RAID Cache Backup via Nonvolatile Memory Parity

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

RAID systems face challenges in ensuring data integrity and reliability during power outages due to limited backup unit capacity, leading to increased costs and reduced processing power after power restoration.

Innovation Solution

A backup method that writes cache data from volatile memory to nonvolatile memory, generates parity data, verifies errors, and rewrites data as needed, using a one-to-one connection between backup units and controller modules to ensure power supply and reduce the write-through state duration.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If the capacity of the backup unit is increased, then the reliability of data guarantee is improved, but the cost and size of the RAID system increase

Engineering Contradiction:
Improvedata guarantee reliabilityVSAvoidbackup unit capacity
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

The patent writes cache data to nonvolatile memory in advance before power outage occurs, so that when power is lost, the data is already preserved and does not require large backup units to maintain reliability during extended outages

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent creates a copy of cache data in nonvolatile memory as a backup, allowing the system to recover without needing large capacitive backup units. The nonvolatile memory serves as a duplicate storage location that preserves data integrity

Inventive Principle:
Principle #26Copying

2Reliability

If the capacity of the backup unit is increased, then the reliability of data guarantee is improved, but the size of the RAID system increases

Engineering Contradiction:
Improvedata guarantee reliabilityVSAvoidRAID system size
Core Design Contradiction:
ReliabilityVSVolume of stationary object

Solution Approach 1:

Data is written to nonvolatile memory before power failure, eliminating the need for large backup units and reducing overall system size while maintaining reliability

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

By copying data to nonvolatile memory, the system achieves reliable data protection without requiring additional large-capacity backup components, thus reducing system size

Inventive Principle:
Principle #26Copying

3Productivity

If power is restored after outage, then the RAID system can resume operation, but the write-through state reduces processing power significantly

Engineering Contradiction:
Improveprocessing powerVSAvoidwrite-through state duration
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

Cache data is written to nonvolatile memory before power outage, so when power is restored, the system does not need to maintain a prolonged write-through state, reducing the time productivity is reduced

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system uses parity data to verify data integrity in nonvolatile memory and automatically rewrites corrupted data, allowing the system to exit write-through state faster and resume normal processing power

Inventive Principle:
Principle #23Feedback

4Reliability

If error in cache data cannot be recovered by parity data, then data integrity is compromised, but rewriting in different area increases nonvolatile memory usage

Engineering Contradiction:
Improvedata integrityVSAvoidnonvolatile memory usage
Core Design Contradiction:
ReliabilityVSVolume of stationary object

Solution Approach 1:

The nonvolatile memory is divided into multiple areas, and when error recovery fails, data is rewritten to a different area segment, preserving data integrity while managing memory resources efficiently

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Error correction and rewriting operations are applied locally to specific data areas rather than the entire memory, maintaining data integrity in affected regions while minimizing overall memory usage impact

Inventive Principle:
Principle #3Local quality

Data Source

PatentUS8286028B2Backup method and disk array apparatus
Publication Date: 2012.10.09 FUJITSU LTD
  • US8286028B2 patent drawing
  • US8286028B2 patent drawing
  • US8286028B2 patent drawing

AI summary

A backup method makes a backup of cache data to a nonvolatile memory by using a controller, the cache data being stored in the volatile memory. The backup method includes writing the cache data stored in the volatility memory in a selected area of the nonvolatile memory, generating party data by operating the parity operations between each of the predetermined parts of the cache data in the volatile memory, verifying whether an error found in the part of the cache data in the nonvolatile memory can be recovered by using the parity data, and rewriting the part of the cache data when the error found in the part of the cache data in the nonvolatile memory cannot be recovered by using the parity data in an area of the nonvolatile memory different from the selected area.