SD-WAN Uplink Selection via Path Quality Thresholds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In Software Defined Wide Area Networks (SD-WANs), the naive round-robin approach to distributing traffic across multiple uplinks does not account for varying usage patterns, priority of connections, and dynamic uplink behaviors, leading to inefficient bandwidth usage and potential overloading of certain uplinks.
Innovation Solution
The implementation of a method that uses deep packet inspection (DPI) to determine criticality scores for applications, classifies them, and calculates path quality threshold scores for each uplink based on link health and tolerance levels, allowing for the selection of primary, secondary, and standby uplinks for each application category, ensuring balanced traffic distribution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If naive round-robin approach is used to distribute traffic across multiple uplinks, then traffic distribution is simple to implement, but bandwidth usage becomes inefficient and certain uplinks may be overloaded
Solution Approach 1:
The system changes the parameters of traffic distribution by introducing application criticality scores, uplink path quality scores, and tolerance levels. Instead of simple round-robin, the system dynamically adjusts traffic routing based on these parameters to optimize bandwidth usage while maintaining ease of operation through automated scoring and classification mechanisms
Solution Approach 2:
The system implements dynamic traffic distribution where uplink selections are not fixed but adapt based on real-time conditions. The network orchestrator continuously evaluates path quality scores and tolerance levels to dynamically route traffic, allowing the system to respond to changing network conditions while maintaining simple operation through centralized control
2Productivity
If criticality scores and path quality threshold scores are calculated for each application and uplink, then bandwidth usage is optimized and network efficiency improves, but system complexity increases
Solution Approach 1:
The network orchestrator serves as an intermediary that centralizes the complex calculations of criticality scores and path quality threshold scores. By placing this computational burden in a centralized controller rather than distributed across all network devices, the system achieves optimized bandwidth usage while managing complexity through centralized coordination
Solution Approach 2:
The system segments the complexity by dividing it into distinct components: application criticality scoring, uplink path quality scoring, tolerance level definition, and traffic routing decisions. Each component can be independently configured and managed, allowing the overall complex system to be broken down into manageable segments that collectively optimize network efficiency
3Reliability
If multiple uplinks are used for high availability and load sharing, then network reliability improves, but uplink selection without proper classification leads to unequal traffic distribution and potential overloading
Solution Approach 1:
The system applies local quality by assigning different criticality scores to different application categories and different path quality scores to different uplinks. This allows traffic distribution to be tailored to the specific requirements of each application and the characteristics of each uplink, ensuring both high availability through multiple uplinks and balanced load distribution through localized quality assessments
Data Source
AI summary
An example non-transitory, computer-readable medium includes instructions that cause a device to determine, for uplinks of a branch gateway, a link health baseline. The instructions further cause the device to determine, for a set of criticality classes, a class link health baseline for each link health baseline, based on the link health baseline and a tolerance level of each criticality class. The instructions further cause the device to calculate, based in part on weighted parameters of the class link health baselines and an uplink cost, a path quality threshold score for each application category and for each uplink. The instructions further cause the device to select, for each application category, a primary uplink and a secondary uplink based on the path quality threshold scores. The instructions further cause the device to route network traffic through the primary uplink of the application category assigned to the network traffic.