Adaptive Content-Aware Deduplication for Similar Data Zones

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing deduplication systems lack content awareness, leading to poor deduplication ratios and increased storage and bandwidth consumption, especially when dealing with encrypted or re-indexed data, and struggle with managing large numbers of small metadata objects, resulting in performance degradation.

Innovation Solution

A content-aware and adaptive deduplication processor that identifies and manages zones of contiguous bytes, generating zone stamps to detect similarity and perform delta compression, reducing metadata overhead and enabling efficient storage and transmission of similar data zones.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If deduplication operates on small kilobyte-sized objects, then deduplication granularity is improved, but metadata management complexity increases significantly

Engineering Contradiction:
Improvededuplication granularityVSAvoidmetadata management complexity
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent segments data into zones of contiguous bytes rather than fixed-size chunks, allowing variable-length segments that are grouped into application objects. This segmentation approach reduces the total number of metadata entries while maintaining fine-grained deduplication capability within each zone, resolving the contradiction between granularity and metadata complexity.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent merges multiple small zones into larger application objects that can be managed as single metadata entities. By combining related zones that belong to the same application context into unified objects, the system reduces metadata overhead while preserving the ability to perform granular deduplication operations on individual zones within the merged structure.

Inventive Principle:
Principle #5Merging (Combining)

2Reliability

If deduplication is applied to encrypted or re-indexed data, then data security is improved, but deduplication ratio deteriorates

Engineering Contradiction:
Improvedata securityVSAvoiddeduplication ratio
Core Design Contradiction:
ReliabilityVSLoss of substance

Solution Approach 1:

The patent performs deduplication operations before data is encrypted or re-indexed by downstream applications. By conducting deduplication on plaintext data in advance, the system achieves high deduplication ratios while allowing subsequent encryption or re-indexing operations to proceed without interfering with the deduplication benefits, thus resolving the contradiction between security and deduplication effectiveness.

Inventive Principle:
Principle #10Preliminary action

3Speed

If all metadata is retained in memory, then access speed is improved, but memory consumption increases

Engineering Contradiction:
Improvemetadata access speedVSAvoidmemory consumption
Core Design Contradiction:
SpeedVSQuantity of substance

Solution Approach 1:

The patent merges multiple small metadata entries into consolidated application object metadata structures. By combining metadata for related zones into unified objects, the total metadata volume is reduced, allowing all metadata to fit in memory while maintaining fast access speeds. This merging approach resolves the contradiction between memory-resident metadata and memory consumption.

Inventive Principle:
Principle #5Merging (Combining)

Data Source

PatentEP2256934B1Method and apparatus for content-aware and adaptive deduplication
Publication Date: 2018.10.10 EXAGRID SYST
  • EP2256934B1 patent drawingFigure 1
  • EP2256934B1 patent drawingFigure 2a
  • EP2256934B1 patent drawingFigure 2b

AI summary

A method, a system, an apparatus, and a computer readable medium for transmission of data across a network are disclosed. The method includes receiving a data stream, analyzing the received data stream to determine a starting location and an ending location of each zone within the received data stream, based on the starting and ending locations, generating a zone stamp identifying the zone, the zone stamp includes a sequence of contiguous characters representing at least a portion of data in the zone, wherein the order of characters in the zone stamp corresponds to the order of data in the zone, comparing the zone stamp with another zone stamp of another zone in any data stream received, determining whether the zone is substantially similar to another zone by detecting that the zone stamp is substantially similar to another zone stamp, delta-compressing zones within any data stream received that have been determined to have substantially similar zone stamps, thereby deduplicating zones having substantially similar zone stamps within any data stream received, and transmitting the deduplicated zones across the network from one storage location to another storage location.