Customer experience and operational automation were the top two priorities of communications service providers (CSPs) according to an Omdia survey in 2024. In an AI-flooded world, where telecom networks are expected to become more advanced, smarter, and faster, the demands placed on customer experience and automation are likely to increase dramatically.
Subscribers will be consuming heavier data requiring more bandwidth, higher accuracy and better quality, whether for gaming, video conferencing, real-time interacting or immersive experiences. Emergency responders will require ultra-reliable, low-latency communications, while industry and manufacturing will need integrated sensing and deterministic latency for massive machine-type communications equipment from connected sensors, cameras, and IoT devices.
Paving the way for the network to operate efficiently, the disparate types of data need to be collected, captured, and analyzed – preferably in one place. This is where the 3GPP Network Data Analytics Function (NWDAF) comes in. The NWDAF is a function within the 5G network core architecture that standardizes data collection and analytics. It enables real-time detection of user experience issues and dynamic adjustment of network quality of service configurations. Ingesting structured data from the network function provides real-time and predictive analytics. It is designed for closed-loop automation, feeding insights directly to network functions (NFs) and management functions (MnFs) to drive policy and optimize decisions in real time. NWDAF also provides analytics to external application servers via standard interfaces, enabling network operators to monetize the analytics generated by NWDAF.
NWDAF:
- Offers real-time data-driven decision making with analytics focused on network operations – such as subscribers’ predicted quality of experience (QoE), load level of a cell or network slice, as well as real-time detection of abnormal subscriber or network behavior. These can be analyzed, and the network optimized to drive immediate network adjustments.
- Enables automation by embedding intelligence into closed-loop processes at the network level. This means that NWDAF detects anomalies, or load issues, and notifies a network function like the policy control function (PCF), a management function such as the Network Slice Management Function (NSMF) or RAN intelligent controller (RIC), which then executes a policy change. This operates on its own within seconds.
- Offers enhanced customer experience and ensures quality of experience targets are met. NWDAF provides near real-time insights into user centric issues, e.g., detecting abnormal user experience behavior or sudden increase in latency and notifications can then help adjust quality of service or reroute traffic accordingly.
- Ensures SLAs are maintained with network slicing. If a service level agreement (SLA) violation occurs, an alarm can be triggered, allowing the operator to take corrective action using the provided information. This can be verified through closed-loop automation and is even reflected in performance reports.
- Prepares the network for 5.5G and 6G by adding a layer of intelligence to its core network and structuring massive data volumes into actionable insights. This paves the way for AI-native and next-generation networks such as 5.5G and 6G.
- Reduces operational costs through network optimization. Using performance metrics to improve network performance means higher network efficiency and lower operational costs. In some cases, NWDAF has the potential to reduce the total cost of ownership by up to 45%.
RADCOM’s NWDAF is a vendor-neutral analytics platform that can serve as NWDAF in the core or as an MDAF in the management layer. It integrates advanced AI/ML and is completely standards-compliant with 3GPP Rel. 16-18 use cases – even offering extended and customizable analytics.
One of the key differentiators is its data source flexibility and subscriber level insights, ingesting data from any source including: standard service-based interfaces, proprietary NF vendor’s event data records (EDRs), OSS data (PM/FM), and service assurance solutions. It captures control plane events from multi-vendor networks and sources and correlates granular per-subscriber data (at IMSI-level events/traffic flows) across the RAN and the core. The architecture is microservice-based and cloud-native, with modular components for data collection, storage, analytics, and exposure.
RADCOMs NWDAF is highly flexible, allowing operators to deploy analytics on-premises or in the cloud, as well as scale each component independently. RADCOM’s NWDAF is open, offering an independent, multi-vendor view of the network that can integrate data from different sources, including third-party OSS tools.
RADCOM’s NWDAF supports various deployment models including:
- Basic 3GPP NWDAF mode with standard network function interfaces
- An extended mode that ingests service assurance data and offers additional closed-loop interfaces
- A hybrid mode with a lightweight NWDAF proxy deployed in multiple 5GC Core sites and a centralized NWDAF engine.
- Cloud-based NWDAF with multi-tenancy, providing network analytics to hundreds of private networks from a single NWDAF instance.
Summary
NWDAF is the “brain” of the 5G intelligent network. It operates within the 5G core as a network function that collects data from the core network elements and from the RAN. It essentially adds a layer of intelligence to make informed decisions, enhancing customer experience and driving automation.
For more information on RADCOM’s NWDAF, view here
