1.2 Scalability Considerations

Consider the following before installing the AppManager for Avaya Communication Manager module:

  • For CDR-processing tasks for the AvayaCM_Call* Knowledge Scripts, the primary scalability factor is the number of active phones you are monitoring. Active phones are configured phones that are off the hook and in use. Active phones generate call detail records (CDRs), which increase CPU usage on the proxy agent computer. Inactive phones, those that are configured but not in use, do not increase CPU usage on the proxy agent computer and therefore are not a scalability consideration.

    One proxy agent computer can support the monitoring of approximately 21,000 active phones in one cluster or across multiple clusters. It is possible to monitor more active phones by increasing the interval in which the Communication Manager server sends RTCP packets to the managed object on the proxy computer. Increasing the interval can reduce the strain on the proxy agent computer and allow it to monitor more active phones.

  • For SNMP-related tasks for the remaining AvayaCM Knowledge Scripts, one proxy agent computer can support up to five Communication Manager clusters. The efficiency of the SNMP code in these Knowledge Scripts changes only slightly based on the number of phones you are monitoring. From an SNMP perspective, there is little difference in monitoring ten phones or 10,000 phones.

  • For the PhoneQuality Knowledge Script, the primary scalability factor is the number of data points the script generates. The script generates seven data points per monitored active phone every 30 seconds, which is a large amount of data when multiplied by thousands of monitored phones. Although the proxy agent computer can handle that much data, the AppManager repository (QDB) and management server may eventually be unable to handle the volume of data.

    You can decrease the flow of data by turning off some data streams and by changing the value of the Data collection interval for voice quality metrics parameter to 1 minute.

    The management server can handle a larger volume of data if you increase the size of the map queue. The default size is 25 MB, but you can increase it to 250 MB with no adverse affect on performance. An even larger value may work in your environment.

    To change the size of the map queue:

    1. On the management server computer, open the Registry Editor and navigate to HKEY_LOCAL_MACHINE\SOFTWARE\NetIQ\AppManager\4.0\NetIQms\Config.

    2. In the right pane, double-click PIOC Data Map File Size MB.

    3. In the Edit DWORD Value dialog box, change the value in the Value data field.

    4. Restart the NetIQ AppManager Management Service, NetIQms.exe. The change in file size does not take effect until you restart the service.

  • Based on the preceding scalability considerations, your proxy agent computer should meet the following minimum hardware requirements:

    • CPU: 2 HT

    • Speed: 3.20 GHz

    • RAM: 8 GB