Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Enhancing Cloud-Native Resource Allocation with Probabilistic Forecasting Techniques in O-RAN

Vaishnavi Kasuluru, Luis Blanco, Engin Zeydan, Albert Bel, Angelos Antonopoulos

TL;DR

This paper addresses uncertainty-aware, dynamic resource provisioning in cloud-native O-RAN for 6G networks. It proposes embedding probabilistic forecasting rApps in the Non-Real Time RIC, evaluating SFF, DeepAR, and Transformer against a deterministic LSTM baseline for predicting next-$24$ hours of PRBs, using distributions $\mathcal{Y}$ and point forecasts $\hat{y}$. Results show that DeepAR and Transformer achieve lower $MSE$ and balanced compute footprints, while SFF is fast but unable to capture temporal dependencies and LSTM performs worst. The findings demonstrate the practical viability of probabilistic forecasting in cloud-native O-RAN, enabling proactive, uncertainty-aware resource allocation to improve QoS, energy efficiency, and reliability in real-world networks.

Abstract

The need for intelligent and efficient resource provisioning for the productive management of resources in real-world scenarios is growing with the evolution of telecommunications towards the 6G era. Technologies such as Open Radio Access Network (O-RAN) can help to build interoperable solutions for the management of complex systems. Probabilistic forecasting, in contrast to deterministic single-point estimators, can offer a different approach to resource allocation by quantifying the uncertainty of the generated predictions. This paper examines the cloud-native aspects of O-RAN together with the radio App (rApp) deployment options. The integration of probabilistic forecasting techniques as a rApp in O-RAN is also emphasized, along with case studies of real-world applications. Through a comparative analysis of forecasting models using the error metric, we show the advantages of Deep Autoregressive Recurrent network (DeepAR) over other deterministic probabilistic estimators. Furthermore, the simplicity of Simple-Feed-Forward (SFF) leads to a fast runtime but does not capture the temporal dependencies of the input data. Finally, we present some aspects related to the practical applicability of cloud-native O-RAN with probabilistic forecasting.

Enhancing Cloud-Native Resource Allocation with Probabilistic Forecasting Techniques in O-RAN

TL;DR

This paper addresses uncertainty-aware, dynamic resource provisioning in cloud-native O-RAN for 6G networks. It proposes embedding probabilistic forecasting rApps in the Non-Real Time RIC, evaluating SFF, DeepAR, and Transformer against a deterministic LSTM baseline for predicting next- hours of PRBs, using distributions and point forecasts . Results show that DeepAR and Transformer achieve lower and balanced compute footprints, while SFF is fast but unable to capture temporal dependencies and LSTM performs worst. The findings demonstrate the practical viability of probabilistic forecasting in cloud-native O-RAN, enabling proactive, uncertainty-aware resource allocation to improve QoS, energy efficiency, and reliability in real-world networks.

Abstract

The need for intelligent and efficient resource provisioning for the productive management of resources in real-world scenarios is growing with the evolution of telecommunications towards the 6G era. Technologies such as Open Radio Access Network (O-RAN) can help to build interoperable solutions for the management of complex systems. Probabilistic forecasting, in contrast to deterministic single-point estimators, can offer a different approach to resource allocation by quantifying the uncertainty of the generated predictions. This paper examines the cloud-native aspects of O-RAN together with the radio App (rApp) deployment options. The integration of probabilistic forecasting techniques as a rApp in O-RAN is also emphasized, along with case studies of real-world applications. Through a comparative analysis of forecasting models using the error metric, we show the advantages of Deep Autoregressive Recurrent network (DeepAR) over other deterministic probabilistic estimators. Furthermore, the simplicity of Simple-Feed-Forward (SFF) leads to a fast runtime but does not capture the temporal dependencies of the input data. Finally, we present some aspects related to the practical applicability of cloud-native O-RAN with probabilistic forecasting.
Paper Structure (14 sections, 1 equation, 3 figures, 1 table)

This paper contains 14 sections, 1 equation, 3 figures, 1 table.

Table of Contents

  1. Introduction
  2. Transition to Cloud Native Architectures
  3. Contributions
  4. O-RAN Architecture
  5. Methodology
  6. Overview of Cloud Native Resource Allocation in O-RAN
  7. Introduction to Probabilistic Forecasting Techniques
  8. Integration Approach of Probabilistic Forecasting as Radio Application (rApp) in O-RAN
  9. Experimental Setup and Results
  10. Experimental Setup
  11. Results and Analysis
  12. Real-World Applications
  13. Conclusion
  14. Acknowledgements

Figures (3)

  • Figure 1: O-RAN architecture with probabilistic forecast as rApp.
  • Figure 2: Training (a) CPU Usage , (b) Memory Usage
  • Figure 3: Prediction (a) CPU Usage , (b) Memory Usage