5G network intelligence
February 10, 2021 - Sara Philpott
Sara Philpott, Data Product Director, discusses the growing focus on Network Data Analytics Function (NWDAF)
Now that a number of service providers have taken their first steps to deploying a 5G Core, attention is focusing to the strategic value of using network data to manage 5G services and customers’ experiences.
5G network intelligence – Needs more than probe data
Probe vendors suggest that some of the 3GPP defined use cases can be performed today, providing sufficient functionality to manage the network load levels. They question the need for a centralised predictive analytics function automatically guiding resource optimisation as it continually inhales the fresh statistics from the network. Probes feeding from the access data perform a focused and siloed role, providing user plane information on carried load to enable policy decisions to ease congestion, for example. However at the control plane level, a much more diverse and complex ecosystem of activity is necessitating the need for a central function to provide the intelligence to automatically guide resource allocations to ensure sufficient capacity for individual quality of experience requirements. This is where the 5G 3GPP guided Networks Data Analytic Function (NWDAF) comes into play.
5G Intelligence – Step forward NWDAF
NWDAF has considerable impact on the 5G business as the continuous tuning and optimisation of resources influence not only end user experience, but also the entire total cost of ownership of the 5G network. Scaling resources as required means that capacity is provisioned based on anticipated demand thus saving on valuable network resources. Also service providers are growingly increasing conscious of their carbon and digital footprint with genuine and concerted efforts to providing more ‘green’ services as part of their contribution to the environment. Focus and guidance on optimised use case implementation from the 3GPP standards, particularly Release 17, is nudging the industry in this direction.
Interweaving AI functionality
The NWDAF interweaves AI functionality, while collecting the various network functions and operational management data, and outputting anticipated scoring and anticipated insights. Every few minutes, recalculated insight is served up as the models continually consume the fresh data to deliver real-time network intelligence which can be actioned by AI. This insight provided by neural network algorithms not only provides a vital input to inform and guide network resources, but also serves the business in helping to monetise 5G services.
NWDAF and policy and charging integration
The intimacy of the NWDAF with the policy and charging functions of the network will enable use cases such as ability to detect unusual consumption patterns. A 5G SIM may be priced and packaged to support a particular service, for example. The AI algorithms will be able to detect unusual or abnormal behaviour suggesting, perhaps, commercial abuse of the service offer or identifying a new opportunity to monetise in accordance with unexpected service adoption and consumption patterns.
What is very true, as the industry savours the possibilities with NWDAF, is that many business applications will benefit from the enhanced insight provided by this centralised function and although some use cases may be difficult to anticipate due to current low volumes of 5G traffic, there is plenty of opportunity for those service providers with ambition.