Hierarchical Block Storage Deduplication via Object Store Hashing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In a scenario where multiple block storage devices share an object storage device without deduplication functionality, the same data from these block storage devices are duplicated and stored in the object storage, leading to inefficiencies and increased storage costs.

Innovation Solution

A block storage device configuration that includes a hierarchical logical disk structure with physical devices of varying access performance, a deduplication unit, and a hierarchy changing unit. The deduplication unit searches the object storage for data with reduced access frequency and, if not found, transmits it to the object storage, while omitting transmission if the data is already present, thus preventing duplication.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If multiple block storage devices share an object storage device without deduplication functionality, then storage capacity is increased, but data duplication occurs leading to increased storage costs and reduced efficiency

Engineering Contradiction:
Improvestorage capacityVSAvoidstorage efficiency
Core Design Contradiction:
Adaptability or versatilityVSLoss of substance

Solution Approach 1:

The block storage device performs preliminary deduplication checks before transmitting data to the object storage device. The deduplication unit calculates hash values of data blocks and checks against previously stored hash values in the deduplication table, preventing duplicate data from being transmitted to object storage in the first place.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system uses hash value copies instead of actual data copies for deduplication comparison. By storing only hash values in the deduplication table and comparing these lightweight copies, the system efficiently identifies duplicate data blocks without requiring actual data comparison, reducing overhead while maintaining storage efficiency.

Inventive Principle:
Principle #26Copying

2Reliability

If data is frequently transferred between block storage and object storage, then data availability is maintained, but network bandwidth is consumed and transfer time increases

Engineering Contradiction:
Improvedata availabilityVSAvoiddata transfer time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system performs preliminary hash value comparison before actual data transfer. By checking the deduplication table for existing hash values upfront, the system determines whether data transfer is necessary, avoiding unnecessary network transmissions and reducing overall data transfer time.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The deduplication table acts as an intermediary between block storage and object storage operations. It stores hash value mappings that mediate data transfer decisions, allowing the system to quickly determine data uniqueness without direct data comparison, thus reducing transfer overhead and time.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Quantity of substance

If all data blocks are stored in object storage, then storage capacity is maximized, but access performance for frequently accessed data deteriorates

Engineering Contradiction:
Improvestorage capacityVSAvoidaccess performance
Core Design Contradiction:
Quantity of substanceVSSpeed

Solution Approach 1:

The storage system is segmented into two distinct layers: block storage for frequently accessed data requiring high performance, and object storage for archival data requiring high capacity. The hierarchical logical disk structure divides data blocks between these layers based on access patterns, with the active layer providing fast access and the archive layer providing extensive capacity.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Different storage locations are assigned different quality characteristics: block storage provides high-speed access quality for active data, while object storage provides high-capacity quality for archival data. The system dynamically assigns data blocks to appropriate storage locations based on their access requirements, ensuring each data block resides in the most suitable storage environment.

Inventive Principle:
Principle #3Local quality

Data Source

PatentUS10185495B2Block storage device having hierarchical disks with different access frequencies
Publication Date: 2019.01.22 NEC CORP
  • US10185495B2 patent drawing
  • US10185495B2 patent drawing
  • US10185495B2 patent drawing

AI summary

A first block storage device sharing an object storage device with a second block storage device has: plural kinds of physical devices having different access response performances; and a hierarchical logical disk created by using the storage areas of the physical devices and the storage area of the object storage device. When the access frequency of a logical block stored in a hierarchy one level above a lowest hierarchy falls below a threshold, the first block storage device searches the object storage device for a logical block having the hash of the logical block as a content address. When the search fails, the first block storage device transmits the logical block to the object storage device to change a hierarchy to store the logical block to the lowest hierarchy, whereas when the search succeeds, omits the transmission and changes a hierarchy to store the logical block to the lowest hierarchy.