Automated Database Data Classification for Security
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing information security systems face challenges in accurately and proactively classifying sensitive data in databases due to subjective decision-making, employee turnover, flexible job assignments, and the labor-intensive nature of data classification, leading to inconsistent or nonexistent classifications.
Innovation Solution
The system proactively monitors databases for changes, extracts newly created or modified data columns, and applies data classifications by matching their attributes to pre-trained classification models, automatically assigning classifications and protecting sensitive information from unwanted access.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If manual data classification is used, then flexibility in classification decisions is maintained, but classification consistency deteriorates due to employee turnover and subjective judgment
Solution Approach 1:
The system enables self-service classification by automatically detecting data changes and applying pre-trained classification models without requiring manual intervention. The automated system monitors database changes, extracts affected columns, and applies appropriate classifications based on pre-established models, eliminating the need for continuous human judgment while maintaining consistency.
Solution Approach 2:
The system performs preliminary action by pre-training classification models before actual data classification is needed. These pre-trained models are ready to immediately apply consistent classification rules when data changes are detected, avoiding the variability introduced by manual classification decisions made at the time of data creation or modification.
2Ease of operation
If manual data classification is performed, then human judgment can be applied, but labor intensity increases and productivity decreases
Solution Approach 1:
The system replaces the mechanical manual classification process with an automated computational system. Instead of requiring human analysts to manually review and classify data, the system uses automated detection mechanisms to identify data changes and applies pre-trained machine learning models to perform classification, dramatically increasing productivity while reducing labor intensity.
Solution Approach 2:
The automated system performs self-service by independently detecting data changes, extracting affected columns, selecting appropriate classification models, and applying classifications without requiring human intervention. This self-service capability eliminates the bottleneck of manual processing while maintaining the quality of classification through pre-trained models.
3Loss of time
If reactive data classification is used, then existing data can be classified, but new or modified data remains unclassified until detected
Solution Approach 1:
The system performs preliminary action by continuously monitoring database changes before classification is needed. When data changes are detected, the system immediately extracts affected columns and applies pre-trained classification models, ensuring that new or modified data receives timely classification and protection rather than waiting for periodic reactive review.
Solution Approach 2:
The system maintains continuity of useful action through continuous monitoring of database changes. Rather than performing classification at discrete periodic intervals, the system operates continuously to detect and respond to data changes in real-time, ensuring that classification is always ready to be applied immediately when needed.
4Reliability
If automated classification models are applied, then classification consistency improves, but system complexity increases
Solution Approach 1:
The system applies segmentation by dividing the classification task into distinct modular components: change detection module, data extraction module, model selection module, and classification application module. Each component handles a specific aspect of the classification process independently, making the overall complex system more manageable and easier to implement while maintaining consistent classification results.
Data Source
AI summary
A system includes a memory and a processor. The memory stores classification models that indicate a data classification to apply to data columns of one or more databases. The processor is configured to periodically access one or more databases and determine that a data table has been newly created or modified. The processor is further configured to determine whether a data column of the data table has an existing data classification, and in response to determining that the data column does not have an existing data classification, determine from the plurality of classification models a particular classification model to use for the data column. The processor is further configured to send instructions to a classification server to apply a particular data classification of the particular classification model to the data column in the classification database.