AS-Level Bottleneck Detection for Application-Aware Traffic Rerouting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current SD-WAN systems lack the ability to identify and avoid bottlenecks within autonomous systems, leading to suboptimal routing decisions that affect the quality of service for applications, particularly in dynamic network environments with varying performance characteristics.
Innovation Solution
A supervisory service that collects quality of experience metrics for online applications, maps sessions to paths traversing autonomous systems, identifies bottlenecks, and reroutes traffic to avoid these systems, using machine learning and path tracing to predict and prevent SLA failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If end-to-end path metrics are used for routing decisions, then routing automation is achieved, but the ability to identify and avoid bottlenecks in specific autonomous systems is lost
Solution Approach 1:
The patent segments the end-to-end path into individual autonomous system segments. Instead of treating the entire path as a single metric unit, the system breaks it down into discrete AS segments that can be individually evaluated. This segmentation enables the routing system to identify which specific AS contains bottlenecks while maintaining automated routing decisions.
Solution Approach 2:
The patent adds a new dimension to routing metrics by introducing autonomous system-level measurements alongside traditional end-to-end path metrics. This multi-dimensional approach allows the system to simultaneously maintain routing automation and achieve bottleneck identification precision by evaluating performance at both the path level and the AS level.
2Ease of operation
If the entire path is monitored without AS-level granularity, then routing decisions are simplified, but the ability to explore and utilize alternative paths excluding hotspots is limited
Solution Approach 1:
By segmenting the network path into autonomous system components, the system maintains simplified routing decisions at the overall path level while enabling detailed bottleneck analysis at the segment level. This allows the routing system to easily select paths based on AS-level bottleneck information without complex manual intervention.
Solution Approach 2:
The patent implements feedback mechanisms that provide bottleneck information from monitored paths back to the routing decision system. This feedback loop enables the system to automatically adapt and explore alternative paths that avoid identified hotspots, enhancing path exploration capability while keeping the operation simple through automated responses.
3Device complexity
If traditional routing systems operate without application-aware information, then system complexity is reduced, but quality of service for specific applications is degraded
Solution Approach 1:
The patent implements a universal bottleneck detection mechanism that operates across multiple autonomous systems and application types without requiring application-specific customization. This multi-functional approach maintains relatively simple system architecture while improving quality of service for various applications by identifying and avoiding bottlenecks that affect them.
Solution Approach 2:
The patent introduces an intermediary bottleneck detection and analysis layer between the traditional routing system and the applications. This intermediary component provides application-aware quality of service improvements by identifying AS-level bottlenecks and guiding routing decisions, without requiring direct integration complexity between the routing system and each application.
Data Source
AI summary
In one embodiment, a supervisory service for a network obtains quality of experience metrics for application sessions of an online application. The supervisory service maps the application sessions to paths that traverse a plurality of autonomous systems. The supervisory service identifies, based in part on the quality of experience metrics, a particular autonomous system from the plurality of autonomous systems associated with a decreased quality of experience for the online application. The supervisory service causes application traffic for the online application to avoid the particular autonomous system.