If you own a car, you know you should have it serviced regularly, not only to get all the routine maintenance tasks done but to have an expert assess what could go wrong in the future. It’s always a good idea to listen when your mechanic warns you about the tread on your tires, or there is unusual wear and tear that could cause severe problems if not addressed in the near future.

The telecommunications world is no different. As well as regular checks and maintenance, we also need to recognize issues before they occur.

With networks evolving faster than ever, this need is growing by the day. The agility made possible by DevOps, or its telco customized version used in Nokia - DelOps (Delivery and Operations) and Continuous Integration, Continuous Delivery (CI/CD) has allowed new features and services to be launched within weeks or days. All stages of introducing these services - design, deploy, test and operate – need to keep pace and network monitoring is no exception.

To make the most of DevOps and CI/CD and the speed advantages they bring, predictive network monitoring is essential. It’s vital to rapidly assess and predict what could go wrong, so that networks can work without interruptions.

The power of prediction

Nokia’s Predictive Network Monitoring (PNM) Service predicts anomalies and their implications, isolates affected areas, analyses consequences and recommends actions. Through continuous monitoring, PNM can use real-time data, addressing issues before they affect the business.

The service offers a complete overview of performance, allowing continuous verification in near real time, to all levels of the telecommunication ecosystem.

So what is in it for the operators? PNM has shown it can cut the number of incidents affecting networks by 37%, faults per subscriber by 85% and reduce restore time by 48% - this then improves time between failures by 64%.

These numbers are to be improved using Artificial Intelligence (AI) and Machine Learning (ML). By using AI and ML to monitor network data in a closed loop system, we can automate the detection of abnormal behaviors and system anomalies.

So how does it work?

Predictive Network Monitoring operates based on three cornerstones:

1. Data collection: Network elements do produce thousands of counters and logs.
2. Intelligent correlation and analysis of such data with alarms and KPIs creates possibility to detect underlying problems of the network and predict future trends.
3. Utilising closed loop automation to connect this intelligent analysis with automated healing actions enables CSP to gain an early warning mechanism and capability to act prior to a network impacting incident.

All above is based on logical chain of analysis drilling down to root cause level, suggesting remedial actions, all in minutes instead of days.

So, while cars might not commonly have it just yet, telecommunication networks and respective solutions can already be smarter in predicting problems and take action before the issue affects the business, saving time and effort for operators. To learn more, please watch our video and visit our website.