Cloud Remediation System Using Masked Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cloud platform errors can lead to downtime and security vulnerabilities, affecting multiple cloud applications and clients, with existing remediation methods being inefficient and potentially exposing sensitive information.
Innovation Solution
An automated remediation system using a machine learning model trained with masked log files, incident notes, and voice recordings to identify errors and suggest remediation actions, while ensuring data security by masking personally identifiable information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If automated remediation systems process raw log files and incident data, then remediation accuracy improves, but data security risks increase due to exposure of personally identifiable information
Solution Approach 1:
The patent introduces an intermediary masking layer between the raw data and the machine learning model. A masking module processes log files, incident notes, and voice recordings to replace personally identifiable information with anonymized tokens before feeding data to the remediation system. This intermediary structure allows the system to maintain high remediation accuracy while eliminating data security risks associated with PII exposure.
2Reliability
If multiple data sources including voice recordings are integrated for training, then remediation reliability improves, but system complexity increases
Solution Approach 1:
The patent segments the complex multi-source data processing system into distinct functional modules: a voice recording module that processes audio inputs, a log file processing module, an incident note processing module, and a masking module. Each module handles specific data types independently before integration. This segmentation reduces system complexity by creating manageable, specialized components while maintaining the reliability benefits of multi-source data integration.
3Object-affected harmful factors
If manual review of remediation actions is required, then data security is improved, but response time increases
Solution Approach 1:
The patent implements preliminary masking of personally identifiable information in all input data before processing. By pre-anonymizing log files, incident notes, and voice recordings through automated masking operations, the system eliminates security risks at the data intake stage rather than requiring time-consuming manual reviews later. This preliminary action maintains data security while enabling rapid automated remediation.
Data Source
AI summary
In some implementations, a remediation system may receive, from a set of monitoring software, information associated with the cloud platform. The remediation system may apply a set of data security rules to the information to generate masked information. The remediation system may provide the masked information to a machine learning model to receive a classification associated with the masked information and an indication of a suggested remediation action. The remediation system may determine that the classification is associated with an error. The remediation system may transmit, to the cloud platform, a command to perform the suggested remediation action based on determining that the classification is associated with the error.