Adopting a platform-based approach to implementing the NWDAF brings CSPs significant value

This article was commissioned by Oracle Communications. Usage is subject to the terms and conditions in our copyright notice. Analysys Mason does not endorse any of the vendor’s products or services.

Communications service providers (CSPs) require deep insights into 5G network performance and activities in order to best capitalise on their 5G investments and drive operational efficiencies. However, constraints regarding accessing data and delivering insights to the required locations within the network will limit the benefits that CSPs can achieve. The network data analytics function (NWDAF), defined by 3GPP, addresses these challenges by providing standard interfaces to streamline analytics workflows. NWDAF implementation should be considered to be part of CSPs’ broader analytics strategies, and they should plan to take a platform-based approach.

The NWDAF addresses CSPs’ difficulties in accessing data in the 5G core

The 5G standalone (SA) core network is central to CSPs gaining full access to the key 5G capabilities, such as network slicing, that promise to unlock new revenue opportunities. As CSPs deploy these 5G core networks, data analytics will be critical to the operational functions, such as the closed-loop automation of network functions and network slice lifecycle management, that are required to assure a high quality of experience (QoE) and meet service-level agreements (SLAs).

However, it is somewhat challenging to access and deliver data to analytics systems in order to analyse data and consume insights to address a range of use cases. The non-standardisation of the interfaces that enable the collection and delivery of data and insights across the network has contributed to this challenge. In addition, the distributed nature of the 5G SA network requires CSPs to rethink their current analytics strategies.

3GPP has defined the NWDAF, a network function for the 5G core, to streamline the operation of interfaces that support analytics workflows. Standard interfaces implemented within the NWDAF ensure that the access, consumption and analysis of data and the delivery of insights are simplified so as to empower CSPs to achieve their business objectives. A key function of the NWDAF is to use advanced technologies such as machine learning to enable analytics use cases such as predictive analytics in order to achieve objectives such as preventive network maintenance and proactive operations.

CSPs should view the implementation of the NWDAF in a broader analytics context and not as the creation of just another silo

The value of the insights generated by the NWDAF exceeds that of the use cases currently defined by 3GPP because the NWDAF has the potential to address other analytical use cases spanning a CSP’s organisation. 3GPP is primarily concerned with use cases that improve the internal operations of the core network, including user data congestion, user equipment (UE) mobility, communication analytics and pattern prediction. However, maintaining focus on these use cases limits the extent to which CSPs can maximise their investments in the NWDAF. Nonetheless, insights such as UE mobility currently provide CSPs with monetisation opportunities. For example, A1 Austria (with A1 Mobility Insights) and SK Telecom (with DataSpark) use these insights to enrich the services that they deliver to enterprise customers.

As such, the NWDAF should be part of a CSP’s journey to fulfil its company-wide vision for analytics. The NWDAF’s architecture supports this approach because its key components are disaggregated into two logical functions, providing data management and analytics services, respectively. The disaggregation of the NWDAF presents multiple benefits including creating the potential for the NWDAF to co-exist with, and contribute to, CSPs’ broader vision for analytics and providing CSPs with the opportunity to reuse pre-existing analytics components, thereby reducing the time and cost to implement the NWDAF.

Furthermore, the standard interfaces of the NWDAF can help CSPs to reduce the effort and time required for data engineers and scientists to access network data and integrate it with other data sets to drive company-wide objectives. CSP executives such as Verizon’s Chief Data Officer and Senior Vice President, Linda Avery, have noted that data scientists spend around 70% of their time accessing and preparing data for analysis. The use of proprietary interfaces to network assets is a key factor in this. The NWDAF can therefore be positioned as an important data and analytics repository with standard interfaces that can store and expose information to CSPs’ internal development teams or third-party developers to support higher-level analytics use cases.

CSPs should take a platform-based approach to implementing the NWDAF

The implementation of the NWDAF needs to support the function’s potential to contribute to current and future analytics use cases in order to enable CSPs to realise the full benefits that it can bring to their broader analytics strategies. Taking a platform-based approach to implementing the NWDAF provides CSPs with an opportunity to implement this network function as part of their broader analytics ecosystem and derive the benefits of adopting such a strategy.

Analysys Mason defines a platform as a structure that allows multiple applications participating in a ‘solution’ to be delivered within a common and shared technical and business framework. Consequently, an NWDAF implemented as a platform-based analytics solution enables the creation and operation of multiple analytics applications using a combination of common and shared resources to meet company-wide analytics needs. These common resources include data collection agents, a data/analytics repository, AI/analytics development and execution environments and interfaces that enable analytics applications to share data and insights. The benefits associated with adopting a platform-based approach include having a consistent environment for the accelerated development of the NWDAF, its associated use cases and other analytics applications.

CSPs can adopt open-source platforms, vendor-proprietary platforms or a mix of the two to implement the NWDAF. There are key characteristics that the platform will need to have, regardless of the approach used. These include having the required cloud-native credentials, aligning with the 3GPP standards defined for the NWDAF and possessing the key features that accelerate the development of the NWDAF and other analytics applications that will run on it.

By assessing vendors’ solutions based on the extent to which they support these characteristics, CSPs can ensure that the analytics platform that they invest in will support both the requirements of the NWDAF and the future analytics requirements of the business.