WAN Flow Distribution Using Dynamic Load Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing load balancing strategies in communication networks are static and do not adapt to dynamic traffic flows, leading to load imbalances and resource contention, which violate Service Level Agreements (SLAs) and result in suboptimal resource utilization.
Innovation Solution
A method for distributing data flows across multiple networks by measuring network loads over a reference period, calculating target and adjusted proportions using machine learning, and minimizing distribution divergence to ensure optimal load balancing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If static distribution strategies are used to simplify routing decisions, then device complexity is reduced, but adaptability to dynamic traffic patterns deteriorates
Solution Approach 1:
The patent implements dynamic load balancing by continuously monitoring network load metrics and adjusting flow distribution in real-time. The system transitions from static ECMP hashing to a dynamic controller that adapts distribution proportions based on current network conditions, resolving the contradiction between simplicity and adaptability.
Solution Approach 2:
The patent employs feedback mechanisms where the controller receives load information from network devices, processes this information, and adjusts flow distribution accordingly. This closed-loop feedback system enables the network to adapt to changing traffic patterns while maintaining manageable complexity through centralized control.
2Reliability
If homogeneous multi-path routing (ECMP) is used to ensure equal cost paths, then routing fairness is improved, but load balancing effectiveness deteriorates due to fixed distribution
Solution Approach 1:
The patent changes the distribution parameter from fixed equal-cost hashing to dynamic proportions based on actual network load. The controller adjusts the proportion of flows directed to each network path based on real-time load metrics, maintaining fairness while improving load balancing effectiveness through parameter adaptation.
Solution Approach 2:
The system transitions from static ECMP to dynamic flow distribution where the controller continuously adjusts path selection proportions based on network conditions. This dynamic approach maintains routing fairness through controlled randomness while significantly improving load balancing by adapting to actual network capacity.
3Productivity
If non-homogeneous multi-path routing is used to use more paths, then path utilization is improved, but distribution control precision deteriorates
Solution Approach 1:
The patent implements precise control through feedback mechanisms where the controller monitors actual flow distribution and network load, then adjusts distribution proportions to achieve target utilization. This feedback-based control maintains precision even when utilizing multiple paths with different characteristics.
Solution Approach 2:
The controller acts as an intermediary between flow sources and network paths, centralizing the decision-making process. This intermediary coordinates flow distribution across multiple paths with different costs and capacities, maintaining precise control through centralized management while maximizing path utilization.
4Manufacturing precision
If policy-based routing is used to direct flows to technical criteria, then flow distribution control is improved, but adaptability to heterogeneous traffic profiles deteriorates
Solution Approach 1:
The patent implements dynamic adaptation to heterogeneous traffic by continuously monitoring flow characteristics and network conditions. The controller adjusts distribution strategies in real-time based on traffic profile variations, maintaining precise control while adapting to diverse application requirements and traffic patterns.
Solution Approach 2:
The system dynamically changes distribution parameters based on traffic profile analysis. The controller adapts hashing parameters, distribution proportions, and selection criteria according to observed traffic characteristics, enabling precise control that adapts to heterogeneous traffic profiles rather than applying fixed policies.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention relates to a method for distributing data flows in a communication system comprising separate networks coupled to switching equipment. The method comprises the steps: • obtaining (100) a target distribution of incoming flows between the networks; • measuring (102) the flows of the networks in order to calculate the effective load of the networks over a reference period; • obtaining (104) an incoming load in the networks over a period following the reference period; • obtaining (106) a persistent proportion α of flows whose propagation time in the networks is greater than the reference period; • estimating (108) a distribution setpoint comprising an adjusted proportion for each network; • distributing (112), according to the adjusted proportions, the incoming load between the networks via the equipment.