Open RAN CU-UP Dynamic Scaling Without Load Balancers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing OpenRAN architectures face challenges in dynamically scaling user plane throughput to accommodate peak and non-peak hours, leading to potential service disruptions and resource inefficiencies, especially in critical applications like emergency calls and eHealth, without impacting subscriber sessions.
Innovation Solution
Implementing CU-UP as a cloud-native network function (CNF) with intelligent dynamic scaling, allowing for pod addition/deletion without load balancer, and signaling endpoint changes to maintain service continuity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional OpenRAN CU-UP architecture is used, then service continuity can be maintained, but resource utilization is inefficient and scaling is difficult during peak and non-peak hours
Solution Approach 1:
The patent implements dynamic scaling of CU-UP instances based on real-time traffic conditions. The system automatically adds or removes CU-UP instances (scaling from 1 to 5 pods) according to network load, enabling the architecture to adapt dynamically between peak and non-peak hours while optimizing resource utilization.
Solution Approach 2:
The system changes operational parameters by adjusting the number of active CU-UP pods based on traffic patterns. During peak hours, more pods are deployed to handle increased load, while fewer pods operate during non-peak hours, thereby optimizing resource allocation and improving productivity.
2Speed
If manual intervention is used for CU-UP scaling, then control is simplified, but response time increases and service disruption risk increases
Solution Approach 1:
The system performs self-service scaling by automatically detecting traffic conditions and initiating pod deployment or termination without manual intervention. The controller monitors network metrics and autonomously adjusts CU-UP instance counts, achieving fast response speed while reducing operational complexity through automation.
Solution Approach 2:
The patent implements feedback mechanisms where the system continuously monitors network traffic and resource usage, then adjusts CU-UP scaling decisions based on this feedback. This closed-loop control enables rapid response to changing conditions while maintaining manageable complexity through automated decision-making.
3Reliability
If load balancer is used for CU-UP scaling, then traffic distribution is improved, but system complexity and latency increase
Solution Approach 1:
The patent extracts and removes the load balancer component from the traditional scaling architecture. Instead of using a separate load balancing layer, the system directly manages pod deployment and traffic routing through the controller, simplifying the architecture while maintaining service continuity and reducing latency.
Solution Approach 2:
The system segments the scaling function into discrete, manageable operations where the controller independently manages pod lifecycle and traffic routing. This segmentation eliminates the need for a centralized load balancer while maintaining reliable traffic distribution across multiple CU-UP instances.
4Productivity
If more CU-UP pods are deployed, then throughput capacity increases, but power consumption and resource waste increase during non-peak hours
Solution Approach 1:
The system dynamically adjusts the number of active CU-UP pods based on real-time traffic conditions. During peak hours, more pods are deployed to maximize throughput capacity, while during non-peak hours, the system scales down to minimal necessary pods, thereby reducing power consumption and avoiding resource waste.
Solution Approach 2:
The patent changes the operational parameter of pod quantity based on network demand. The system transitions between different scaling states (1-5 pods) according to traffic patterns, optimizing the balance between throughput capacity and energy consumption by deploying resources only when needed.
Data Source
AI summary
A system is disclosed for providing Open RAN CU-UP high availability, the system comprising: at least one active CU-CP; at least one active CU-UP in communication with the at least one active CU-CP; and at least one standby CU-UP in communication with the at least one active CU-CP; wherein when a message may be received from a CU-CP that detects a failure of the at least one active CU-UP, the at least one standby CU-UP may be configured to take over and become an active CU-UP, thereby providing failover redundancy for the at least one active CU-UP.