Six ways to troubleshoot voice quality monitoring

VQM works by using both active (synthetic) and passive (live call/session) testing and analysis. As with any other application data, network congestion can delay voice packets or even cause some to be lost along the way.  While this is just a minor inconvenience for most data applications – e.g., an email taking slightly longer to process - even a few lost RTP (media) packets can introduce unacceptable echo or disruption in a conversation. VQM solves these problems by analysing live/in-service sessions, taking detailed notes of session state, start and end times, called party and configuration information as well as a robust suite of underlying service-affecting fault details. 

In addition, synthetic voice traffic generation efforts ensure that service-affecting issues are identified during both on and off-hours, constantly monitoring infrastructure state and reporting on quality. This approach ensures that any underlying issues are caught by the VQM tool, before end-users even notice service degradation, making calls uninterrupted and communication much easier.

There are effectively two schools of thought when it comes to voice quality monitoring, either active monitoring or passive analysis. Active monitoring is application monitoring which occurs by sending synthetic voice traffic back and forth between software agents and reporting on its quality. Think of this as the ‘secret shopper’ model. It’s imperative that these agents not only send real media (RTP) but also support registration and signalling to allow them to effectively become valid addressable endpoints on your network. They can then exercise not just the media plane, but also signalling infrastructure to provide a complete view into back-end performance and reliability. This enables organizations to become aware of the issues before the end-user reports them.

Passive analysis describes the process of analysing live traffic or in-service calls. Using this model, VQM is able to effectively monitor both the signalling and media planes to check for issues with any sessions traversing the enterprise network, identify any common threads and ideally alert operations personnel immediately. This lets them effectively address identified issues to minimize impact to the business.

Clearly, both approaches are complementary and provide value, ensuring that any issues that arise are identified immediately. According to research from Gartner, 86% of organisations are conducting virtual interviews to hire candidates during the coronavirus pandemic. Connection and call quality issues can be a significant problem when it comes to candidates being able to perform at their best under pressure. 

To help IT and network managers improve communications on video conferencing software across their company, here are six ways to troubleshoot any problems with voice or video quality in order of operation that can applied to either active or passive monitoring:

1. Check session MOS scores – these are typically reported both per session as well as per direction. When scores are significantly degraded, examine calls placed from the location in question over a specific time period to quickly identify the common degradation factor.
2. Further correlate session degradation factors to determine where to head next. Degradation sources can be used to effectively isolate session impairments to help guide troubleshooting efforts.
3. Using the results from step two, dig deeper into the tool to extract per-session & per-direction statistics to gain a better understanding of the underlying issues as perceived by the measurement point/end-user. SQmediator provides expert analysis on a per-call basis to assist with effectively interpreting low-level diagnostics.
4. As a final step, compare UC sessions that traversed similar paths during the same time period to confirm whether other sessions experienced similar symptoms and performance.
5. Using these results, take corrective actions to rectify network and application performance.
6. Continue to monitor incoming performance data points, both active as well as passive, to verify the identified issue has been fixed.

Network architects and managers need to address call quality and performance management problems during the planning and deployment phases, but they should be aware that these problems also occur regularly during normal day-to-day network operation post-deployment. Many quality-related unified communications problems tend to be short-lived, temporary in nature, and can occur anywhere along the network path. For example, a user accessing a file from a remote server or a home-based worker watching a YouTube video or downloading large files from the server may cause a temporary and brief bottleneck.

This can sometimes cause short-term degradation in call quality for other users on the network without them realising. It is therefore important that network managers use performance management tools, that are able to detect and measure these types of network impairments and alert them if there are any issues or network downtime.

Network managers could attempt to use techniques such as packet loss and jitter metrics to estimate call quality, but this approach is reactive, it doesn’t actually correlate to the end-user experience and doesn't provide enough diagnostic information to be able to determine the cause of the problem. That’s why the six steps mentioned above are as a useful place to start when trying to troubleshoot problems with voice quality on video calls.  Whilst no online software is perfect, there’s still work to do to make communications seamless on these kinds of platforms in the future.