5G Core Network Architecture: 9 Key Network Functions

5G Core Network Architecture: 9 Key Network Functions

Reference Point System Architecture

The reference point architecture is a traditional model defined by explicit interfaces between network functions. Each function communicates with others through standardized reference points, such as N1 for UE-AMF and N2 for AMF-RAN interactions. This model preserves the clear delineation of control and data planes, allowing operators to follow a structured and interoperable deployment.

This architecture simplifies integration with existing 4G/LTE networks, supporting scenarios like non-standalone (NSA) deployments. It provides traceability and clarity in signaling paths, which is useful in troubleshooting and policy enforcement. However, its rigid structure can limit flexibility and scalability in dynamic 5G environments.

Service Based System Architecture

The service based architecture (SBA) is a shift in 5G core design. Instead of fixed reference points, network functions expose services via standardized APIs using protocols like HTTP/2. This enables functions to interact flexibly and dynamically, fostering microservices-based deployment and enabling cloud-native scalability.

SBA supports features like network slicing and on-demand resource provisioning. Each function—such as AMF, SMF, or PCF—acts as both service consumer and provider, registering and discovering services via the network repository function (NRF). This architecture enables automation, simplified orchestration, and efficient lifecycle management.

Related content: Read our guide to core network 5g (coming soon)

Core 5G Network Architecture Functions

1. Access and Mobility Management Function (AMF)

The AMF manages all control-plane signaling related to user equipment (UE) access and mobility. It handles:

Registration management: Processes initial UE registration, tracking area updates, and de-registration procedures.
Connection management: Establishes and releases NAS (Non-Access Stratum) connections with the UE.
Mobility management: Coordinates handovers between cells or base stations, interfacing with the RAN and maintaining UE context.
Authentication relay: Initiates and relays authentication requests and responses to/from AUSF.
Security context management: Establishes security parameters for UE communication, including ciphering and integrity protection.
Network slice selection assistance: Interacts with NSSF to allocate the UE to the appropriate network slice.

AMF is stateless regarding session data, focusing solely on control functions, and supports SBA by interacting with other network functions via service APIs.

2. Session Management Function (SMF)

The SMF oversees all session-related control functions, including:

PDU session management: Establishes, modifies, and terminates PDU sessions, which represent UE data connectivity.
IP address allocation: Assigns IP addresses or other identifiers to UEs during session setup.
User plane management: Selects and configures UPF(s) to route user data traffic according to policy.
Policy and charging enforcement: Applies rules received from PCF, such as QoS parameters or charging instructions.
Interworking: Supports interoperation with legacy EPC (Evolved Packet Core) in 4G environments for dual connectivity scenarios.

SMF aids in orchestrating the user data path and adapting it to dynamic network and service needs.

3. User Plane Function (UPF)

UPF handles all user data traffic and is responsible for:

Traffic routing and forwarding: Directs traffic between the RAN and data networks (DNs), such as the internet or enterprise LANs.
Packet inspection and QoS enforcement: Applies traffic classification, filtering, and QoS enforcement at the packet level.
Usage reporting: Generates usage data records for billing and analytics.
Session anchoring: Serves as an anchor point for session continuity during handovers or IP address changes.
Local breakout support: Routes traffic locally to reduce latency and offload central infrastructure, useful for edge computing.

UPF can be deployed centrally or at the edge, enabling flexible network design optimized for latency, cost, or performance.

4. Authentication Server Function (AUSF)

AUSF is responsible for UE authentication and aids in network security:

Authentication processing: Executes 5G-AKA or EAP-AKA’ procedures using credentials from UDM.
Session key derivation: Coordinates cryptographic key generation with AMF for secure communication.
Service authorization: Verifies UE eligibility to access requested services or slices.
Roaming support: Works with SEAF (security anchor function) and potentially external HSS in roaming scenarios.

AUSF ensures that only authenticated users gain access, forming a foundation for secure 5G communication.

5. Unified Data Management (UDM)

UDM serves as the centralized data repository for subscriber-related information. It provides:

Subscriber profile management: Stores subscription data, including access rights, service preferences, and slice information.
Authentication credentials: Manages long-term keys and credentials for secure access authentication.
Policy data provisioning: Supplies relevant data to PCF and SMF for policy and session management.
Support for number portability and slicing: Ensures users are routed to appropriate slices and networks based on subscription.

UDM enables consistent and personalized service delivery across a distributed 5G network.

6. Policy Control Function (PCF)

PCF provides centralized policy management across the 5G Core:

Policy decision and enforcement: Issues rules to SMF and AMF for QoS handling, charging, and access control.
Event exposure: Consumes network events (e.g., location change, usage thresholds) to adjust policies.
Application-aware control: Supports application function (AF) interaction to apply app-specific QoS and routing.
Slice-aware policing: Delivers distinct policy sets per network slice to maintain service differentiation.

PCF is crucial for managing network behavior dynamically.

7. Network Repository Function (NRF)

NRF enables service-based architecture (SBA) operations by managing:

Service registration: Allows network functions to register their availability, version, and service capabilities.
Service discovery: Enables lookup and connection between service consumers and providers.
Status monitoring: Maintains updated availability states of all registered functions for orchestration.

NRF enables dynamic scaling and resilience by supporting real-time service composition across the 5G core.

8. Network Exposure Function (NEF)

NEF provides a secure interface for third-party applications and external networks to interact with the 5G core. It:

API gateway: Exposes network functions such as QoS control, event notifications, and session data via standardized APIs.
Security and policy enforcement: Validates requests and ensures that access complies with operator policies.
Analytics and event subscription: Allows external apps to subscribe to network events (e.g., UE location, traffic status).

NEF assists in enabling service innovation by allowing safe and controlled external access to core network capabilities.

9. Network Slice Selection Function (NSSF)

NSSF manages network slicing by:

Slice selection assistance: Guides AMF and SMF on assigning UEs to appropriate slices based on subscription and network context.
Slice availability monitoring: Tracks operational status of slices to enable allocation and failover.
Configuration policy enforcement: Applies operator policies for slice assignment, prioritization, and isolation.

NSSF ensures efficient and policy-compliant use of network slices.

Best Practices to Implement 5G Core Network Architecture

Network operators should keep the following practices in mind when implementing a network architecture with a 5G core.

1. Adopt Cloud-Native Principles

Cloud-native architecture forms the foundation of modern 5G core deployments. This involves designing network functions (NFs) as stateless microservices that are loosely coupled and independently scalable. Functions are packaged into containers (e.g., Docker) and orchestrated using Kubernetes, allowing dynamic scaling based on traffic demand.

Microservices communicate through service meshes that manage networking, security, and observability at scale. With CI/CD pipelines, updates can be rolled out rapidly with minimal downtime. The use of container registries, declarative configuration, and infrastructure-as-code further standardizes and automates deployments, ensuring consistency across environments. 

2. Implement Control and User Plane Separation (CUPS)

CUPS is a critical architectural principle in 5G that improves network flexibility and efficiency. The control plane—comprising components like AMF and SMF—handles signaling and session management, while the user plane (UPF) manages the actual data transfer. By decoupling these planes, operators can scale and manage them independently based on workloads.

This separation enables distributed deployment models, such as placing UPFs near the network edge to serve latency-sensitive applications while keeping control functions centralized. CUPS also simplifies traffic engineering and policy enforcement, allowing differentiated treatment for various services.

3. Leverage Network Function Virtualization (NFV) and Software-Defined Networking (SDN)

NFV replaces purpose-built hardware appliances with software-based network functions running on general-purpose servers. This allows rapid deployment, cost savings, and vendor flexibility. NFV infrastructure includes a virtualized infrastructure manager (VIM), hypervisors, and orchestration layers to handle VNFs (virtualized network functions).

SDN complements NFV by decoupling the control plane from the data plane and centralizing network intelligence in software-based controllers. These controllers programmatically manage traffic flows and enforce policies across a distributed environment. SDN enables features like automated failover, load balancing, and real-time path optimization.

4. Integrate Edge Computing

Edge computing is essential to meet the ultra-low latency demands of applications like autonomous driving, remote surgery, and real-time analytics. By deploying compute and storage resources near the radio access network (RAN), operators can process data locally, reducing the need to send traffic to centralized data centers.

Integration with the 5G core involves deploying UPFs at the edge to enable local breakout of traffic. This allows selected services to bypass the core network and connect directly to local applications or internet gateways. Coupled with multi-access edge computing (MEC) platforms, operators can expose edge services via APIs and manage workloads based on location and demand.

5. Monitor and Optimize Network Operations

A well-functioning 5G core relies on continuous monitoring and optimization to maintain performance and reliability. Key performance indicators (KPIs) such as packet loss, latency, jitter, and throughput should be tracked in real-time across all network functions. Automation tools should integrate with orchestration systems to apply configuration changes, deploy patches, or scale resources without manual intervention.

Using AI and machine learning, operators can detect traffic anomalies, predict capacity bottlenecks, and trigger automated responses through closed-loop assurance systems. Observability frameworks should include logs, metrics, and distributed tracing to diagnose issues quickly and correlate events across components. 

5G IoT Connectivity with floLIVE

floLIVE is strategically positioned to support and enable 5G Internet of Things (IoT) connectivity, integrating it into its comprehensive global cellular networking solutions. The company emphasizes 5G’s role in future-proofing IoT deployments, enhancing performance, and enabling new use cases.

General 5G Support and Future-Proofing floLIVE’s cloud-native core network  is designed to support a wide range of cellular technologies, including 2G, 3G, 4G/LTE, and 5G. This ensures that floLIVE’s platform is future-proof and can adapt to evolving IoT needs. The company’s technology is constantly evolving, and customers automatically benefit from updates and upgrades to underlying technologies, including 6G. floLIVE’s global connectivity service is specifically designed for IoT domains and supports the common needs of global enterprises, IoT service providers, and mobile operators in the 5G space.

5G Core Network and Infrastructure floLIVE’s own core network infrastructure, is cloud-native and optimized specifically for IoT, including 5G capabilities. This design provides flexibility and scalability for supporting a massive number of connected devices to each cell, greatly improving data rates and adding special support for IoT and M2M use cases in 5G. The cloud-native nature allows for elastic, behavior-based scaling, which is crucial for handling the growth of 5G IoT deployments. It can run on various IT environments, including bare metal, virtual machines, containers, Docker, and Kubernetes.

Specific 5G Technologies: RedCap floLIVE is strategically positioned to support emerging LPWA (Low Power Wide Area) technologies, including 5G RedCap (Reduced Capability), also known as NR-Light. RedCap is a new 5G technology designed to bridge the gap between high-performance 5G solutions and lightweight massive IoT deployments, offering reduced complexity, lower costs, and moderate data rates while leveraging 5G capabilities. floLIVE’s adaptable technology integration ensures smooth transitions as the market shifts towards LTE Cat 1 bis and RedCap. The long-term outlook for LPWA suggests a complete integration of functionalities into the 5G ecosystem, with RedCap and LTE Cat 1 bis emerging as primary solutions.

Performance and Benefits of 5G for IoT 5G networks are built for IoT, offering significant improvements in bandwidth, performance, and the ability to link devices, networks, and systems. floLIVE’s global infrastructure, with its local Points of Presence (POPs) and local LTE/5G breakouts, is designed to reduce latency and ensure high throughput for 5G IoT applications. This is particularly important for mission-critical IoT use cases that require real-time data exchange, such as autonomous driving and remote diagnostics. For industrial IoT facilities, floLIVE offers Private 5G Network solutions that are fast, secure, and lightweight, seamlessly integrating into its global connectivity offering.

VoLTE and Vo5G Capabilities floLIVE aims to offer a comprehensive VoLTE (Voice over LTE) solution tailored for IoT devices, enhancing its overall cellular ecosystem. This includes supporting outbound and inbound calls for use cases like law-related communications, safety alerts, and remote health monitoring, as well as emergency call support for the automotive industry. The expansion of 5G networks is expected to further boost VoLTE adoption, introducing “Vo5G” services. floLIVE’s roadmap includes deploying IMS (IP Multimedia Subsystem) on its core network to support voice initiation and termination services, which will be integrated with its Connectivity Management Platform (CMP) for call preference, billing, and reporting management.

Deployment and Integration floLIVE’s 5G solutions are offered as-a-service with a pay-as-you-grow business model. Its cloud-native architecture allows for flexible deployment options, including on-premise, in the cloud, or in a hybrid model. This flexibility extends to integrating with existing MNO networks and customer platforms through its rich REST API suite.

In summary, floLIVE’s approach to 5G IoT connectivity focuses on providing a scalable, high-performance, and compliant solution that addresses the diverse and evolving needs of global IoT deployments, leveraging its owned and operated cloud-native infrastructure.