RAID Redundant Code Generation Without Bit Shift Overhead

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing RAID systems face processing burdens due to the high computational demands of Galois field operations required for generating and restoring redundant codes, particularly when using general-purpose CPUs, which necessitate either inefficient bit shift and data mask operations or costly dedicated hardware.

Innovation Solution

A method and device that perform Galois field operations without special hardware by executing exclusive OR operations in data units, eliminating the need for bit shift and data mask operations, using a CPU with an exclusive OR arithmetic circuit and memory to generate and restore redundant codes efficiently.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If Galois field operations are performed using general-purpose CPUs, then device complexity is reduced, but processing speed deteriorates due to the need for bit shift and data mask operations

Engineering Contradiction:
Improvehardware complexityVSAvoidprocessing speed
Core Design Contradiction:
Device complexityVSSpeed

Solution Approach 1:

The patent changes the operational parameters of Galois field operations by performing XOR operations on data units (e.g., bytes or words) rather than individual bits. This parameter change eliminates the need for bit shift and data mask operations, allowing general-purpose CPUs to execute Galois field operations at speeds comparable to dedicated hardware while maintaining device simplicity

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent substitutes the mechanical bit-level manipulation system (bit shift and mask operations) with a more efficient data-unit level XOR operation system. This substitution leverages the native capabilities of general-purpose CPUs to perform operations on entire data units simultaneously, eliminating the sequential bit-by-bit processing that slowed down previous implementations

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Speed

If dedicated hardware is used for Galois field operations, then processing speed is improved, but device complexity and cost increase

Engineering Contradiction:
Improveprocessing speedVSAvoidhardware complexity
Core Design Contradiction:
SpeedVSDevice complexity

Solution Approach 1:

The patent makes general-purpose CPUs universal by enabling them to perform Galois field operations efficiently through data-unit XOR operations. This eliminates the need for separate dedicated hardware circuits, allowing the same CPU to handle both general computing tasks and RAID redundancy operations without requiring specialized Galois field processing units

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The patent copies the functional capability of dedicated Galois field hardware into software-based XOR operations that run on general-purpose CPUs. By replicating the computational function rather than requiring physical dedicated hardware, the system achieves similar processing speeds without the complexity and cost of specialized circuits

Inventive Principle:
Principle #26Copying

3Adaptability or versatility

If bit shift and data mask operations are used for Galois field operations, then compatibility with general-purpose CPUs is maintained, but processing time increases

Engineering Contradiction:
ImproveCPU compatibilityVSAvoidcomputation time
Core Design Contradiction:
Adaptability or versatilityVSLoss of time

Solution Approach 1:

The patent extracts and eliminates the time-consuming bit shift and data mask operations from the Galois field computation process. By removing these intermediate steps and directly performing XOR operations on data units, the patent maintains CPU compatibility while dramatically reducing computation time

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

Instead of breaking down data into individual bits and performing sequential operations (the traditional approach), the patent inverts the approach by operating directly on complete data units. This inversion eliminates the need for bit-level manipulation and allows the CPU to process entire data units in parallel, significantly reducing computation time

Inventive Principle:
Principle #13The other way round (Inversion)

Data Source

PatentUS8499224B2Redundant code generation method and device, data restoration method and device, and raid storage device
Publication Date: 2013.07.30 NEC CORP
  • US8499224B2 patent drawing
  • US8499224B2 patent drawing
  • US8499224B2 patent drawing

AI summary

A redundant code generation method includes: dividing original data into data strings; dividing each data string into a number of bit strings that accords with an extended Galois field operation; storing each of the bit string in a different memory area of a memory; and executing an exclusive OR operation among vectors, which are extracted from the respective bit strings stored in the memory, according to an operational expression to compute bit strings that make up redundant code data strings without carrying out a bit shift operation within the vectors. A predetermined plural number of bits is taken as a data unit and the number of bits as elements constituting each vector is equal to the data unit. The operational expression includes a companion matrix of a primitive polynomial of the Galois field and defined the generation of the redundant code data strings.