Network Controller Intent-Graph Replay for Historical Fault Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network management systems lack the ability to replay and analyze historical network metrics and intent graphs, making it difficult to identify root cause faults and network issues outside of real-time operations.
Innovation Solution
A network controller is configured to store and manage a plurality of intent graphs and telemetry data over time, allowing it to generate metrics for a specific past time by selecting an appropriate intent graph and telemetry data subset, which can be analyzed to identify potential root cause faults and network performance issues.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If real-time metrics are generated using current intent graph and telemetry data, then network monitoring is available, but historical network issues cannot be analyzed
Solution Approach 1:
The system creates copies of intent graphs at different points in time and stores them in a database. When analyzing historical issues, the system retrieves the appropriate copied intent graph matching the historical time period and combines it with corresponding historical telemetry data, enabling accurate reconstruction and analysis of past network states without affecting current operations
Solution Approach 2:
The system performs preliminary actions by continuously capturing, storing, and versioning intent graphs and their associated telemetry data as they occur. This preliminary data collection and organization enables rapid historical analysis when needed, eliminating the time loss that would otherwise occur in investigating past network issues
2Adaptability or versatility
If intent graphs are stored for historical replay, then historical analysis capability is improved, but system complexity increases
Solution Approach 1:
The system segments the network state data by dividing intent graphs into distinct temporal versions, each associated with specific time-range telemetry data. This segmentation allows the database to store organized, queryable segments of historical network state rather than monolithic data structures, reducing management complexity while maintaining comprehensive historical analysis capability
Solution Approach 2:
The intent graph database is designed with multi-functionality to serve both real-time network operations and historical analysis purposes. The same database infrastructure supports current network state management while simultaneously enabling historical metrics replay and root cause analysis, reducing overall system complexity by consolidating functions rather than creating separate systems
3Measurement precision
If complete telemetry data is stored for all network devices, then analysis accuracy is improved, but data storage requirements increase
Solution Approach 1:
The system extracts only the relevant telemetry data corresponding to specific time ranges and associates it with the appropriate intent graph versions. Rather than storing and processing all possible telemetry data continuously, the system extracts and stores only the portions needed for historical replay and analysis, maintaining measurement precision while reducing overall data storage requirements
Solution Approach 2:
The system implements partial action by storing telemetry data selectively based on relevance to intent graph changes and potential analysis needs. Rather than comprehensively storing every telemetry point indefinitely, the system retains data at the appropriate level of detail for historical analysis without excessive redundancy, balancing accuracy with storage efficiency
Data Source
AI summary
A method includes storing, by one or more processors, a plurality of intent graphs for a network and receiving, the by one or more processors, a query indicating a time. The method further includes determining, by the one or more processors, a subset of telemetry data that corresponds to the time range using the time indicated by the query. The method further includes generating, by the one or more processors and based on an intent graph of the plurality of intent graphs that is associated with a time range that includes the time indicated by the query and based on the subset of the telemetry data, one or more metrics and outputting, by the one or more processors, an indication of the one or more metrics.