Network Testing System Using Normalized Feature Tagging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for testing computer network functionality are inadequate in accurately determining the performance status of network nodes, as they fail to effectively combine and analyze diverse data types from nodes with varying characteristics, leading to inefficiencies in identifying normal or abnormal network behavior.
Innovation Solution
A computer-implemented method and system that determines limit values for each data type, normalizes test data, combines node features, and uses machine learning to tag performance status, enabling the system to assess network functionality by training on tagged data sets until a success criterion is met.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional network testing methods are used, then the testing process is simple, but the accuracy of determining network performance status is insufficient
Solution Approach 1:
The testing system segments network data into multiple feature types (configuration features, performance features, network features) and processes each type separately through dedicated normalization modules before combining them for comprehensive analysis
Solution Approach 2:
A machine learning model serves as an intermediary between raw network data and performance status determination, learning complex patterns and relationships that traditional methods cannot capture, thereby improving measurement precision
2Measurement precision
If diverse data types from nodes with varying characteristics are analyzed, then the accuracy of performance determination improves, but the difficulty of data processing increases
Solution Approach 1:
The system applies different normalization techniques tailored to each data type and node characteristic, using node-specific limit values and normalization parameters that adapt to local data properties rather than applying uniform processing
Solution Approach 2:
The system dynamically adjusts normalization parameters including limit values, scaling factors, and offset values based on node characteristics and data type properties, transforming diverse data into a unified format suitable for machine learning analysis
3Adaptability or versatility
If normalization based on node characteristics is implemented, then the system accommodates diverse node capabilities, but the computational overhead increases
Solution Approach 1:
The system performs preliminary normalization of node data before feeding it to the machine learning model, pre-processing configuration, performance, and network features to reduce the computational burden during real-time inference and improve overall processing efficiency
Data Source
AI summary
Method of training and using a computerized system for testing a network that includes computer nodes, to determine functionality of the network, including: determining respective limit values for each of a plurality of data types; obtaining at least one sample of test data of a given data type, converting the sample value into a corresponding normalized sampled value, so that each normalized sampled value is within limit values of the given data type determined based on at least the characteristics of the given data type and the given node; combining normalized values of different types of the given node into a set of combined node features; tagging each set of combined node features with a performance tag; generating a training set that includes a plurality of tagged sets of combined features that pertain to at least one node; and inputting the training set to train the system.