Seasonality is an important factor for network and service assurance. It’s also known as time-of-day/week/month/year specific activity.
For example, we often monitor network health through the analysis of performance metrics (eg CPU utilisation) and set up thresholds to alert us if those metrics go above (or below) certain levels. The most basic threshold is a fixed one (eg if a CPU goes above 95% utilisation, then raise an alert). However, this might just create unnecessary activity. Perhaps we do an extract at 2am every evening, which causes CPU utilisation to bounce at nearly 100% for long perids of time. We don’t want to receive an alert in the middle of the night for what might be expected behaviour.
Another example might be a higher network load for phone / SMS traffic on major holidays or during disaster events.
The great thing about modern analytics tools is that as long as they have long time series of data, then they can spot patterns of expected behaviour at certain times/dates that humans might not be observing and adjust alerting accordingly. This reduces the number of spurious notifications for network assurance operators to chase up on.Read the Passionate About OSS Blog for more or Subscribe to the Passionate About OSS Blog by Email