*By Martin Wessel
For telecom operators, it has always been necessary to faithfully understand the performance of their networks. To put this into practice, they have historically used different KPIs that allow them to supervise operation and predict where and when it is necessary to act so that the service does not fail. With the technological evolution of networks that happened in parallel with the arrival of 4G, Artificial Intelligence (AI) algorithms began to be used to detect anomalies and predictive analyzes that allowed for more efficient management of networks and the customer experience. Thus, solutions have emerged to predict problems of network congestion and early detection of failures.
Now, with new 5G networks increasingly present in Latin America, there are many new opportunities for both operators and the entire ecosystem. It will become increasingly important (and easier) to do more predictive analytics, anomaly detection, and trend analysis for use cases such as online customer experience management and personalized marketing.
With billions of IoT devices connected over 5G networks, operators will be in a leading position to make use of the massive volume of network data generated to sell valuable information to their business customers and thereby generate new revenue streams. Therefore, the compilation and processing of network data to obtain meaningful information and its subsequent treatment with analytics and AI tools will be essential.
NWDAF: The 5G Analytics Standard
To make all this possible, there is a protocol called NWDAF (Network Data Analytics Function). By using tools that comply with this protocol, operators will be at an advantage. But to do so, you need an integrated NWDAF solution that delivers the advanced analytic capabilities of AI to address all use cases.
NWDAF introduces a higher level of intelligence into 5G networks. With data abstraction and machine learning, KPIs based on inference by use case provide critical information to enable new use cases.
An example is the use of drones for public safety or industrial uses. They rely on 5G connectivity and low latency to not only stream video directly, but also calculate the drone's position, provide situational awareness, avoid collisions and receive instructions from a remote operator, all in real time.
To enable this type of use, the analysis must be available at the network terminal. NWDAF enables real-time analytics at the endpoint to optimize the user experience in ultra-low latency use cases such as this one, thereby creating new revenue opportunities for operators and new applications of the technology.
*Martin Wessel is Principal Industry Consultant at SAS for Latam
Notice: The opinion presented in this article is the responsibility of its author and not of ABES - Brazilian Association of Software Companies
EN 
PT 
