Striped Data Rebalancing via Parallel Chunk Allocation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for rebalancing striped data across multiple storage devices are inefficient, requiring serial data movement and temporary storage, leading to prolonged processing times, especially when adding or removing storage devices.

Innovation Solution

The method involves allocating contiguous storage spaces into chunks, where the number of chunks is the least common multiple of specific values, and then distributing these chunks across logical stripes, allowing for parallel data movement and reduced data transfer, thereby optimizing the rebalancing process.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If conventional relayout techniques are used to add stripes to striped local volume, then data is moved serially through temporary storage devices, but the process takes excessively long time (2.5 hours for 1200GB)

Engineering Contradiction:
Improvedata integrity during relayoutVSAvoidrelayout completion time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent segments the relayout process into multiple parallel operations by dividing data into different stripe sets that can be moved simultaneously across multiple storage devices. Instead of moving all data serially through temporary storage, the system performs segmented parallel transfers, reducing the overall relayout time from hours to minutes while maintaining data integrity through coordinated chunk allocation and verification.

Inventive Principle:
Principle #1Segmentation

2Reliability

If conventional relayout techniques are used, then temporary storage devices are required to hold data during transfer, but this increases device complexity and storage overhead

Engineering Contradiction:
Improvedata safety during rebalancingVSAvoidstorage system architecture
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent uses a metadata structure (chunk allocation table) as an intermediary to coordinate data movement between source and destination storage devices. Instead of requiring physical temporary storage devices, the system uses the metadata layer to track and manage parallel data transfers, eliminating the need for additional hardware while maintaining reliable rebalancing operations.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Reliability

If conventional relayout techniques are used, then all striped data must be moved serially, but this reduces productivity and increases operational time by 94%-95%

Engineering Contradiction:
Improvecomplete data migrationVSAvoidstorage management efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent implements continuous parallel data movement by maintaining multiple active transfer streams simultaneously. While conventional methods complete one serial transfer at a time, this system continuously moves multiple stripe sets in parallel across available storage devices, maximizing throughput and reducing operational time by eliminating idle periods and maintaining constant productive action throughout the rebalancing process.

Inventive Principle:
Principle #20Continuity of useful action

Data Source

PatentEP3785105B1Systems and methods for rebalancing striped information across multiple storage devices
Publication Date: 2025.03.26 VERITAS TECHNOLOGIES LLC
  • EP3785105B1 patent drawingFigure 1
  • EP3785105B1 patent drawingFigure 2
  • EP3785105B1 patent drawingFigure 3

AI summary

The disclosed computer-implemented method for rebalancing striped information across multiple storage devices may include (1) allocating "Y" contiguous storage spaces on "Y" physical devices, (2) dividing, at the computing device, the "Y" continuous storage spaces into "N" subvolumes, (3) allocating the "N" subvolumes to "X" logical stripes, (4) allocating each of the first "Y" logical stripes to a respective physical stripe in each of the "Y" physical stripes, (5) distributing remaining subvolumes to respective logical stripes, and (6) allocating the respective subvolumes across the "Y" physical stripes. Various other methods, systems, and computer-readable media are also disclosed.