Adaptive SD-WAN Policies via Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing SD-WAN systems rely on global policies that do not account for site-specific differences in network hardware, WAN link performance, and WAN link costs, leading to suboptimal network path selection and traffic steering.
Innovation Solution
A machine-learning engine receives WAN link characterization data from SD-WAN edge devices and routers, using this data to automatically generate or update local policies for each site, optimizing network performance and cost based on site-specific conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If global policies are used for all sites, then system complexity is reduced and ease of operation is improved, but network performance and cost optimization deteriorate due to lack of site-specific adaptation
Solution Approach 1:
The patent applies local quality by generating site-specific local policies tailored to each location's unique network conditions, hardware characteristics, and cost structures. Instead of applying a uniform global policy across all sites, the system customizes policies for each site based on local WAN link performance metrics, available bandwidth, latency characteristics, and operational costs, thereby optimizing network performance and cost efficiency for each specific site while maintaining centralized management capabilities.
2Productivity
If site-specific local policies are generated, then network performance and cost optimization are improved, but device complexity and data processing requirements increase
Solution Approach 1:
The patent introduces an intermediary policy generation system that acts as a mediator between centralized global policy definitions and distributed site-specific implementations. This intermediary component automatically generates localized policies by processing global policy templates against site-specific network characteristics and performance data, thereby eliminating the need for manual policy configuration at each site while still achieving optimized, location-specific network performance and cost management.
Solution Approach 2:
The system implements self-service by enabling automatic generation of site-specific policies through machine learning models that process network performance data and independently determine optimal policy configurations for each site. The system autonomously analyzes local network conditions, evaluates multiple policy options, and generates optimized policies without requiring manual intervention from network operators, thereby reducing operational complexity while achieving superior network performance.
3Adaptability or versatility
If machine learning models are used to generate local policies, then adaptability to site-specific conditions is improved, but computational requirements and processing time increase
Solution Approach 1:
The patent applies preliminary action by pre-training machine learning models offline using historical network performance data and site characteristics. The models are prepared in advance with learned patterns and relationships between network conditions and optimal policy configurations. During runtime, the pre-trained models rapidly infer site-specific policies from current network data without requiring intensive real-time computation, thereby achieving high adaptability to local conditions while minimizing computational resource consumption and processing time during actual policy generation.
Data Source
AI summary
An example method includes receiving, by an SD-WAN system, WAN link characterization data for a plurality of WAN links of the SD-WAN system over a time period; and for each site of a plurality of sites of the SD-WAN system, generating, by the SD-WAN system, a local policy for the site, wherein generating the local policy is based on a machine learning model trained with the WAN link characterization data for the plurality of WAN links, and providing the local policy to an SD-WAN edge device of the site.