Network Data Model Mapping via Semantic Matching Strength
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The networking industry faces challenges in managing and operating networks due to data model diversity, with vendors having proprietary models and standard bodies maintaining common models, leading to a labor-intensive and error-prone process for manual mapping, which is not easily achieved due to differences in labels and hierarchical organization of network data models.
Innovation Solution
The solution involves an automatic mapping of device data models through semantic matching using a prioritized and weighted matching strength threshold, employing a 'bottom-up approach' to identify matching nodes and filter out false positives, leveraging algorithms that generate representations of network data models and utilize lexical databases to compare labels and compute matching strengths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual mapping of network data models is performed, then mapping accuracy can be maintained, but labor intensity and time consumption increase significantly
Solution Approach 1:
The system performs automatic mapping of network data models without requiring manual intervention. The mapping module autonomously compares labels of leaf nodes between different data models, computes matching strengths, and generates mappings based on predefined thresholds, enabling the system to serve itself in the mapping task.
Solution Approach 2:
The patent replaces manual mechanical mapping processes with automated computational algorithms. The system uses algorithms to generate representations of data models, compare labels, compute matching strengths, and determine mappings, substituting human effort with automated information processing.
2Productivity
If automatic mapping is implemented, then productivity increases, but mapping precision may deteriorate due to false positives
Solution Approach 1:
The system implements feedback through threshold-based filtering. Mappings are generated only when the computed matching strength exceeds a predefined threshold, allowing the system to filter out false positives and maintain accuracy while operating automatically.
Solution Approach 2:
The patent uses parameter changes by adjusting the matching strength threshold to control the balance between productivity and precision. By setting appropriate thresholds, the system can tune its operation to achieve desired levels of both automation and accuracy.
3Measurement precision
If comprehensive label comparison is performed to ensure accuracy, then computational complexity increases
Solution Approach 1:
The patent segments the mapping process into distinct stages: generating representations of data models, comparing labels of leaf nodes, computing matching strengths, and generating mappings. This segmentation allows each stage to be handled independently with appropriate algorithms, managing overall complexity.
Solution Approach 2:
The system applies local quality by focusing comparison efforts on specific elements (leaf node labels) that are most critical for determining mapping accuracy. Rather than comparing entire data models uniformly, the algorithm concentrates computational resources on label comparisons where they matter most.
Data Source
AI summary
In one embodiment, a method includes processing network data models at a network device configured to operate in a network comprising one or more network components associated with one of the network data models, generating representations of the network data models, the representations comprising labels for elements in the network data models, comparing the labels associated with leaf nodes of the network data models to identify matching leaf nodes, comparing the labels associated with parent nodes of the matching leaf nodes to identify a strength of matching, and mapping at least two of the network data models at the network device based on the strength of matching for use in a network application. An apparatus and logic are also disclosed herein.