Source-linked AI summary

Integrating Wi-Fi into 3GPP 5G Network Slicing: An Experimental Prototype Study

Nelson Ion de Oliveira, Marília Costa Muniz, William M. C. do Nascimento, João Pedro Brasil, Victor Farias Monteiro, Sérgio Barros, Maykon Silva, Augusto Venâncio Neto e Vicente A. de Sousa

arXiv:2609.02625v1cs.NI

TL;DR

Existing 3GPP slicing architectures provide limited empirical evidence for unified Wi-Fi and cellular slice instances. This paper adapts TNGF and related functions in an open-source testbed to integrate 5G and WLAN access. The prototype demonstrates dynamic Wi-Fi resource control and QoS differentiation across multi-RAT slices, while recognizing limits in scaling automation for heterogeneous networks.

  • Problem

    Empirical verification of unified network slicing across WLAN and 5G using real radio access protocol stacks remains sparse.

  • Method

    The study extends TNGF and NSSMF capabilities in a free5GC-based testbed to integrate Wi-Fi access into 5G Network Slice Instances.

  • Results

    The testbed demonstrates real-time multi-RAT slicing with dynamic bandwidth management and QoS differentiation in a shared Wi-Fi environment.

  • Takeaways & Limitations

    The prototype shows that Non-3GPP domains can receive 5G-like security, management, isolation, and transmission-resource control through integrated slicing.

  • Takeaways & Limitations

    The study requires more dynamic and comprehensive integration to meet complex large-scale heterogeneous-network requirements.

Abstract

from arXiv · show

Network Slicing (NS) is a fundamental pillar of 5G and beyond networks, enabling the provisioning of isolated, logical networks tailored to specific Quality of Service (QoS) requirements. While 3GPP standards comprehensively define slicing architectures over cellular access networks, the seamless integration of Non-3GPP technologies such as Wi-Fi into a unified slice instance remains an active area of investigation, particularly regarding empirical validation. This paper presents an end-to-end prototyping study that integrates 5G Standalone (SA) and Wi-Fi networks by adapting the Trusted Non-3GPP Gateway Function (TNGF) to extend NS to WLAN networks, enabling the unified management of Wi-Fi transmission resources. We implement a functional testbed leveraging an open-source 5G Core and an explicit Non-3GPP access to validate multi-Radio Access Technology (multi-RAT) slice operation. Our empirical results showcase the dynamic viability of multi-RAT slicing under varying bandwidth allocations and traffic steering policies, providing a concrete proof of concept for unified 3GPP and Non-3GPP service delivery.

I. INTRODUCTION

5G network slicing enables virtual networks tailored to diverse services, but extending unified slice instances across Wi-Fi requires empirical validation. This study builds a multi-RAT prototype using TNGF to integrate WLAN resources with 5G slicing.

  • Network slicing partitions physical infrastructure into virtual networks optimized for URLLC, eMBB, and mMTC use cases.
  • Wi-Fi must be incorporated into Network Slice Instances because it carries much local and indoor data traffic and serves many Wi-Fi-based XR and IoT devices.
  • Most prior work uses mathematical modeling or software simulation, leaving real-world signaling, encapsulation overhead, and processing latency insufficiently validated.
  • The prototype implements trusted Non-3GPP access through TNGF, establishing secure IPsec tunnels from Wi-Fi-connected UEs into the core user plane.
  • The study extends 5G benefits including logical isolation, service differentiation, and traffic management to legacy devices connected through WLAN.
  • The empirical proof of concept demonstrates real-time multi-RAT NSI operation and observes throughput effects from dynamic bandwidth adaptation under distinct traffic conditions.

II. RELATED WORK

Prior research covers Wi-Fi slicing and heterogeneous connectivity, but real protocol-stack validation of unified Wi-Fi and 5G network slicing remains sparse. This study addresses that gap with an empirical multi-RAT NSI testbed.

  • The literature review searched Scopus and Web of Science for studies combining Wi-Fi, wireless access, and network slicing.The search returned 136 records, which were analyzed to identify investigated strategies.
  • The review screened 35 papers in full text and found that only 23 specifically addressed network slicing in Wi-Fi networks.
  • Early LWA and MR-DC work supports heterogeneous connectivity and traffic splitting but operates mainly at the RAN layer rather than providing end-to-end 5G virtualization and isolation.
  • Reviewed studies pursued QoS, spectral-efficiency, and resource-management goals, yet frequently lacked sufficient Wi-Fi–5G integration or collaboration.
  • Empirical verification on real radio access protocol stacks remains sparse, with existing testbeds often limited to 5G-only slicing or unsliced multi-RAT configurations.

III. PROTOTYPING ARCHITECTURE AND TESTBED DEPLOYMENT

The testbed architecture is designed so one S-NSSAI identifier spans 3GPP and Non-3GPP access technologies, providing a unified multi-RAT slice representation.

  • A single S-NSSAI identifier spans all 3GPP and Non-3GPP access technologies in the multi-RAT network slicing testbed.

A. Architectural Framework

The architectural framework follows 3GPP principles with an isolated user plane and TNGF-based trusted Non-3GPP access, using free5GC as the experimental 5G Core platform.

  • Architectural Framework: The unified NSI is identified by an S-NSSAI containing a Slice/Service Type and Slice Differentiator.
  • Architectural Framework: TNGF enables Non-3GPP devices to communicate with the 5GC through N2 and N3 while maintaining communication security and integrity.
  • Architectural Framework: The experimental setup centers on free5GC, an open-source Linux Foundation project intended to provide a 3GPP-compliant 5G Core.
  • Architectural Framework: The platform comparison identifies unavailable features across open-source 5GC solutions, while free5GC is selected as a suitable balance against Magma’s higher deployment complexity.
  • Architectural Framework: The framework supports platform selection for laboratory experiments, new-functionality development, and simulated deployments.

B. Multi-RAT Network Slicing Testbed

The testbed extends 5G network slicing to trusted Wi-Fi access by coordinating slice management, Wi-Fi transmission resources, and 5G security through enhanced NSSMF and TNGF functions.

  • Architecture: The assumed OSS/BSS lifecycle functions initiate slice instantiation, while the testbed extends TNGF and NSSMF capabilities for Wi-Fi integration.The CSMF, NSMF, and NSSMF translate service requirements, manage the slice lifecycle, and control component subnets.
  • Wi-Fi resource management: Each Wi-Fi NSI maps to a dedicated transmission queue identified by its SST/SD tuple.The AP configures per-slice airtime and bandwidth controls using 5QI and DL AMBR parameters.
  • Security and connectivity: The TNGF and UE establish trusted Non-3GPP access through EAP-5G authentication, derived keys, and an IPsec tunnel for user-plane traffic.After registration and PDU session approval, IKEv2 creates a dedicated IPsec Child SA carrying encapsulated user data.
  • Limitations: The solution requires a Linux-based Wi-Fi AP with suitable airtime scheduling and a UE supporting WPA2/WPA3-Enterprise authentication.These requirements enable HTB queue manipulation, per-connection airtime scheduling, and 802.1X/EAP trusted-access authentication.

IV. EXPERIMENTAL SETUP AND TEST CASES

The experimental testbed centers on free5GC and combines open-source 5G and Wi-Fi components with TNGF-based signaling, authentication, and slice-resource configuration.

  • Infrastructure: The infrastructure uses a free5GC 5G Core extended to terminate TNGF connections and manage Non-3GPP registrations, authentications, and sessions.Separate UPF instances can be instantiated per NSI to enforce slice traffic isolation.
  • Infrastructure: The Non-3GPP access point runs OpenWrt on a Wi-Fi 4 802.11n device with a maximum transmission capacity of 56 Mbps.Two notebook UEs run the TNGFUE client to establish wireless communication with the AP.
  • Configuration: During UE registration, the extended TNGF extracts DL AMBR, 5QI, and SST/SD values and asynchronously sends bandwidth, airtime-priority, and slice-identification parameters to the AP.This connects 5GC slice configuration with Wi-Fi resource controls.
  • Procedure: The Wi-Fi procedure comprises sequential phases involving IP connectivity, IKEv2/IPsec signaling, and Non-3GPP registration.The UE receives a local AP address, initiates IKEv2 with the TNGF, and carries NAS registration messages containing the S-NSSAI.

A. Test Cases

The test cases evaluate a Wi-Fi network slice with two UEs across six operating modes, varying slicing, bandwidth limits, and QoS priorities.

  • Test cases: The experiment registers two Wi-Fi UEs in the 5GC and applies individualized security and QoS treatment using IMSIs, 5QI, and DL AMBR.Scenarios 1 and 6 use best-effort operation, while Scenarios 2–5 activate the NSI with different bandwidth-limit combinations.
  • Test cases: Scenarios 4 and 5 assign 5QI 87 and 5QI 4 to evaluate traffic prioritization, reversing the prioritized UE between scenarios.Scenario 4 prioritizes UE1, whereas Scenario 5 gives preferential treatment to UE2.

V. RESULTS AND DISCUSSION

The experiments show that integrated slicing can stabilize or differentiate Wi-Fi throughput through bandwidth limits and airtime/QoS policies, while excessive limits recreate contention.

  • Evaluation: The evaluation measures throughput stability, service prioritization, and bandwidth enforcement under concurrent UDP traffic generated by iPerf3.Flows run from the 5GC toward the UEs.
  • Bandwidth enforcement: 20 Mbps limits in Scenario 2 stabilize bandwidth distribution and constrain throughput to the configured value for both users.The scenario uses homogeneous QoS with the same 5QI.
  • Bandwidth enforcement: A 40 Mbps limit in Scenario 3 allows requested rates to exceed AP capacity, producing transmission-resource contention similar to best-effort operation.The result contrasts with the more stable constrained allocation in Scenario 2.
  • Discussion: The results validate dynamic resource management and QoS differentiation for Wi-Fi integrated with the 5GC in a shared wireless environment.The integrated mechanism applies these policies under the tested traffic conditions.

VI. CONCLUSIONS

The study demonstrates that extending network slicing to Non-3GPP domains can provide devices without native 5G support with security, management, and isolation capabilities. Its testbed shows that 5GC QoS parameters can dynamically control Wi-Fi bandwidth and transmission time, while larger heterogeneous deployments require more comprehensive integration.

  • The prototype demonstrates that network slicing can extend to Non-3GPP domains, supporting security, management, and isolation for devices without native 5G support.
  • The free5GC–OpenWrt testbed shows that extending the TNGF enables effective control of Wi-Fi transmission resources.
  • 5QI and DL AMBR parameters enable dynamic bandwidth management through HTB queues and transmission-time management through Airtime Fairness.
  • The integration remains insufficiently dynamic and comprehensive for the more complex requirements of large-scale heterogeneous networks.
Loading 2609.02625v1…