EVPN Endpoint Debugging Using Intent Graph Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing EVPN networks face challenges in diagnosing traffic flow failures due to misconfigurations and hardware/software bugs, which require extensive knowledge across various vendor implementations and result in inefficient, resource-intensive debugging processes.
Innovation Solution
A network management system (NMS) uses intent graph data to identify network anomalies by comparing endpoint information from network devices, such as MAC address tables, to detect anomalous devices and perform automated diagnosis, reducing resource consumption and scaling to large networks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional per-device diagnosis methods are used to check for configuration consistency and network information consistency, then local device anomalies can be detected, but comprehensive network anomalies cannot be identified and extensive knowledge across various vendor implementations is required
Solution Approach 1:
The patent implements a universal debugging system that aggregates endpoint information from multiple network devices of different vendors into a unified data structure. The comparison module universally compares endpoint information across all devices using consistent criteria, eliminating the need for vendor-specific debugging knowledge while maintaining comprehensive anomaly detection capability.
Solution Approach 2:
The patent introduces a centralized debugging system that acts as an intermediary between network devices and analysts. This system collects endpoint information from various devices, performs automated comparison and anomaly detection, and presents results in a unified format, thereby simplifying the debugging process and reducing the complexity of direct per-device analysis.
2Measurement precision
If comprehensive endpoint information is collected and compared across all network devices to identify anomalies, then accurate network-wide anomaly detection is achieved, but resource consumption increases
Solution Approach 1:
The patent extracts only the essential endpoint information (MAC address tables and associated metadata) from network devices for comparison, rather than collecting and analyzing complete device configurations or all operational parameters. This selective extraction maintains anomaly detection accuracy while significantly reducing the volume of data that requires processing and comparison.
Solution Approach 2:
The patent implements partial comparison by focusing only on specific endpoint information fields that are most indicative of network anomalies, rather than performing exhaustive comparison of all device parameters. This approach achieves sufficient detection accuracy with reduced computational resources compared to complete device-state analysis.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A system includes a storage device configured to store intent graph data associated with a plurality of virtual networks and a plurality of network devices and processing circuitry in communication with the storage device. The intent graph data identifies a subset of the plurality of virtual networks associated with each network device. The processing circuitry is configured to determine whether the intent graph data indicates a first network device and a second network device are configured for a common virtual network, in response to determining that the graph indicates the first network device and the second network device are configured for the common virtual network, determine a network anomaly based on a comparison of first endpoint information obtained from the first network device and second endpoint information obtained from the second network device, and output an indication of the network anomaly.