Additional ITSP in CUCM/VXML IOS

danclaw · ‎03-07-2018

First off, thanks to the community!

Is it possible to add a second ITSP using SIP (not backup) to a Cisco PCCE environment?

My Network:

What needs to be done:

Company moving from one SIP provider (ITSP1) to new SIP provider (ITSP2).
Network admin wants to test the connection to ITSP2 before transitioning to ITSP2.
There is no test environment.

Previous state:

Single ITSP to VXML2.
VXML1 and VXML2 used to be in separate physical locations. ITSP1 provided two SIP terminations (one to each location).
VXML1 and VXML2 were moved into a single physical location. ITSP1 then only provided one network connection for the SIP trunk.

Current state:

ITSP1 is connected through VXML2.
ITSP2 has been connected to a port on VXML1
A static NAT (and firewall) has been made through from the public-facing IP to the internal interface for ITSP2 on VXML
ITSP2 can get a SIP Options response from the interface on VXML1.

What needs to be done:

Register phone with ITSP2
receive/make calls with ITSP2

What should I look to do next? Are there topics I should read up on? Does anyone know of a guide to add a 2nd ITSP (not backup) for use in an environment?

Thank you,

-Daniel

R0g22 · ‎03-07-2018

I am expecting that you would obviously have a CUBE at the edge since your topology does not highlight one.
These VXML1 and VXML2 have individual SIP trunks in a active/active fashion load balancing ?

danclaw · ‎03-08-2018

Thank you for responding.

I believe our VXML routers are the CUBE routers. They have the DIAL PEER configurations in the configs.

CUCM has the VXML/CUBE routers in separate route groups. I believe this means they are NOT active/active.

I'll post part of my response to another responder in this thread:

"I've made a few changes in CUCM:

Created a new device pool under the existing VXML01 trunk

Created a new Route Group under the VXML01 trunk

Put a physical phone profile into that device pool. THe phone boots, registers. However, it calls out, and receives calls, from the old pool.

Where should I be looking?"

Thank you,

-Daniel

Andrew West · ‎03-07-2018

If you have your SIP trunks coming in via CUBEs then its easy to route to whatever via dialpeers.

RIght now I have two different PCCE environments running which means two different versions of CVP. All easy to account for via dialpeers on CUBE.

At a higher level I have two different SIP trunks hitting my CUBEs with one set heading to PCCE and one to a different CUCM cluster.

The key thing is as long as you hit CUBE first what you want to do is not hard at all.

danclaw · ‎03-08-2018

I believe the VXML routers (Cisco 3900 series devices running IOS 15.4) are the CUBE routers. They have the dial peers in the configuration.

I've attached redacted versions of the configuration files. The redactions are an attempt to hide SSH keys/password hashes, and to obfuscate the identity of my employer.

Changes I've made (Note, these are the changes and/or additions).

Note: The ITSP provider is doing both signaling and stream via one channel. the IP is 68.68.117.116
They have provided a test number that starts with a different area code than the blocks I control. I control two four-digit blocks, and several toll free numbers.

class-map match-any VoIP-Ctrl+DNS+SIP
 match access-group name ITSP2-DNS
 match access-group name ITSP2-SIP
class-map match-any VoIP-RTP
 match protocol rtp audio
!
dial-peer voice 10001 voip
 description ***Inbound from ITSP2 PSTN SIP***
 incoming called-number 720.......
 no vad
!
dial-peer voice 11001 voip
 description ***Outbound to ITSP2 PSTN SIP; US/N. America Call***
 session target ipv4:68.68.117.116
 
sip-ua
 sip-server ipv4:68.68.117.116
 
dial-peer voice 11003 voip
 description ***Outbound to ITSP1 PSTN SIP; 911 Call***
 session target ipv4:68.68.117.116

dial-peer voice 11004 voip
 description ***Outbound to ITSP1 PSTN SIP; 911 Call***
 session target ipv4:68.68.117.116

I've made a few changes in CUCM:

Created a new device pool under the existing VXML01 trunk

Created a new Route Group under the VXML01 trunk

Put a physical phone profile into that device pool. THe phone boots, registers. However, it calls out, and receives calls, from the old pool.

Where should I be looking?

R0g22 · ‎03-08-2018

When you say it uses old pool ? How do you know it is using the old pool ? Do you use SLRG ?

danclaw · ‎03-08-2018

When I say "Old Pool", I mean that when I make a call from the IP phone, calling a 3rd party phone number (Verizon cell), there are no active calls listed on VXML1, as indicated by 0 results on VXML1 using the "show call active voice compact" command.

I cannot see whether my individual call was going out through VXML2, as there are always calls going through on VXML2.

Looking at my CUCM configuration, it doesn't appear that we are using SLRG.

Attached are screencaps from my CUCM configuration. I have made attempts to pipe the new SIP from ITSP2 to the IP phone.

Thank you,

-Daniel Clawson

R0g22 · ‎03-08-2018

Do a DNA on CUCM to see which RP is being used ? Or if you have descriptions added properly you can check the RP that you anticipate would be used for these call and check the associated RL and RG.

danclaw · ‎03-19-2018

Thank you for responding. I've been OOO for several days.

Assumptions from your last post:
RP = Route Pattern
RL = Route List
RG = Route Group

We only have one RP for normal 10-digit dialing. It points to one Partition, and one route list.

The DNA analysis shows:

The Tranalation Pattern strips off the leading "7", going from "712125559876"to "12125559876"
The Route Pattern adds the leading "9", resulting in "912125559876"
The Route List selects the PSTN-RL list, indicated above.
In PSTN-RL are RGs: "500-SIP", "NEST-SIP", and "Standard Local Route Group"

Does this provide the information you wanted?

Thank you,

-Daniel Clawson