CVP report vs. Telco report

Unanswered Question
Sep 22nd, 2009
User Badges:


I have a question with CVP reporting:


We have 4 CVP servers being load balance on the CSS. One pair located in Region 1 (side A), and a pair located Region 2 (sideB). Both locations have ingress gateways. Lately, we've been having problems with performance with our Nuance servers therefore we have to allocate DTMF and ASR between locations (ie. 50% to 50%, or 100% DTMF side A, and 100% ASR side B). Presently, we also have a problem with our Service level agreement going down because of the large number of call going in. Here are the facts: Calls are normal volume, we have more than enough agents, and there are no abnormality reported by the agents. In addition, we compare the Telco report numbers (calls) and the calls report from CVP.

Is there a complication with our frequent allocation of DTMF and ASR between regions that can cause the numbers of call to be askew? The report from the CVP is higher by 800 calls, by the average, compared to TELCO.

The CVP pool is behind a VIP (virtual IP). When we switch from region to region, those calls in the middle of transaction will have to request or reconnect again? Could it be counted twice? Can there be a possibility of phantom calls?

There may be a number of variables missing from my post. Your opinion is valuable to me.

Thanks.

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Loading.

What statistic are you using to count the CVP calls? Is it an ICM call type report?


If you are using statistics off the CVPs you may be counting the switch leg and the VRU leg as two calls - when the Telco says there is one.


The call type report would be accurate.


Regards,

Geoff

jeanpaul.delmun... Wed, 09/23/2009 - 05:34
User Badges:

Thanks for your response. I will check our reporting server configuration. I'm almost certain we are collecting the metrics from the IVR (CVP).

jeanpaul.delmun... Wed, 09/23/2009 - 12:12
User Badges:

Geoff:


This is the response I got from the vendor who implemented our system.


There can be more than two legs reported for a call in the database. Physically you have the VRU leg and the switch leg but in the database you will find that there is the switch leg, the VRU leg, a Call Manager leg when the call is connected to the desk phone, and additional legs for any transfer and consultation.

However, even if we counted the switch leg and VRU leg as the same call, the percentages would remain the same - correct? 25% answered within the 60 seconds interval would be true regardless if we were counting one call or two calls.

We don't do that though. We are using the Termination_Call_Detail records and filtering for a distinct call record ID. This record is shared across all legs of the call so pulling only one ensures that we only pull information for one leg of the call.

We also pull from the half hour tables to show us the stats for the logged in agents, time worked and individual calls supplied to their phones, etc.


Is the Call record ID the same as the Correlation ID passed to the VRU? He did not respond to that question. I'm also making an assumption that we are using VRU type 10 for our latest and greatest IPCC Enterprise environment. I will double check. I'm still learning this system(s).

Actions

This Discussion