Server Fault is a question and answer site for system and network administrators. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I work for a school district with the following configuration:

Two remote sites connect to a main site via T-1s, then the main site has 3 T-1s which provide Internet access, advanced content filtering and hosts our CallManager.

We are adding bandwidth via cable modem to each site and keeping our T-1s in place, primarily for VoIP , but there are also some services (web, file shares, etc.) hosted at the main site.

How can I accomplish the following:

  1. allow certain subnets (for teachers, administrators, secretaries) at the remote sites to use the faster Internet connections
  2. the student and voip subnets at the remote sites should continue to use the T-1s
  3. if the cable connection goes down, the teacher/admin subnet should failover to the T-1s.

At first I was thinking I could use a static IP route with acls to deny the student subnet to use the cable connection, but I think that cause the student network to fail.

share|improve this question
We need a lot more details to figure this out. Are you running NAT on anything? How do you plan to be able to determine when the 'cable modem' goes down? Static ip route with acls would be an extremely kludgy way of doing it. Why does it matter which particular connection your traffic uses, generally speaking? – Aaron Aug 3 '11 at 15:57

We have similar config in Juniper, not sure if cisco allows that:

  1. Source based routing basically, so you configure what networks use which interface or gateway as its route.

  2. Interface monitoring via ping(multiple pings to multiple outside gateways with different threshold and weight), if the interface goes down, source based static route is automatically taken out and default route is used.

share|improve this answer
I assume these internet connections have different IP address ranges. There needs to be some NAT setup also to make sure that a packet exiting through one connection will have a source address belonging to that connection. This is not simple to setup. (A proper way would be to have provider independent IP space and to run BGP with both providers.) – snap Aug 3 '11 at 5:26
Well Original questions mentions that these are separate subnets. – R D Aug 3 '11 at 14:32

Sounds like a typical example where you'd want to use Policy Routing. You'll need a Policy rule that sets the next hop address and (ideally) enough routing protocols running that you can use (say) the loopback of the remote router as "next hop". If the next hop you're trying to set is unreachable, Policy routing will fall back to the normal routing table.

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.