP with the exception of the default routes.

Richard Burts Fri, 03/16/2007 - 14:29

Brian

Yes that sounds correct. In fact that is pretty much what I thought I was describing in my first response and what I thought you were asking for in your original post.

If you are running EIGRP then you want to be sure that EIGRP is running over both circuit 1 and circuit 2. If circuit 2 is to be primary for most traffic you want it to have the better metric. The easy way is to increase the delay metric on circuit 1 (you will need to do it on both ends of the circuit). When you do this EIGRP will use circuit 2 as preferred route but will fail over to circuit 1 if circuit 2 fails.

After you get the routing working the way that you want you can configure Policy Based Routing. Configure an access list which will permit traffic to or from the specific host. Then configure a route map which will use the access list to identify the traffic and will use a set (set interface or set next-hop) to direct traffic over circuit 1. The nice thing about PBR is that if circuit 1 fails and PBR can not use it, then the traffic from the specific host will use normal routing and go over circuit 2.

HTH

Rick

Actions

This Discussion