Failover Design Help

bregimand · ‎02-21-2007

I have attached a diagram of a proposed network.

I need to route all VoIP traffic through R1, which is no problem.

However, the client wants a failover solution so that if the serial link on R1 goes down, voice and data will fail over to R2. Conversely, if S0 goes down on R2, data will go to R1 (voice already is). I'm sure that this would be possible with BGP, but that's not an option in this situation.

Does anyone have any suggestions on how to do this otherwise?

Thanks in advance,

Ben

jheckart · ‎02-21-2007

What's doing your call signaling here? Are either of these routers CCME, or do you have CCM somewhere? Are you using a routing protocol? Do you have a PRI/POTS lines somewhere? In the event of a failure, do you only need the phones to be able to call each other, or do the need to get out to the PSTN?

The actual routing should be simple, but if the phones cannot register with something on the side of R2, it's not going to work.

bregimand · ‎02-21-2007

It's not call manager - it's all skinny back to my core through either connection to an asterisk system. Both serials are PPP. I can run any other protocol necessary between the 2 routers.

There are no POTS, and, yes, they need to register and make calls to the PSTN.

They can register either through R1 or R2.

Thanks,

Ben

jheckart · ‎02-21-2007

What you can do is run HSRP between the two routers for the voice VLAN and a second for the data VLAN. You'll want to make the priority higher on R1 so that the voice/data traffic will go through it under normal circumstances. In the event of R1 failing, voice/data traffic will go through R2 automatically.

HSRP supports tracking, which you could configure on each router under the standby configuration to fail in the event of the serial interface going down.

http://www.cisco.com/en/US/products/ps6350/products_configuration_guide_chapter09186a008042fbb3.html

Is this what you were looking for?

bregimand · ‎02-21-2007

It looks.... similar. I'm going to try to set this up in my lab and see what happens. I'll let you know.

Thanks for your help,

Ben