Fair Hierarchical QoS Marking with Throughput-Value Matrices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network resource management techniques, particularly in HQoS scheduling, face challenges with computational complexity, memory demands, and configuration complexity, especially in managing heterogeneous traffic mixes and dynamic bottlenecks in modern networks, leading to unfairness and inefficiencies in resource sharing.
Innovation Solution
Implementing a Throughput-Value Function (TVF) that maps throughput values to packet values for subflows, using a matrix-based approach to periodically update resource sharing weights and region determination, allowing for efficient packet marking that ensures weighted fairness among subflows without requiring scheduler modifications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If HQoS scheduling is implemented to manage heterogeneous traffic mixes, then resource sharing control and QoS management are improved, but computational complexity and memory demands increase
Solution Approach 1:
The patent segments the complex HQoS scheduling problem into two independent components: (1) packet marking at the edge router using HPPV algorithms, and (2) simple packet value-based scheduling at bottleneck routers. This segmentation allows the complex resource sharing policies to be defined once at the edge without burdening the bottleneck routers with computational complexity, while still achieving effective HQoS management.
Solution Approach 2:
The patent extracts the complex policy decision-making function from the bottleneck routers and relocates it to the edge router. The edge router performs the computationally intensive HPPV packet marking, while bottleneck routers only need to perform simple comparisons and scheduling based on packet values. This extraction eliminates the computational complexity burden from bottleneck routers while preserving resource sharing control capabilities.
2Productivity
If packet marking is used to implement resource sharing policies, then bandwidth sharing control is improved, but configuration complexity increases
Solution Approach 1:
The HPPV packet marking algorithm automatically adapts to changing traffic conditions and dynamically adjusts packet values based on observed throughput and fairness requirements. This self-service capability eliminates the need for manual configuration of packet marking parameters, as the system automatically learns and adjusts to optimize bandwidth sharing control without increasing configuration complexity.
Solution Approach 2:
The patent changes the configuration approach from static parameter settings to dynamic parameter adaptation. The HPPV algorithm continuously adjusts packet marking parameters based on real-time traffic conditions, eliminating the need for complex manual configuration while maintaining effective bandwidth sharing control. The system adapts parameters automatically rather than requiring operators to set and tune them.
3Reliability
If static reservation is implemented to ensure minimum QoS, then QoS guarantee is improved, but resource utilization decreases
Solution Approach 1:
The patent replaces static reservation with dynamic packet marking using HPPV algorithms. Instead of pre-allocating fixed bandwidth, the system dynamically adjusts packet values based on current traffic conditions, observed throughput, and fairness requirements. This dynamic approach ensures minimum QoS guarantees when needed while automatically increasing resource utilization when traffic conditions allow, eliminating the waste inherent in static reservation.
Solution Approach 2:
The HPPV packet marking mechanism incorporates feedback from observed traffic patterns and throughput measurements to continuously adjust packet values. This feedback loop ensures that QoS guarantees are maintained when necessary while optimizing resource utilization by allowing more aggressive bandwidth allocation when traffic conditions permit, thus resolving the trade-off between reliability and productivity.
Data Source
AI summary
A computing device (150) identifies a plurality of regions of a Throughput Value Function, TVF (200). The TVF (200) maps throughput values to packet values for a plurality of subflows. The computing device (150) periodically updates a plurality of matrices comprising a matrix of resource sharing weights and a region determination matrix. The computing device (150) marks a packet (180) with a packet value based on the plurality of matrices. The packet (180) is received in one of the subflows.