In previous posts, I described the need for machine intelligence to drive closed-loop automation in complex 5G networks and why the 3GPP has defined a new standard – the Network Data Analytics Function (NWDAF) – which runs in the 5G Core and continuously analyzes network data, generating the critical real-time operational intelligence that is used to automate service orchestration and network operation functions.
Now let’s consider what is involved in defining an open standard for 5G network automation by digging into how NWDAF works.
Input Data Types and Providers
NWDAF collects network-related data provided by multiple sources:
- Network Functions (NFs) in the 5G Core
- Application Functions (AFs) in the 5G Core
- Management Data Analytics Functions (MDAFs) in the OAM layer
- Non-standard data providers, such as network monitoring probes
NWDAF in 3GPP Release 16 defines the format and semantics for each data type provided by NFs in the 5G Core. However, NWDAF does not yet define standard data types for other sources. This will be addressed in future 3GPP releases, determined by operational use case requirements.
NWDAF performs two types of analytics processing on collected data: statistical analytics for determining what is currently happening in the network (or has happened in the past) and predictive analytics to forecast what is likely to happen in the future, based on current and historical trends.
NWDAF defines a set of analytics types (differentiated by Analytics ID) and the output data for each type, specifying the format and semantics. NWDAF is not prescriptive about the implementation details of AI/ML algorithms, but it is unambiguous regarding the statistical and predictive analytics outputs that must be generated.
Architectural Framework and Operational Models
The 3GPP’s 5G Service-Based Architecture (SBA) defines the architectural framework and detailed operational models for how real-time intelligence generated by NWDAF is consumed and utilized by NFs in the 5G Core to automate service orchestration and network operations. The operational models are defined by a set of use cases, each characterized by the input data types provided by specific sources and the analytics processing performed on the data to generate the outputs utilized by consuming NFs.
The 5G SBA defines a set of Service-Based Interfaces (SBIs) for communication between all functions in the 5G Core. These are open APIs which are used to pass data between functions, including the data collected by NWDAF from provider functions and the analytics outputs delivered to consuming functions. The operational model for a given use case consists of a standard sequence of API calls between 5G Core functions to perform specific service orchestration and/or network automation tasks.
Four Pillars of 5G Network Automation
You can think of 5G network automation as resting on four pillars, defined in the 3GPP’s 5G SBA:
- A common architectural framework
- A set of use-case driven operational models
- Standard data types and semantics
- Open APIs between 5G core functions
The result is a well-defined system for automating network and service operations in the 5G Core.
If you are thinking that it might be overly ambitious for the 3GPP to attempt to define such a comprehensive standard, consider the ultimate payoff for mobile network operators (MNOs).
For too long, monitoring and analytics has been the realm of non-standard, vendor-proprietary solutions for generating monitoring data, collecting and analyzing it, and using the analytics outputs to inform service orchestration and network automation. As a consequence, 4G/LTE network operators have been locked into supplier-dependent operational silos, limiting their business and operational flexibilty.
With NWDAF and the 3GPP’s 5G SBA, operators won’t have to go down that path again. A standard approach to 5G network automation will allow operators to deploy multi-vendor networks with confidence, leveraging the unique capabilities of best-of-breed products. This will ultimately benefit all players in the 5G value chain: customers, operators and their suppliers.
Read more on 5G Analytics & NWDAF:
- Guavus 5G-IQ NWDAF
- 5G Analytics Standards: Who Needs ‘Em?
- Will 5G Complexity Overwhelm MNOs?
- NWDAF: Automating the 5G Network with Machine Learning & Data Analytics
- 5G Rollout Challenges & Analytics-Driven Remedies for MNOs
To learn more about Guavus’ 5G NWDAF approach, schedule a meeting with Guavus today.
About the Author
|Andrew Colby is Head of 5G Strategy and Product Management at Guavus, a pioneer in AI-based analytics for communications service providers.
As a member of the Guavus Office of the CTO, Andrew leads initiatives with customers to identify ways to apply analytics to improve and transform their operations and customer experience.
He has worked in the areas of telecom and IP networking, operational support systems, and data analytics, for more than 30 years.
Image attribution: iStock