Managing a Massive Voice Infrastructure with Cisco Prime Collaboration
Cisco IT monitors and manages a huge voice infrastructure, with over 200,000 UC endpoints, and the Cisco Prime Collaboration solution helps us do this work efficiently.
For example, a common problem for my team is identifying which devices are provisioned in the Cisco Unified Communications Manager (Cisco UCM), but are no longer in use. This issue is getting more complex as Cisco employees have multiple devices associated with their one directory number. In a typical case, a salesperson might have a desk phone and a Cisco TelePresence personal video endpoint in the office, another phone in their home office, and use Cisco Jabber clients on a laptop and smartphone at home, at customer sites, or while traveling. Cisco Prime Collaboration lets me easily view this information and verify that the employee is actively using all of these devices.
Hardware phones in particular can become inactive when an employee leaves or transfers and no one else moves to that desk. Cisco Prime Collaboration lets me easily identify and remove that phone. We can also detect which employees haven’t downloaded the latest Jabber client version and encourage them to update their devices to the currently supported software.
Cisco Prime Collaboration gives me a very easy graphical interface to see into the whole global network, and then allows me to drill down to any components to see what’s going on.
Figure 1: Sample CPC Network Topology, enabling drilldown on each location and device
We had an issue a few weeks back with a lot of dropped calls between our contact center and a partner contact center. We moved the calls to backup paths as we sorted out what was going on with our pair of SIP trunks connecting us to the partner. A quick look at the Prime map showed 100% utilization on one trunk, and 0% utilization on the other, and as call loads increased the first trunk filled up and dropped calls, which made it very easy to determine the root cause of the dropped calls. Turns out that we had configured the SIP trunks to load balance all voice across both trunks, and the partner had configured the second trunk as a failover trunk. We worked with the partner to synchronize our SIP trunk configurations and fixed the problem. Cisco Prime Collaboration saved us a good afternoon of sorting through dropped packets to find the underlying problem.
Figure 2: CPC monitors areas of concern, thresholds, alarms, and overall network health
We have implemented two instances of the Cisco Prime Collaboration solution for voice and video monitoring; each instance is implemented on Cisco UCS servers. One instance is primary and installed in our Allen, Texas data center. The second instance is installed in our Amsterdam data center for disaster recovery purposes. The Cisco IT network operations team uses a separate Cisco Prime Infrastructure instance for monitoring and managing the circuits and elements in our network infrastructure.
We have deployed two instances of CPC (primary and backup) on two UCS servers because of the size of Cisco’s network and our commitment to delivering very high levels of voice and video service availability to our users around the world. For most Cisco customers, the high availability levels delivered by the Cisco UCS server will be adequate, so multiple Cisco Prime Collaboration instances won’t be necessary. Even so, these two CPC instances have replaced five sets of Cisco UOM, USM, UPM, and USSM servers, 25 servers in all, in five major clusters in our global network.
What impresses me about Cisco Prime Collaboration is that we’re able to use just this one solution to manage one of the largest enterprise voice infrastructures in the world. And because we discuss all of our management needs and lessons learned with the product development teams, customers benefit directly from our experience.