Layered Parity Protection for Storage Data Reconstruction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Storage systems face challenges in tolerating multiple concurrent disk failures without data loss, leading to increased overhead in parity information and prolonged reconstruction times, which complicates data recovery and increases the risk of secondary failures.

Innovation Solution

A method is introduced that calculates multiple set and group parity units, which are cached in a storage system's memory, allowing for efficient reconstruction of data units upon disk failures by utilizing non-lost set and group parity units, and strategically destaging these units across multiple disks to minimize data loss and reconstruction time.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If the number of parity units is increased to tolerate more concurrent disk failures, then the reliability of the storage system is improved, but the overhead of parity information increases and reduces the storage capacity for user data

Engineering Contradiction:
Improvetolerance to concurrent disk failuresVSAvoidstorage capacity for user data
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

The patent segments the parity protection into multiple hierarchical levels: set-level parity units (protecting against 1 failure) and group-level parity units (protecting against additional failures). This segmentation allows the system to provide multi-level protection while optimizing the distribution of parity overhead, enabling tolerance to multiple concurrent failures without requiring all parity units to be present simultaneously for all operations.

Inventive Principle:
Principle #1Segmentation

2Quantity of substance

If the stripe size is increased to reduce parity overhead percentage, then the storage efficiency is improved, but the reconstruction time increases and the complexity of data recovery increases

Engineering Contradiction:
Improvestorage efficiencyVSAvoidreconstruction time
Core Design Contradiction:
Quantity of substanceVSLoss of time

Solution Approach 1:

The patent divides the large stripe into smaller sets, each with its own parity unit for single-failure recovery. This segmentation enables faster reconstruction for common single-failure scenarios by limiting the reconstruction scope to only the affected set rather than the entire large stripe, while still achieving high storage efficiency through the overall large stripe structure.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent pre-calculates and stores multiple levels of parity units (set-level and group-level) in advance. When failures occur, the system can immediately use the pre-computed parity units for reconstruction without performing complex real-time calculations, significantly reducing reconstruction time while maintaining high storage efficiency.

Inventive Principle:
Principle #10Preliminary action

3Quantity of substance

If the stripe size is increased to reduce parity overhead, then the storage efficiency is improved, but the device complexity increases due to excessive data processing requirements

Engineering Contradiction:
Improvestorage efficiencyVSAvoiddata processing complexity
Core Design Contradiction:
Quantity of substanceVSDevice complexity

Solution Approach 1:

The patent segments the data and parity units into hierarchical sets and groups, simplifying the processing logic. Each set has its own parity unit, and groups of sets share group-level parity units. This segmentation reduces the complexity of data processing by breaking down large-scale operations into smaller, more manageable units that can be processed independently and in parallel.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system performs preliminary parity calculations and stores multiple levels of parity units in advance. This pre-computation eliminates the need for complex real-time calculations during normal operations and failure recovery, significantly reducing device complexity while maintaining high storage efficiency through the optimized stripe structure.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS9229812B2Layered failure protection scheme for protecting destaged data units
Publication Date: 2016.01.05 INFINIDAT LTD
  • US9229812B2 patent drawing
  • US9229812B2 patent drawing
  • US9229812B2 patent drawing

AI summary

A method that includes calculating multiple set parity units for multiple sets of data units that are cached in a storage system thereby providing a group of units that includes multiple sets of units. Each set of units includes a set of data units and at least one associated set parity unit that facilitate a reconstruction of the set of units when a first number of units of the set of units is lost. Calculating at least one group parity unit for the group of units. A combination of the at least one group parity unit and at least one associated set parity unit facilitate a reconstruction of the set of units upon a failure of a second number of units of the set of units. The second number of units of the set of units exceeds the first number of units of the set of units. Destaging the group of units and the at least one group parity unit to multiple disks of the storage system.