Network Path Clustering for High-Predictability Routing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing SD-WAN systems struggle with inconsistent path predictability, leading to reactive routing decisions and high SLA failure rates due to the dynamic nature of network paths and varying QoS characteristics, which are not effectively addressed by current predictive routing methods.
Innovation Solution
Implementing a predictive application aware routing engine that uses clustering to group network paths with similar characteristics into clusters, determining a predictability metric for each cluster, and enabling predictive routing based on these metrics to optimize resource usage and reduce SLA failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If predictive routing is applied to all network paths, then routing predictability should improve, but resource consumption increases and SLA failures persist due to varying path predictability
Solution Approach 1:
The patent segments network paths into distinct clusters based on their predictability characteristics. By grouping paths with similar predictability profiles together, the system can apply predictive routing selectively to high-predictability clusters while avoiding low-predictability paths, thereby improving overall routing predictability without uniformly consuming resources across all paths.
Solution Approach 2:
The patent implements local quality by assigning different routing strategies to different path clusters based on their specific predictability characteristics. High-predictability clusters receive predictive routing treatment while low-predictability clusters use alternative routing methods, optimizing resource consumption according to the local quality of each path group's predictability profile.
2Productivity
If predictive routing is applied uniformly across all paths, then routing decisions should be more proactive, but SLA failures increase due to varying path characteristics
Solution Approach 1:
The patent segments paths into clusters with distinct predictability profiles, allowing the system to identify which path groups are suitable for predictive routing. This segmentation enables proactive routing decisions to be applied only where they will be effective, avoiding SLA failures that would result from attempting predictive routing on unsuitable paths.
Solution Approach 2:
The patent changes the parameter of routing strategy selection based on path cluster characteristics. By adjusting which routing approach is applied to each cluster based on their predictability parameters, the system optimizes both the effectiveness of routing decisions and compliance with SLA requirements.
3Use of energy by moving object
If clustering is applied to group network paths, then resource consumption is reduced, but system complexity increases
Solution Approach 1:
The patent merges multiple network paths into consolidated clusters based on their predictability characteristics. This merging reduces the number of individual path evaluations required, thereby reducing resource consumption while the clustering framework provides a structured approach to managing the complexity of path groupings.
Data Source
AI summary
In one embodiment, a device forms a plurality of clusters of network paths used to convey traffic for an online application by applying clustering to telemetry data for those network paths. The device determines a predictability metric for a particular cluster in the plurality of clusters. The device provides an indication of the predictability metric for the particular cluster for display. The device enables, based in part on the predictability metric, predictive routing for the network paths in the particular cluster.