Uplink Grant Scheduling for Computation Resource Requests
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing radio communication systems fail to adequately consider computation resources when scheduling uplink transmissions for computationally intensive services like AI and ML, leading to service failures and suboptimal resource allocation.
Innovation Solution
Incorporating computation resource request information into the uplink grant process, allowing user equipment to report computation resource needs alongside traditional air interface requests, using mechanisms like MAC CEs and BSRs to ensure comprehensive resource scheduling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional uplink grant scheduling is used without computation resource consideration, then air interface resource allocation is simple and efficient, but computation resource allocation is insufficient leading to service failures
Solution Approach 1:
The patent combines air interface resource scheduling and computation resource scheduling into a unified uplink grant process. The network device now schedules both types of resources simultaneously through integrated signaling mechanisms, ensuring that computation resources are allocated alongside traditional radio resources, thereby improving service reliability without creating separate complex scheduling systems
Solution Approach 2:
The uplink grant mechanism is enhanced to serve multiple functions: it now simultaneously allocates air interface resources and computation resources. The scheduling system becomes multi-functional by incorporating computation resource request indication fields and buffer status reports that convey both types of resource needs, eliminating the need for separate scheduling procedures
2Measurement precision
If computation resource request information is integrated into uplink scheduling, then resource allocation accuracy improves, but signaling overhead increases
Solution Approach 1:
The patent merges computation resource request information with existing uplink scheduling signaling structures. By integrating computation resource indicators into the uplink grant and buffer status report mechanisms already present in the system, the patent achieves accurate resource allocation while minimizing additional signaling overhead through efficient reuse of existing message formats
Solution Approach 2:
The patent introduces computation resource request indication fields with optimized parameter structures. These fields use compact representations and efficient encoding schemes to convey computation resource needs accurately while maintaining minimal signaling overhead, balancing precision requirements with communication efficiency
3Productivity
If computation resources are allocated for AI/ML services, then service quality improves, but resource allocation complexity increases
Solution Approach 1:
The network device's scheduling function is extended to handle both traditional air interface resources and computation resources within a unified framework. This multi-functional scheduling approach enables efficient AI/ML service delivery by allocating computation resources alongside radio resources, improving service productivity without creating separate complex management systems
Solution Approach 2:
The patent combines computation resource management with existing radio resource management procedures. By integrating computation resource allocation into the uplink grant process and using unified scheduling mechanisms, the patent simplifies resource management while ensuring AI/ML services receive adequate computation resources, thereby improving service efficiency
Data Source
AI summary
A resource request information processing method includes: reporting computation resource request information, where the computation resource request information is configured for requesting computation resource from a network side.