ORAN-HAutoscaling: A Scalable and Efficient Resource Optimization Framework for Open Radio Access Networks with Performance Improvements
Author:
Open Access
- 23 March 2025
- journal article
- research article
- Published by MDPI AG in Information
- Vol. 16 (4) , 259
Abstract
Open Radio Access Networks (ORANs) are transforming the traditional telecommunications landscape by offering more flexible, vendor-independent solutions. Unlike previous systems, which relied on rigid, vertical configurations, ORAN introduces network programmability that is AI-driven and horizontally scalable. This shift is facilitated by modern container orchestrators, such as Kubernetes and Red Hat OpenShift, which simplify the development and deployment of components such as gNB, CU/DU, and RAN Intelligent Controllers (RICs). While these advancements help reduce costs by enabling shared infrastructure, they also create new challenges in meeting ORAN’s stringent latency requirements, especially when managing large-scale xApp deployments. Near-RTRICs are responsible for controlling xApps that must adhere to tight latency constraints, often less than one second. Current orchestration methods fail to meet these demands, as they lack the required scalability and long latencies. Additionally, non-API-based E2AP (over SCTP) further complicates the scaling process. To address these challenges, we introduce ORAN-HAutoscaling, a framework designed to enable horizontal scaling through Kubernetes. This framework ensures that latency constraints are met while supporting large-scale xApp deployments with optimal resource utilization. ORAN-HAutoscaling dynamically allocates and distributes xApps into scalable pods, ensuring that central processing unit (CPU) utilization remains efficient and latency is minimized, thus improving overall performance.Keywords
This publication has 57 references indexed in Scilit:
- RLOps: Development Life-Cycle of Reinforcement Learning Aided Open RANIEEE Access, 2022
- Applications of Machine Learning in Resource Management for RAN-Slicing in 5G and Beyond Networks: A SurveyIEEE Access, 2022
- Intelligent zero trust architecture for 5G/6G networks: Principles, challenges, and the role of machine learning in the context of O-RANComputer Networks, 2022
- Evolution of Non-Terrestrial Networks From 5G to 6G: A SurveyIEEE Communications Surveys & Tutorials, 2022
- Machine Learning Approaches for Reconfigurable Intelligent Surfaces: A SurveyIEEE Access, 2022
- Optimal Resource Allocation Considering Non-Uniform Spatial Traffic Distribution in Ultra-Dense Networks: A Multi-Agent Reinforcement Learning ApproachIEEE Access, 2022
- A Survey on Reinforcement Learning-Aided Caching in Heterogeneous Mobile Edge NetworksIEEE Access, 2022
- O-RAN AI/ML Workflow Implementation of Personalized Network Optimization via Reinforcement Learning2021
- Dynamic multi-workflow scheduling: A deadline and cost-aware approach for commercial cloudsFuture Generation Computer Systems, 2019
- Recent Advances in Cloud Radio Access Networks: System Architectures, Key Techniques, and Open IssuesIEEE Communications Surveys & Tutorials, 2016