Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Noisy Neighbor Influence in the Data Plane of Beyond 5G Networks

Rodrigo Moreira, Larissa Ferreira Rodrigues Moreira, Tereza C. Carvalho, Flavio de Oliveira Silva

TL;DR

This work tackles Noisy Neighbor effects in beyond-5G data planes by introducing kernel-level instrumentation using $bpftrace$/eBPF to measure per-packet GTP-U decapsulation latency in a containerized UPF, attributing latency to specific TEID/$QFI$ flows. An empirical framework with nonparametric tests and quantile regression quantifies how NN and traffic patterns shape latency tails, revealing that even high-priority QoS flows can be degraded and that CPU headroom mitigates but does not completely eliminate tail risk. The findings provide actionable operator guidance for telemetry and runtime mitigation, paving the way for telemetry-driven control and more robust slice isolation in B5G networks.

Abstract

Virtualization and containerization enhance the modularity and scalability of mobile network architectures, facilitating customized user services and improving management and orchestration across the network. In the context of the 5th Generation Mobile Network (5G), these advancements contribute to reduced Operational Expenditures (OPEX) and enable sliced-based networking for novel applications and services. However, as beyond fifth-generation (B5G) networks aim to address the remaining challenges regarding network slice isolation, the shared underlying hardware can lead to data plane contention among slices, resulting in the Noisy Neighbor (NN) effect, which may compromise network slicing and Service-Level Agreements (SLAs). We propose a kernel-level instrumentation of the User Plane Function (UPF) to assess the impact of noisy slices on data plane processing. Our findings reveal that even prioritized slices are susceptible to degradation induced by NN, with observable effects on latency metrics pertinent to user experience.

Noisy Neighbor Influence in the Data Plane of Beyond 5G Networks

TL;DR

This work tackles Noisy Neighbor effects in beyond-5G data planes by introducing kernel-level instrumentation using /eBPF to measure per-packet GTP-U decapsulation latency in a containerized UPF, attributing latency to specific TEID/ flows. An empirical framework with nonparametric tests and quantile regression quantifies how NN and traffic patterns shape latency tails, revealing that even high-priority QoS flows can be degraded and that CPU headroom mitigates but does not completely eliminate tail risk. The findings provide actionable operator guidance for telemetry and runtime mitigation, paving the way for telemetry-driven control and more robust slice isolation in B5G networks.

Abstract

Virtualization and containerization enhance the modularity and scalability of mobile network architectures, facilitating customized user services and improving management and orchestration across the network. In the context of the 5th Generation Mobile Network (5G), these advancements contribute to reduced Operational Expenditures (OPEX) and enable sliced-based networking for novel applications and services. However, as beyond fifth-generation (B5G) networks aim to address the remaining challenges regarding network slice isolation, the shared underlying hardware can lead to data plane contention among slices, resulting in the Noisy Neighbor (NN) effect, which may compromise network slicing and Service-Level Agreements (SLAs). We propose a kernel-level instrumentation of the User Plane Function (UPF) to assess the impact of noisy slices on data plane processing. Our findings reveal that even prioritized slices are susceptible to degradation induced by NN, with observable effects on latency metrics pertinent to user experience.
Paper Structure (13 sections, 2 equations, 3 figures, 6 tables, 1 algorithm)

This paper contains 13 sections, 2 equations, 3 figures, 6 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Proposed Method
  4. Influence of Noisy Neighbors on Slice Isolation
  5. Statistical Analysis
  6. Evaluation Setup
  7. Results and Discussion
  8. Descriptive Analysis
  9. Group Comparisons
  10. Quantile Regression Analysis
  11. Visualization Analysis
  12. Discussion and Insights
  13. Concluding Remarks

Figures (3)

  • Figure 1: Proposed Instrumentation Method.
  • Figure 2: Experimental setup and placement of our instrumentation within the UPF data path.
  • Figure 3: Latency characteristics across traffic classes and CPU load.