Segment Routing Optimization Across Multiple Traffic Matrices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing routing schemes in networks are inflexible and fail to account for varying traffic patterns, leading to congestion due to insufficient robustness in handling different anticipated traffic demands.
Innovation Solution
A machine learning-based approach that utilizes a ML model to determine a single routing scheme optimized for a network by transforming deflection parameters into a non-linear domain and normalizing them, allowing unconstrained optimization of link utilization across multiple traffic matrices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional routing schemes are used, then the routing setup is simple, but the network becomes congested when traffic patterns vary
Solution Approach 1:
The patent transforms deflection parameters into a non-linear domain using exponential functions, allowing the ML model to learn optimized routing parameters that handle varying traffic patterns. This parameter transformation enables the system to achieve robust routing performance across different traffic conditions while maintaining manageable complexity through automated learning.
Solution Approach 2:
The patent replaces traditional mechanical routing algorithms with a machine learning-based approach. The ML model takes traffic matrices as input and automatically learns optimal routing schemes by minimizing link utilization, substituting complex manual routing configuration with automated intelligent decision-making that adapts to varying traffic patterns.
2Productivity
If a single routing scheme is used for all traffic patterns, then the routing configuration is simple, but link utilization is not optimized under varying traffic demands
Solution Approach 1:
The patent implements a dynamic routing scheme where the ML model adapts the routing configuration based on input traffic matrices. The system dynamically adjusts deflection parameters and routing decisions according to varying traffic demands, enabling optimized link utilization across different traffic patterns rather than using a static fixed routing scheme.
Solution Approach 2:
The transformation of deflection parameters into the non-linear domain allows the ML model to capture complex traffic patterns and optimize routing parameters accordingly. This enables the system to adapt to varying traffic demands while maintaining a unified routing framework, improving both productivity and adaptability.
3Reliability
If deflection parameters are used directly in the linear domain, then the optimization is constrained, but transforming to non-linear domain increases computational complexity
Solution Approach 1:
The patent applies exponential transformation to convert linear domain deflection parameters into the non-linear domain, enabling unconstrained optimization of the loss function. This parameter transformation allows the ML model to explore a wider solution space and find optimal routing configurations that would be inaccessible in the linear domain, despite the increased computational complexity being managed through automated learning processes.
Data Source
Figure 1A
Figure 1B
Figure 1C
AI summary
In some embodiments, there may be provided a method that includes receiving a first traffic matrix; receiving information regarding links associated with each segment of the network; determining a total amount of segment flow using the at least one non-linear deflection parameter applied to the traffic demand of the first traffic matrix; determining a link flow for each of the links using the total amount of segment flow and the second input to the machine learning model; determining link utilization for each of the links using the link flows and a capacity for each of the links; learning, by the machine learning model using a gradient descent, a minimum of a maximum amount of the link utilization over the links by at least adjusting a value of the at least one non-linear deflection parameter. Related systems, methods, and articles of manufacture are also disclosed.