Incremental Configlet Updates for Network Model Consistency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Managing configuration changes in complex networks is challenging due to large and complex device configuration databases, leading to substantial overhead in updating and reverting configurations, often resulting in changes being implemented directly in the actual network without proper modeling or documentation.
Innovation Solution
The system facilitates incremental updates by logging and managing each change as a 'configlet', which is processed independently and integrated into the overall configuration, allowing for easy roll-backs and reducing the overhead of maintaining a configuration database, with the same data used to update both the model and the actual system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the device configuration database is updated with each change, then the modeled network reflects the actual network, but the overhead of updating and recompiling the configuration database becomes substantial
Solution Approach 1:
The configuration database update process is segmented into incremental updates. Instead of recompiling the entire configuration database with each change, the system applies individual configlets (configuration snippets) independently. Each configlet represents a discrete configuration change that can be processed, simulated, and applied separately, reducing the overall update overhead while maintaining configuration consistency.
Solution Approach 2:
The system performs preliminary actions by maintaining a configuration database that is continuously updated with configlets before actual network changes are implemented. The configlets are logged, tracked, and prepared in advance, allowing for 'what-if' analysis and rollback capabilities without requiring substantial recompilation overhead when changes need to be reverted or modified.
2Ease of repair
If the configuration database is restored to a prior state, then roll-back is achieved, but identifying the particular changes associated with each backup is difficult
Solution Approach 1:
The system implements feedback mechanisms through configlet logging and tracking. Each configlet is assigned a unique identifier and is logged with metadata about the change it represents. This feedback system allows the network manager to track which configlets correspond to which configuration changes, making it easy to identify and revert specific changes without having to manually search through backup configurations.
Solution Approach 2:
The configlet acts as an intermediary between the configuration database and the network devices. Each configlet encapsulates a specific configuration change and maintains a record of what it modifies. This intermediary structure simplifies rollback by allowing the system to selectively remove or undo individual configlets rather than managing complex backup restoration processes.
3Productivity
If changes are made directly in the actual network without modeling, then the time required is reduced, but the overhead of keeping the configuration database up-to-date increases
Solution Approach 1:
The system enables self-service configuration management through the configlet mechanism. Network managers can create, modify, and apply configlets directly to the configuration database without requiring extensive manual updates or recompilation processes. The system automatically processes these configlets, updates the modeled network, and maintains consistency between the actual and modeled networks, reducing the maintenance overhead associated with direct changes.
Data Source
AI summary
In a network simulation system, a compiler is provided to support incremental updates to the configuration data associated with the modeled network. Each incremental change is identified and logged, to facilitate configuration management and select roll-backs to prior configurations. Because each update is processed and managed individually, and integrated automatically into the overall system configuration, the overhead associated with keeping a configuration database up-to-date is substantially reduced, thereby increasing the likelihood that all configuration changes will be reflected in the modeled network. In a preferred embodiment, the same data is used to incrementally update the configuration model and to execute the change in the actual system, thereby further reducing the overhead and assuring a correspondence between the modeled network and the actual network.


