Latency Critical Application Resource Allocation in Wireless Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current wireless network schedulers face challenges in providing low latency for latency-critical applications due to variance in mean latency and jitter, which is exacerbated by best effort and fair scheduling approaches, leading to unsatisfactory performance in scenarios like autonomous vehicles and real-time gaming.
Innovation Solution
A method that provisions latency-critical applications by determining and assigning reference point values for mean latency and throughput, using functions to calculate actual point values and compare them with reference values, optimizing resource allocation in real-time to adapt to current transmission conditions, ensuring prioritization of resources for latency-critical applications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If best effort and fair scheduling approaches are used in wireless networks, then resource allocation is simplified and equitable, but latency variance and jitter increase making latency-critical applications unsatisfactory
Solution Approach 1:
The patent implements dynamic scheduling that adapts to real-time conditions by continuously monitoring latency requirements and adjusting resource allocation. The scheduler transitions from static fair scheduling to dynamic priority-based scheduling based on application latency sensitivity, achieving both operational simplicity and latency consistency.
Solution Approach 2:
The system changes scheduling parameters dynamically based on application requirements. By adjusting scheduling weights, time slots, and resource block allocations according to latency criticality, the system maintains simplicity while achieving consistent low latency for critical applications.
2Reliability
If resources are allocated dynamically based on real-time conditions, then latency for latency-critical applications is reduced, but system complexity increases
Solution Approach 1:
The patent implements self-service mechanisms where applications declare their own latency requirements and the system automatically adjusts scheduling without complex manual configuration. The scheduler autonomously monitors performance metrics and adapts resource allocation, reducing the perceived complexity for users while maintaining high latency performance.
Solution Approach 2:
The system employs feedback loops that continuously monitor latency performance and automatically adjust scheduling decisions. This closed-loop control simplifies the overall system by using straightforward feedback mechanisms rather than complex predictive models, achieving low latency while managing complexity through adaptive response to measured conditions.
3Loss of time
If edge computing is deployed to process data near end users, then latency is reduced for real-time applications, but infrastructure cost and device complexity increase
Solution Approach 1:
The patent segments the network infrastructure into edge computing nodes distributed near end users and centralized data centers. This segmentation allows latency-critical processing to occur locally at edge nodes while non-critical functions remain centralized, reducing transmission latency for time-sensitive operations while managing overall infrastructure complexity through modular architecture.
Solution Approach 2:
The edge computing nodes are designed with multi-functionality, serving both latency-critical applications and standard data processing needs. This universality reduces overall infrastructure complexity by using a single distributed edge network for multiple purposes rather than requiring separate specialized systems for different application types.
Data Source
AI summary
A method includes: a) provisioning at least one latency critical application; b) determining mean latency and mean throughput which are required by a respective latency critical application, and assigning the determined mean latency and mean throughput to a reference point value; c) allocating the reference point value to the respective latency critical application; d) calculating via a time variable point value function an actual point value for the respective latency critical application; e) comparing the reference point value with the actual point value; f) determining a difference value between the reference point value and the actual point value; g) repeating steps d) to f) for a subset of points in time within the time interval; h) summing up all difference values determined in step f) for all points in time of the subset; and i) continuously optimizing current use of resources in the cell.