6G Compute Offloading Control for Dynamic Workload Migration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing cellular networks lack efficient mechanisms for dynamic workload migration and augmented computing, particularly in next-generation networks like 6G, which require flexible and scalable computing solutions to handle diverse computing tasks with varying resource requirements and dependencies across devices and network edges.
Innovation Solution
Implementing a RAN-based compute offloading system with compute control clients and service functions, utilizing user plane transport and dynamic session establishment procedures to manage computing tasks, enabling collocated and non-collocated scenarios, and supporting IP and Non-IP based radio interfaces for efficient workload distribution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If computing tasks are offloaded to network edges, then latency is reduced and computing capabilities are enhanced, but network complexity and resource management difficulty increase
Solution Approach 1:
The patent segments computing tasks into different categories (compute-intensive, data-intensive, real-time) and segments network resources into multiple edge nodes with different capabilities. This allows selective offloading based on task requirements while managing network complexity through structured resource allocation.
Solution Approach 2:
The system dynamically adjusts computing task distribution based on real-time network conditions, device capabilities, and workload requirements. Computing tasks are dynamically migrated between edge nodes and cloud centers, and resource allocation is continuously optimized to balance latency reduction with network complexity management.
2Adaptability or versatility
If dynamic workload migration is enabled, then adaptability and resource efficiency improve, but system complexity and management overhead increase
Solution Approach 1:
The system changes key parameters such as task priority, resource allocation ratios, and migration thresholds based on network conditions and device capabilities. This enables adaptive workload management through parameter adjustment rather than complex reconfiguration, reducing management overhead while maintaining high adaptability.
Solution Approach 2:
The patent implements feedback mechanisms where network conditions, device battery status, and task execution performance are continuously monitored and used to adjust workload distribution decisions. This closed-loop control enables dynamic adaptation while simplifying management through automated feedback-driven optimization.
3Adaptability or versatility
If heterogeneous computing capabilities are leveraged, then computing versatility and scalability improve, but resource management complexity and coordination difficulty increase
Solution Approach 1:
The patent creates a universal resource management framework that can handle multiple computing task types and device categories through standardized interfaces and protocols. This multi-functional approach enables the system to manage heterogeneous computing capabilities uniformly, reducing coordination complexity while maintaining high versatility.
Solution Approach 2:
The system introduces intermediary layers such as network-side compute functions and edge computing platforms that act as mediators between diverse devices and computing resources. These intermediaries abstract the heterogeneity of computing capabilities, simplifying resource management while enabling versatile computing scenarios.
4Power
If compute-intensive tasks are executed at network edges, then latency reduction and computing performance improve, but energy consumption and infrastructure requirements increase
Solution Approach 1:
The patent applies local quality by deploying compute resources at different locations (edge nodes closer to users vs. centralized cloud centers) based on task requirements. Compute-intensive tasks are executed at edges with appropriate computing capabilities, while less demanding tasks are handled at cloud centers, optimizing the energy-performance tradeoff through localized resource utilization.
Solution Approach 2:
The system performs partial computing at the edge rather than complete offloading, depending on task characteristics and edge capabilities. This partial action approach reduces the energy burden on edge infrastructure while still achieving significant latency reduction for critical tasks, balancing performance improvement with energy consumption.
Data Source
AI summary
Various embodiments generally may relate to the field of wireless communications. For example, some embodiments may relate to enabling augmented computing as a service or network capability for sixth-generation (6G) networks. For example, some embodiments may be directed to the operation of a compute control client (Comp CC) at the UE side, and a compute control function (Comp CF) and compute service function (Comp SF) at the network side, which are referred to herein as “compute plane” functions to handle computing related control and user traffic.