Network Traffic Allocation via Dynamic Fairness Parameter
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing traffic engineering techniques face challenges in balancing throughput and fairness in bandwidth allocation, as they often conflict with each other, making it difficult to maximize network resource utilization while ensuring fair allocation to all users.
Innovation Solution
A method and system using a server to determine max-min fairness allocation through binary and linear search, allowing for flexible trade-offs between efficiency and fairness by establishing bounds for fairness degradation, thereby optimizing resource allocation across multiple commodities and users.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If max-min fairness allocation is used to ensure fair bandwidth distribution among all users, then fairness is improved, but network throughput and resource utilization deteriorate
Solution Approach 1:
The patent implements dynamic fairness control by allowing the fairness parameter alpha to be adjusted based on network conditions and user requirements. The system transitions from static max-min fairness to a dynamic framework where the fairness-throughput tradeoff can be optimized in real-time, enabling the network to adapt between fairness-oriented and throughput-oriented operation modes.
Solution Approach 2:
The patent introduces a parameter alpha (0 ≤ alpha ≤ 1) that controls the fairness-throughput tradeoff. By changing this parameter, the system can adjust the allocation strategy: when alpha=1, max-min fairness is achieved; when alpha=0, throughput optimization is achieved. This parameter change enables continuous adjustment between the two opposing objectives.
2Productivity
If network resources are maximally utilized to improve throughput, then productivity is improved, but fairness of resource allocation deteriorates
Solution Approach 1:
The system enables dynamic adjustment of resource allocation strategies by modifying the fairness parameter alpha. When high throughput is required, alpha can be reduced to prioritize network-wide efficiency. When fairness is critical, alpha increases to protect minimum allocations. This dynamic control resolves the contradiction by allowing operational flexibility.
Solution Approach 2:
By introducing the controllable parameter alpha in the utility function, the patent transforms the rigid max-min fairness approach into a flexible framework. The parameter change allows the system to optimize for throughput when resources are abundant while maintaining fairness guarantees when resources are constrained, thus resolving the throughput-fairness contradiction.
3Measurement precision
If traditional linear programming is used to calculate fair allocation, then accuracy is improved, but calculation time and complexity increase
Solution Approach 1:
The patent segments the calculation process into two phases: a binary search phase that quickly narrows down the fairness parameter range, and a linear search phase that refines the solution. This segmentation avoids the computational burden of traditional linear programming while maintaining allocation accuracy, significantly reducing calculation time for large-scale networks.
Solution Approach 2:
The binary search approach allows the system to skip through large portions of the parameter space efficiently, rushing to the optimal fairness parameter value without examining every possible allocation combination. This skipping mechanism dramatically reduces calculation time compared to exhaustive linear programming methods while preserving solution accuracy.
Data Source
AI summary
Systems and methods are provided for allocating resources of a network among a plurality of traffic demands to optimize fairness and throughput. Fairness constraints and traffic commodity importance can be specified to control allocation.