5G vRAN Distributed Unit Updates with Priority-Based Thread Scheduling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge in 5G wireless networks is managing software versioning for distributed units (DUs) in a virtual radio access network (vRAN) when both an existing and an updated DU version attempt to execute threads simultaneously on a single server, leading to malfunctions and data traffic disruptions.
Innovation Solution
Implementing a process scheduler that prioritizes real-time processing threads for an existing DU version and initializes an updated DU version with non-real-time threads, using an update agent to block concurrent real-time thread execution and migrate data traffic efficiently between versions, while utilizing duplicated data packets through virtual network interface cards to bypass the operating system kernel.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If both existing and updated DU versions execute real-time processing threads simultaneously on the same server, then data traffic processing capability is improved, but system stability deteriorates due to thread conflicts and malfunctions
Solution Approach 1:
The patent segments processing threads into real-time and non-real-time categories, and further segments them by DU version (v1 and v2). The process scheduler separates thread execution based on priority levels, with real-time threads for v1 having highest priority and real-time threads for v2 having lower priority. This segmentation prevents conflicts between different DU versions while maintaining processing capability.
Solution Approach 2:
The patent implements dynamic thread priority management where the process scheduler adjusts thread execution based on real-time requirements. The scheduler dynamically determines which threads execute first by comparing priorities, allowing the system to adapt to different processing needs. This dynamic approach enables both versions to process data without causing stability issues through rigid simultaneous execution.
2Device complexity
If a single server architecture is used for DU, then device complexity is reduced, but software versioning support becomes complicated due to concurrent thread execution conflicts
Solution Approach 1:
The patent segments thread execution rights by implementing priority-based access control. The process scheduler segments CPU access based on thread priority levels, allowing multiple DU versions to coexist on a single server without conflicts. This segmentation enables versioning support while maintaining the simplicity of the single-server architecture.
Solution Approach 2:
The process scheduler acts as an intermediary between different DU versions, mediating thread execution conflicts. It intercepts and manages thread creation and execution requests, ensuring that real-time threads from different versions execute without interfering with each other. This intermediary mechanism enables software versioning on a single server without increasing architectural complexity.
3Reliability
If real-time processing threads are blocked for updated DU version during migration, then data traffic disruption is minimized, but processing speed temporarily decreases
Solution Approach 1:
The patent implements preliminary action by pre-creating non-real-time processing threads for the updated DU version (v2) before migrating data traffic. These threads are prepared in advance with lower priority, allowing the system to switch traffic to v2 smoothly without blocking real-time processing. The preliminary thread creation avoids abrupt stops while maintaining traffic continuity.
Solution Approach 2:
The patent uses dynamic thread priority management to handle processing speed during migration. The scheduler dynamically adjusts which threads execute based on priority levels, allowing non-real-time threads to run during migration periods without blocking real-time traffic processing. This dynamic approach maintains data traffic continuity while minimizing overall processing speed impact through efficient resource allocation.
Data Source
AI summary
A distributed unit (DU) in a virtual radio access network (vRAN) of a wireless 5G (fifth generation) network is implemented using software-defined networking (SDN) technology on a single computer server that is physically co-located with a radio unit in a distributed RAN (D-RAN) architecture. An update agent facilitates versioning support through interoperations with different DU versions, v1 and v2, to prevent them from concurrently executing real-time processing threads on the server's processor which would cause malfunctions. DU v2 is initialized using non-real-time threads. The update agent blocks real-time processing thread generation by DU v2 after initialization is complete. The update agent sets DU v1 processing threads to non-real-time and unblocks DU v2 to migrate data traffic between the versions. A layer 2 switch provides duplicated data packets to virtual network interface cards that are implemented for the respective DU versions and which share a common MAC (medium access control) address.