Warakorn - Fotolia
The advent of video conferencing, adoption of softphones and cloud-based services are creating dramatic and exciting developments in unified communications technologies and capabilities. Products such as Microsoft Skype for Business and Cisco Spark are disrupting the industry with easy-to-use, all-in-one UC applications that allow users to communicate however and wherever they want.
While these products can help users' productivity levels, the need to devote more resources and time to UC monitoring is catching many IT organizations by surprise. As a result, IT groups struggle to achieve a high level of visibility into the use and performance of these complex and interdependent UC applications.
IT can overcome this challenge in two key steps: Use a unified performance management platform instead of point-monitoring products, and break down silos that prevent collaboration among IT teams. Failure to take these steps forces IT to spend more money to maintain UC performance for all users and exposes the business to risk.
UC monitoring tools help employees adopt apps
Robin Gareisspresident and founder of Nemertes Research
According to Robin Gareiss, president and founder of Nemertes Research in Mokena, Ill., "Companies devote 33% more IT staff to managing IP telephony and 31% more to UC when they don't use monitoring tools. The tools are instrumental to identifying, isolating and resolving performance issues -- and preventing them from happening again. When apps run efficiently, employees are more likely to use them." That's one reason, according to Nemertes, why 36% more people who use UC in large companies use monitoring tools.
Monitoring media quality and UC performance in a way that identifies and remediates problems as soon as they arise is challenging. IT organizations often miss vital information, and tickets go unresolved or are passed back and forth between support, network and telecom teams.
Two factors are to blame for these problems: the barriers that separate the network, telecom and desktop support teams, and a UC monitoring system comprised of individual point products from multiple vendors that do not communicate with one another. As a result, none of the teams have a complete view into how the network, infrastructure and endpoints are operating.
Take a holistic approach to detect network problems
IT traditionally tries UC monitoring from one of two perspectives -- the telephony endpoint or the network. The problem with the former is you only have visibility into the call manager servers and endpoint devices, such as hard and soft telephones and video systems. So, you know how a problem affects the end-user experience, but you're unable to examine the entire network to determine the cause of problems.
The latter approach, examining the network, requires a wide distribution of probes to monitor network traffic. VoIP and video systems are peer to peer, so you must connect them branch to branch, and should install monitoring in all branches for complete coverage. For some customers with many branch offices, that approach may quickly become cost-prohibitive.
Using either approach alone hampers IT's ability to identify and address the specific cause of a problem. The problem might be with a user's headset, an error the user doesn't realize he made, or a bandwidth bottleneck on the network or infrastructure.
A unified, four-pronged approach to UC monitoring
The ideal resolution should combine the following four approaches into a single strategy:
1. Endpoint telemetry. Most UC vendors instrument their endpoints and provide media reports back to a central collection server. These reports include performance metrics -- such as jitter, packet loss and mean opinion score -- about the inbound media stream received by that endpoint. Typically, these reports are available at the end of a call and incorporated into a call detail record.
2. Network probes. Probes are high-performing, packet-capture appliances deployed strategically throughout the network. Probes process media packets, produce quality metrics for a particular media stream and roll up to summary-level key performance indicators for their assigned location.
3. NetFlow. NetFlow, version 5 or 9, reports summary information about a flow -- communication between source and destination IP and ports -- at a particular location, including bytes transferred and associated DSCP markings. When combined with deep packet inspection, richer metrics may be available using the templating capabilities available in NetFlow version 9. This information is also helpful in providing the media call path.
4. SNMP. Simple Network Management Protocol (SNMP) helps monitor the overall health and performance of nodes in your network, such as routers, switches and gateways. When devices are polled using SNMP, they provide a number of metrics regarding the current state. When applied to troubleshooting VoIP, it can provide essential forensic data when the root cause turns out to be interface errors, CPU or other resource deprivation on the device.
Only by adopting a comprehensive, multipoint approach that leverages all technologies will IT be able to detect and resolve all media quality problems, regardless of whether they are caused by the network, device or user. Also, by enabling decisions based on a common set of metrics from a unified performance management platform, you can break down interdepartmental barriers and converge previously siloed network, telecom and desktop support teams.
David Roberts is the director of product management for unified communications at Riverbed Technology.
UC analytics, metrics must mature before benefits are seen
Complex apps need network-based performance management tools
UC management gives enterprises cloud controls
Dig Deeper on Unified Communications Architecture and Service Models