Virtual Machine Image Sanitization via Sensitivity Labeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for sanitizing virtual machine images in cloud computing environments are incomplete and ad-hoc, leading to data remanence and potential security breaches, as they fail to thoroughly remove sensitive data across all layers of the software stack, resulting in data leakage and privacy issues.
Innovation Solution
A computer-implemented method that uses a labeler module to attach sensitivity level labels to sensitive data based on labeling dependencies and a sanitizer module to sanitize the data according to sanitization policies, ensuring comprehensive removal of sensitive information from virtual machine images, thereby preventing data leakage and security breaches.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If ad-hoc sanitization methods are used, then the process is simple and quick, but sensitive data remains in the virtual machine image causing security breaches
Solution Approach 1:
The sanitization process is segmented into distinct phases: identification of sensitive data locations, labeling of data with sensitivity levels, and systematic removal or masking. This segmentation allows the complex task to be managed through structured, automated steps while ensuring comprehensive security coverage.
Solution Approach 2:
The system performs preliminary identification and labeling of sensitive data before the actual sanitization operation. By pre-mapping data locations and sensitivity levels, the system ensures that no sensitive information is missed during removal, thereby improving reliability without requiring complex real-time decision-making during execution.
2Reliability
If comprehensive sanitization across all software stack layers is performed, then data security is improved, but the time and resources required increase
Solution Approach 1:
The system performs preliminary identification and mapping of sensitive data locations across all software stack layers before actual sanitization. This pre-processing step creates a roadmap that guides the subsequent removal process, ensuring comprehensive coverage while minimizing the time required for the actual sanitization operation.
Solution Approach 2:
The system creates representations or maps of sensitive data locations and sensitivity levels without actually processing the sensitive data itself. This copying approach allows the system to plan and execute sanitization efficiently by working with metadata rather than the actual sensitive content, reducing processing time while maintaining security.
3Productivity
If sensitive data is not thoroughly removed, then the sanitization process is faster and simpler, but data remanence occurs leading to potential security breaches
Solution Approach 1:
The sanitization process is divided into systematic segments that cover different software stack layers and data types. Each segment is processed with appropriate removal or masking techniques, ensuring that no sensitive data is left behind while maintaining an efficient, automated workflow that preserves productivity.
Solution Approach 2:
The system incorporates verification mechanisms that check whether sensitive data has been successfully removed or masked. This feedback loop ensures that data remanence does not occur while maintaining sanitization speed through automated validation rather than manual verification.
Data Source
AI summary
Sanitizing a virtual machine image of sensitive data is provided. Labeling dependencies and sanitization dependencies between a plurality of software components in the virtual machine image are identified based on labeling execution policies located in a labeler module and sanitization execution policies located in a sanitizer module, respectively. The labeler module and the sanitizer module are inserted in the virtual machine image. A sensitivity level label of a plurality of sensitivity labels is attached to identified sensitive data from the sensitive data contained in the virtual machine image based on the identified labeling dependencies. In response to receiving an input to perform a sanitization of the identified sensitive data having attached sensitivity level labels contained in the virtual machine image, the sanitization of the identified sensitive data having the attached sensitivity level labels contained in the virtual machine image is performed based on the identified sanitization dependencies.