O-RAN Overload Control via Dynamic Buffer and Scheduling
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 dropping, especially in stand-alone (SA) 4G/5G and non-stand-alone (NSA) architectures.
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 (RIC) for overload analysis and control actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If scheduling metrics are computed for every active UE in each slot to maintain precise resource allocation, then resource allocation accuracy is improved, but processing overhead increases leading to overload conditions
Solution Approach 1:
The patent segments the UE population into different groups (e.g., URLLC UEs, eMBB UEs) and applies different scheduling metric computation frequencies to each group. High-priority UEs like URLLC users have their metrics computed every slot, while other UEs have metrics computed less frequently or only when needed, thereby reducing overall processing overhead while maintaining allocation accuracy for critical services
Solution Approach 2:
The patent implements periodic scheduling metric computation where UEs are scheduled at different intervals based on their service requirements. Instead of computing metrics for all UEs every slot, the system computes metrics periodically at appropriate intervals for each UE group, reducing processing overhead while maintaining adequate resource allocation accuracy
2Reliability
If buffer space is increased to prevent packet dropping during overload, then system reliability is improved, but resource utilization efficiency deteriorates
Solution Approach 1:
The patent implements dynamic buffer allocation where buffer sizes are adjusted in real-time based on current traffic conditions, UE priorities, and service requirements. During overload conditions, buffer space is dynamically allocated to high-priority UEs (e.g., URLLC) while reducing allocation to lower-priority UEs, thereby maintaining packet delivery reliability for critical services without unnecessarily holding buffers that would reduce overall resource utilization efficiency
Solution Approach 2:
The patent applies different buffer allocation strategies to different UE groups and service types. Instead of uniformly increasing buffer space for all UEs, the system provides enhanced buffer protection locally to specific high-priority UEs or service types (e.g., URLLC traffic) while maintaining normal buffer management for other UEs, thus improving reliability where needed without sacrificing overall resource utilization efficiency
3Reliability
If scheduler weights are adjusted to prioritize certain UEs during overload, then service quality for priority UEs is improved, but fairness among all UEs deteriorates
Solution Approach 1:
The patent applies differentiated scheduler weights locally to specific UE groups based on their service requirements. High-priority UEs such as URLLC users receive enhanced scheduler weights during overload conditions to ensure their service quality, while normal UEs maintain their standard weighting. This localized quality adjustment ensures that fairness is maintained for the majority of UEs while providing priority service where needed
Solution Approach 2:
The patent implements periodic re-evaluation of scheduler weights based on current network conditions and UE behavior. Instead of maintaining fixed priority weights, the system periodically adjusts weights to reflect changing conditions, allowing fairness to be restored when priority UEs no longer require enhanced service. This periodic adjustment mechanism balances service quality for priority UEs with overall fairness among all UEs
Data Source
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.