FHE Database Row Reuse via Deletion Queues
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Fully homomorphically encrypted (FHE) databases face significant memory overhead and performance degradation due to the need to store and manage deleted rows, which increases memory usage and query runtime, especially when deleted rows are not reused.
Innovation Solution
Implementing a system that stores deleted row identifiers in a deletion queue, allowing for deferred actual deletion and reuse of deleted rows, and using insertion and deletion caches to manage rows efficiently, enabling batch operations and reducing noise and chain index consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If deleted rows are stored and managed in an FHE database, then data security is maintained, but memory overhead increases
Solution Approach 1:
The patent extracts the management of deleted rows from the main FHE database by implementing a separate deletion queue. When rows are deleted, their identifiers are moved to this external queue rather than being retained in the encrypted database structure, thereby reducing memory overhead while preserving security through FHE for active data.
Solution Approach 2:
The patent implements a row reuse mechanism where deleted rows are not permanently discarded but recovered and reused. The deletion queue tracks identifiers of deleted rows, and when new rows need to be inserted, the system can reuse the storage space of previously deleted rows, reducing overall memory consumption while maintaining security.
2Reliability
If deleted rows are not reused, then data integrity is maintained, but productivity decreases due to increased memory usage and query runtime
Solution Approach 1:
The patent performs preliminary actions by pre-processing delete operations to extract row identifiers and store them in a deletion queue before actual removal. This preliminary step allows the system to maintain data integrity by tracking what has been deleted while preparing the infrastructure for efficient row reuse, thereby improving future query performance without compromising current integrity.
Solution Approach 2:
The patent recovers deleted rows for reuse by maintaining a mapping between deleted row identifiers and their new positions. This recovery mechanism allows the database to reuse storage space efficiently, reducing memory overhead and improving query runtime performance while maintaining data integrity through the FHE encryption layer.
3Reliability
If FHE encryption is applied to all rows, then security is enhanced, but device complexity increases
Solution Approach 1:
The patent segments the database system into distinct components: an FHE-encrypted main database for active data, a separate deletion queue for tracking deleted rows, and an insertion queue for managing new rows. This segmentation allows FHE encryption to be applied selectively to sensitive data while using simpler data structures for metadata management, thereby enhancing security without proportionally increasing overall system complexity.
Solution Approach 2:
The patent introduces intermediary structures (deletion queue and insertion queue) that mediate between the FHE-encrypted database and external operations. These intermediaries handle the complexity of deletion and insertion operations outside the encrypted space, allowing FHE to maintain security while the intermediary layers manage the operational complexity.
4Productivity
If batch operations are implemented, then productivity improves, but noise and chain index consumption increase
Solution Approach 1:
The patent implements periodic action by processing deletions and insertions in batches rather than continuously. The deletion queue accumulates row identifiers over time, and batch operations are performed at intervals, reducing the frequency of FHE operations and thereby reducing noise accumulation and chain index consumption while maintaining improved productivity through efficient batch processing.
Data Source
AI summary
An example system includes a processor to receive, from a client device, a delete query requesting deletion of a row of in a fully homomorphically encrypted (FHE) database. The processor can store an identifier of the row to be deleted in a deletion queue, where the row is to be replaced with values of a row to be inserted from a received insertion query.