cme call routing

Answered Question
May 2nd, 2010

Dear expert...

How do you direct the voice traffic (both: voip and pots) to go thru certain interface, like PRI link?

for example:

RouterA---------------RouterB---------------RouterC

                                      | (main link - PRI)     | (backup link)

                                      |                                   |

                                RouterD---------------RouterE

If RouterA would like to call RouterE, how do you direct the call to using the "main link - PRI link" on RouterB >> RouterD >> RouterE? instead of going to RouterB >> RouterC >> RouterE as RouterC is set only for "backup link" incase if the "main link - PRI link" is down.

Please advice

Thanks in advance

I have this problem too.
0 votes
Correct Answer by David Hailey about 6 years 7 months ago

Terrence,

You could probably come up with a 1000 scenarios and we could figure

out why or why each will or will not work. So, what is your actual

topology and what do you plan to do with your routing? Let's get a

real scenario and knock this thread out.

Hailey

Sent from my iPhone

On May 4, 2010, at 12:07 AM, vip80y_cisco

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 4 (6 ratings)
Loading.
William Bell Mon, 05/03/2010 - 04:10

If you are using CME on all routers and you have no gatekeeper in the mix then you would need to string the call along, hop-by-hop, using dial peers.

Let's assume the following

1. RouterD-to-RouterE: PRI

2. RouterB: IP address 10.10.10.10

3. RouterC: IP address 11.11.11.11

3. RouterE: Pattern 2xxx

4. h323 between cme's

RouterA Config:

dial-peer voice 10 voip

destination-pattern 2...

session target ipv4:10.10.10.10

<...other stuff for voip dial-peers...>

preference 1

!

dial-peer voice 11 voip

destination-pattern 2...

session target ipv4:11.11.11.11

<...other stuff for voip dial-peers...>

preference 2

!

RouterB:

!

dial-peer voice 10 pots

destination-pattern 2...

<...pots dial peer stuff...>

port 0/1/0:23 !for example

!

RouterD:

!

dial-peer voice 10 pots

destination-pattern 2...

<...pots dial peer stuff...>

port 0/1/1:23

It gets kinda hairy the more you expand this network and it may be prudent that you consider introducing a gatekeeper architecture into this design. It may also be helpful dealing with failover scenarios and will definitely make dial-plan maintenance easier.

HTH.

Regards,

Bill

Please remember to rate helpful posts.

vip80y_cisco Mon, 05/03/2010 - 19:44

Hi William,

Thanks for your reply, I pretty much understand your explanation and yes we dont have gatekeeper.

so what happens if RotuerD got 2 interfaces connected to each router (like the topology below) which IP address should I use for the "session target ipv4" ? Is it 10.0.0.6 or 10.0.0.7 ?

RouterA--------------------RouterB------------------RouterC-|

      10.0.0.1           10.0.0.2  |   10.0.0.3         10.0.0.4             | 10.0.0.9   

                                             |                                                   |

                                             | (pri link - cloud)                      |(back up link)

                                             |                                                   |

RouterF-------------------RouterD-------------------RouterE-| 10.0.0.10

      10.0.0.5             10.0.0.6      10.0.0.7            10.0.0.8

Thanks again,

Terrence

David Hailey Mon, 05/03/2010 - 19:52

Bill, correct me if I'm wrong - but I believe in this case, the link would be 10.0.0.6 as this points (at least it appears to) to the PRI link whereas the 10.0.0.7 is the link between Router D and Router E.

Hailey

William Bell Mon, 05/03/2010 - 20:04

Well, it depends on which interface you wish to bind your call control and media channels to. For h323 (as an example):

interface

!

Given this topology, I'd recommend creating a loopback interface and using that as the bound interface. This way a loss of link 10.0.0.6 OR 10.0.0.7 wouldn't have a negative effect on the non-impacted link. Of course, you would have to make sure that the other routers had either a static route to the loopback address or that you had a dynamic routing protocol exchanging routes.

HTH.

Regards,

Bill

Please remember to rate helpful posts.

vip80y_cisco Mon, 05/03/2010 - 21:07

Hi Guys,

Thank you for your advice...

How about... say If I apply an eigrp / ospf as the routing protocol?

So if one of those routing protocol in place, 10.0.0.6 OR 10.0.0.7  wouldn't be an issue right? meaning I can put either 10.0.0.6 OR 10.0.0.7 for the "session target"

Cheers,

Terrence

Correct Answer
David Hailey Mon, 05/03/2010 - 21:59

Terrence,

You could probably come up with a 1000 scenarios and we could figure

out why or why each will or will not work. So, what is your actual

topology and what do you plan to do with your routing? Let's get a

real scenario and knock this thread out.

Hailey

Sent from my iPhone

On May 4, 2010, at 12:07 AM, vip80y_cisco

vip80y_cisco Mon, 05/03/2010 - 22:32

Hi Dave,

I am really sorry..... I just try to complete my lab one step at the time

will open a new thread then..

Cheers

David Hailey Mon, 05/03/2010 - 22:41

Terrence,

Don't get the wrong idea - if it's for a lab, that's cool.  You just need to let us know upfront.  Keep in mind that we (all of the top engineers who post) do this because we WANT to...we certainly don't get paid for our efforts here.  However, it's not much of an incentive to continue going on an endless thread when we give solid answers and receive 3 star ratings for it.  Not all answers are 5's so that's cool.  I'm just saying that we have full-time jobs and we do this because we enjoy helping others..but if you're prepping for a lab, it certainly seems like there is no end in site to the number of scenarious you might come up with.  There is an Ask the Expert - CCIE Voice thread, you may want to hop on that for a bit and see if you come across something of use to you.

Hailey

Please rate helpful posts!

Actions

This Discussion