Showing results for 
Search instead for 
Did you mean: 
Walkthrough Wednesdays

CVP Migration Strategy to Parent Child Model


It’s a multi-step process, depending on the current state of the system.



What is Parent Child Deployment Model

Parent/Child implies that your “child” is either an existing UCCE “System” model using the System PG or will need to migrate to that model first. Since we only started supporting CVP at the child or System CCE level in 7.5, it’s not likely this existing UCCE with CVP is already in the correct deployment model to support parent/child anyway.

Re-Building ICM

If the current UCCE is a “classic” or standard UCCE with CVP, it will have to be re-deployed to migrate the users to a System PG model to allow the Parent ICM to connect to it. Of course, the Parent ICM will also have to be built along with its CVP at the Parent level.


I’d build the new infrastructure out – new Parent ICM, Parent CVP and look at the existing UCCE – it may be possible to add a System PG to that existing deployment, and start moving agents over to that PG as part of the migration. This would also allow for end-to-end testing with the new parent ICM to prove out the call flow models before moving the inbound numbers to the Parent ICM and moving the agents over to the System PG.


The concept of parent/child normally implies that the parent will be doing all the call routing and queuing, so any logic about open/close hours, agent selection, etc., should be done in the parent ICM routing scripts. This would take some work to replicate those scripts from the existing UCCE, but at least the logic is already defined and can be just re-written in the ICM layer.


The child layer should be looked at to simply deliver the call from the parent to the agent locally, again, all of the complex call routing logic should have moved to the parent, the child should just deliver the call to an available agent. If the agent isn’t available, even the RONA scripting should be handled by the parent ICM using the CVP “router re-query” function and leave the child out of the processing.


No simple answer, it can be done and on some level is similar to what our legacy ICM customers have to deal with when putting ICM on top of their legacy ACD platforms. The additional complexity here is the parent/child connection and requirement for using the System PG at the child in order to connect the two layers to get the parent ICM agent state information for intelligent routing.

Content for Community-Ad