Multi-Level Cloud Deduplication for Storage Security
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In multi-tenanted cloud-based deduplication storage systems, maintaining data security while maximizing storage utilization is challenging due to the need to isolate data from multiple clients, leading to potential duplicate data storage across clients.
Innovation Solution
A multi-level deduplication approach is implemented, where fingerprints are generated from data segments, compared across client-specific and common fingerprint indexes, and stored in backup storage devices, allowing for efficient deduplication at both client-specific and system-wide levels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data from individual clients is isolated from other clients' data for security reasons, then data security is improved, but storage utilization deteriorates due to duplicate data storage across different clients
Solution Approach 1:
The patent segments the deduplication process into multiple levels: client-level segmentation where each client's data is processed separately to maintain security isolation, and system-level segmentation where fingerprints are compared across clients to eliminate duplicates. This multi-level segmentation resolves the contradiction by maintaining security boundaries while enabling cross-client deduplication.
Solution Approach 2:
The patent introduces a new dimension of operation by implementing multi-mode deduplication that can operate in different modes (client-specific mode and system-wide mode). This dimensional change allows the system to switch between prioritizing security isolation and prioritizing storage optimization based on requirements, resolving the inherent contradiction between these two goals.
2Quantity of substance
If a deduplication scheme is employed to reduce storage utilization, then storage efficiency is improved, but data security may be compromised by exposing duplicate data across clients
Solution Approach 1:
The patent introduces fingerprint indexes as intermediaries that enable deduplication without direct data exposure. Instead of comparing actual client data, the system compares cryptographic fingerprints (hashes) of data segments. This intermediary mechanism allows the system to identify duplicates across clients while maintaining security isolation, as the actual data never needs to be exposed or compared directly.
Solution Approach 2:
The patent uses cryptographic fingerprints (copies of data characteristics) instead of the actual data to perform deduplication operations. These fingerprints are mathematical representations that preserve the ability to identify duplicates while containing no actual sensitive information. This copying approach enables storage optimization without compromising data security.
Data Source
AI summary
A system, method, and medium for performing multi-level and multi-mode deduplication in a cloud based backup storage system. Fingerprints are generated for data segments from individual clients at a first level, with a separate fingerprint index for each individual client. The fingerprints are compared to their respective fingerprint indexes, and if the fingerprints are unique, then their associated data segments may be deduplicated across data segments from a plurality of clients at a second level. The fingerprints are compared to a common fingerprint index and deduplicated at a second level. The common fingerprint index contains fingerprints generated from data segments from a plurality of clients connected to the backup storage system.