Adaptive WAN Load Balancing Using Real-Time Telemetry
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing load balancing techniques, such as ECMP and UCMP, lack adaptiveness and flexibility in responding to real-time network conditions, leading to potential traffic congestion and bandwidth bottlenecks due to their static bandwidth assumptions and inability to dynamically adjust to changing network conditions.
Innovation Solution
A dynamic load balancing technique that continuously monitors real-time network telemetry data across multiple paths to optimize traffic flow, selecting the best path based on metrics like latency, jitter, and bandwidth availability, ensuring efficient distribution of traffic and avoiding congestion and bottlenecks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If static load balancing techniques (ECMP/UCMP) are used, then device complexity is reduced and ease of operation is improved, but adaptability to real-time network conditions deteriorates and traffic congestion occurs
Solution Approach 1:
The system performs preliminary actions by continuously collecting network telemetry data (bandwidth, latency, packet loss) before making load balancing decisions. This advance gathering of performance metrics enables the system to proactively identify optimal paths and adjust traffic distribution before congestion occurs, rather than reacting to failures after they happen.
Solution Approach 2:
The patent implements a feedback mechanism where network performance telemetry data is continuously monitored and fed back to the load balancing algorithm. This feedback loop allows the system to dynamically adjust path selection and traffic distribution based on real-time network conditions, automatically adapting to changing bandwidth availability, latency, and congestion levels without manual intervention.
2Productivity
If static bandwidth assumptions are made, then calculation complexity is reduced, but productivity in utilizing available bandwidth deteriorates
Solution Approach 1:
The load balancing system performs self-service by automatically collecting its own performance telemetry data from the network paths it manages. Rather than relying on external monitoring systems or manual configuration, the system autonomously measures bandwidth, latency, and packet loss metrics, then uses this self-generated data to dynamically optimize traffic distribution and maximize bandwidth utilization.
Solution Approach 2:
The patent dynamically changes load balancing parameters based on real-time telemetry data. Instead of using fixed bandwidth assumptions, the system continuously adjusts path selection criteria, traffic distribution ratios, and path weighting parameters according to measured network conditions such as available bandwidth, latency variations, and packet loss rates, thereby optimizing productivity.
3Reliability
If dynamic path selection based on telemetry data is implemented, then network performance and reliability are improved, but device complexity and computational requirements increase
Solution Approach 1:
The system applies partial action by implementing dynamic load balancing selectively for specific traffic flows or path combinations rather than uniformly across all network traffic. This approach allows the system to achieve reliability improvements for critical paths while avoiding the full computational overhead of analyzing and optimizing every possible path for every packet, thereby managing device complexity more effectively.
Data Source
AI summary
Systems and methods for load balancing in a network are disclosed. An illustrative method includes receiving network telemetry data corresponding to network paths of a plurality of coexisting multipaths, performing an adaptive load balancing process by determining whether a network path from the plurality of coexisting multipaths is an adequate network path based on the network telemetry data, and in response to determining the network path is an adequate network path, selecting the network path for a network flow.