O-RAN Overload Control via Dynamic Scheduling Metrics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Open Radio Access Network (O-RAN) networks face challenges in effectively controlling overload situations, leading to undesirable system behavior due to increased processing overhead and potential packet drops, especially as the number of Data Radio Bearers increases, causing TTI-overstretch and degradation of end-to-end throughput.
Innovation Solution
Implementing DL and UL overload control optimizations at the Distributed Unit (DU) and Centralized Unit (CU) by computing scheduling metrics for a subset of UEs within a window of slots, dynamically allocating buffer space, and adjusting scheduler weights, with the assistance of near-Real-Time RIC for overload analysis and control actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If scheduling metrics are computed for every active UE in each slot to maintain optimal resource allocation, then resource allocation efficiency is improved, but processing overhead increases causing TTI-overstretch and system overload
Solution Approach 1:
The patent segments the UE population into multiple groups based on scheduling priority, activity level, and QoS requirements. Different scheduling metric computation frequencies are applied to different groups: high-priority UEs have metrics computed every slot, while lower-priority UEs have metrics computed less frequently. This segmentation resolves the contradiction by maintaining high resource allocation efficiency for critical UEs while reducing overall processing overhead across the system.
2Quantity of substance
If the number of Data Radio Bearers is increased to support more users and services, then network capacity and service diversity are improved, but processing overhead and packet drop risk increase
Solution Approach 1:
The patent dynamically changes scheduling parameters such as metric computation frequency, resource allocation granularity, and priority thresholds based on the current number of active Data Radio Bearers and system load conditions. When the number of DRBs exceeds certain thresholds, the system adjusts parameters to reduce processing overhead and prevent packet drops, thereby maintaining reliability while supporting high network capacity.
3Measurement precision
If scheduling metric computation frequency is increased to respond to rapid channel changes, then scheduling accuracy is improved, but processing time requirements increase causing TTI-overstretch
Solution Approach 1:
The patent implements dynamic adjustment of scheduling metric computation frequency based on real-time channel condition variability and system load. During periods of rapid channel change, the system increases computation frequency for affected UEs to maintain scheduling accuracy. During stable periods or high-load conditions, the frequency is reduced to meet processing time requirements and prevent TTI-overstretch, thus balancing accuracy and timing constraints.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method of implementing data traffic overload control utilizing an E2 node for i) a stand-alone (SA) 4G or 5G architecture wireless network, and a non-stand-alone (NSA) architecture wireless network, the method including: detecting an overload condition at the E2 node; and performing an overload control action including: i) reducing a number of UEs for which corresponding scheduling metric is computed in each transmission time interval (TTI); ii) reducing a number of UEs for which UL grant is given; iii) reducing a size of UL grant given to each UE; (iv) reducing the amount of radio resource the E2 node provides to each cell; (v) dynamically allocating increased buffer spaces to radio link control (RLC) queues in the DU; (vi) reducing an activity factor for selected data radio bearers (DRBs); and (vii) adjusting relative data transmission rates between a 4G leg and a 5G leg of data transmission.