written by ABI Research Senior Analyst Don Alusha
Network Data Analytics Function (NWDAF) is a Third Generation Partnership Project (3GPP)-based analytics function designed to provide analytics to drive actionable insight. NWDAF focuses on the 5G Core (5GC) network. It helps Communication Service Providers (CSPs) obtain clarity in terms of the business value they get from data and analytics from an operations standpoint. Further, NWDAF enables the industry to place data and analytics at the heart of its innovation. NWDAF serves as a stepping-stone for CSPs to build analytics functions that can propel them forward to ride the growth wave associated with the digital economy model.
Use Cases for NWDAF
Data volumes are growing exponentially and the types of data that CSP must collect and analyze are set to also grow. NWDAF will play a key role in providing intelligence for several operational use cases, but also potentially feeding strategic planning and decisions. One (or multiple) NWDAFs can support multiple Operations, Administration, and Maintenance (OAM) functions spanning the 5GC and applications for automation and orchestration management and operational intelligence. Data collection and analytics come from multiple data sources, including 3GPP Service-Based Interface (SBI) and non-standard sources to support various use cases. Guavus’ 5G-IQ NWDAF, for example, is a solution among many others that support many use cases including the following:
- Network Performance – Guavus’ 5G-IQ NWDAF can provide network and User Equipment (UE) performance and congestion-related statistics and predictions for its consumers to understand how network conditions are impacting Quality of Experience (QoE) and make intelligent policy decisions to improve network performance and subscriber QoE. Key application areas are network performance analysis, user data congestion, policy optimization, and automated congestion management.
- Service Experience – Today, the 5GC is, to a large extent, a service business as the 5GC has a new service-based architecture, new Network Functions (NFs), and new Application Program Interfaces (APIs) between it and the Container-as-a-Service (CaaS) layers that were not present in the past. Consequently, a key use case is to effectively facilitate the digital transformation of operations and deliver a superior service experience.
- Load Analysis – With this use case, CSPs obtain real-time operational intelligence in the form of load-related statistics and predictions for network slices and NFs making up those slices to its consumers to adjust NF resources and/or select the optimum traffic path or slice, ensuring efficient resource utilization. Key application areas are NF load analysis, load balancing, and slice load analysis.
- Service Assurance – This helps provide current and expected service experience insights in the form of statistics or predictions to enable operators to deliver and assure network and next-generation service performance. Key application areas are observed service experience, Quality of Service (QoS) sustainability analytics, and application-aware performance optimization with the flexibility to accommodate a rich array of services with heterogeneous QoS needs.
- User Equipment Behavior Analysis – Lastly, with NWDAF solutions coming from Guavus and other vendors, CSPs obtain insight into UE behavior with respect to mobility and usage patterns to enable prediction-driven UE control and management to optimize and protect networks from unexpected behavior and improve subscriber QoE. Key application areas are UE communication analytics, UE mobility, UE expected behavioral parameters, and UE unexpected behavior.
Standards and 3GPP NWDAF Roadmap
NWDAF is one standard, among a few others, which offers both statistical analytics (current and past state) and predictive analytics (future state). Other standards are Near-Real-Time Radio Intelligent Controller (Near-RT RIC) and Non-Real-Time Radio Intelligent Controller (Non-RT-RIC) for Open RAN (O-RAN) analytics and Operations Support System (OSS)-level standards. The level of adoption for each standard remains to be seen. Regardless of the standard that predominates, the industry should seek to define an information data model. In other words, industry bodies should establish a data model that facilitates an accurate and detailed description for the internal topology, procedures, and life cycle of slices spanning the core network and the RAN.
Work remains to be done, particularly on a common standard for data format, which may raise challenges for data unification. A common data standard eliminates data normalization, defines common semantics for Key Performance Indicators (KPIs), and develops common operational models for orchestration and automation. These benefits are bound to have a positive the ripple effect on CSPs and suppliers.
- Select Best-of-Breed Products: Enterprise verticals are likely to have varying requirements for NWDAF analytics. CSPs should have the option to pick best-of-breed products, and suppliers should compete on the ability to innovate and ensure product quality.
- Potentially Reduce Integration Work: A widely adopted standard, such as NWDAF, can potentially reduce the level of custom system integration required and achieve faster time-to-market, while reducing initial and ongoing costs.
- Productize Analytics: Suppliers can better productize analytics, orchestration, and automation functions based on common 5G operational scenarios and use cases.
Figure 1: 3GPP NWDAF Roadmap
(Sources: Guavus, ABI Research)
Conclusion
In addition to having the right tools and analytics functions in place (e.g., NWDAF from an operational intelligence perspective), success for CSPs will come from how NWDAF is used. Using 5G analytics tools and NWDAF to design new services can be done in myriad ways. But if CSPs, and the broader industry, are to be effective with using these tools, they may have to pursue a path that aligns with their unique circumstances—market positioning, growth strategy, and customers. It will not be so much about an NWDAF function as it will be about establishing the right operational context; for example, a fitting culture where virtually everyone in the company seeks ways for functions like NWDAF, Big Data, and analytics to enhance network and business operations.
