Predictive Memory Maintenance for Spare Die Pre-loading

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing memory systems face disruptions and delays when a primary memory die fails, as data recovery and replacement processes can be time-consuming and error-prone, especially in systems that rely on error correction codes for data integrity.

Innovation Solution

Implementing predictive memory maintenance, where a spare memory die is pre-loaded with data from a primary die deemed likely to fail, allowing for seamless replacement with minimal disruption and improved data accuracy by mirroring data before actual failure, using error correction codes to enhance data integrity.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If data recovery techniques are employed after memory die failure, then data integrity can be maintained, but time loss increases due to the reconstruction process

Engineering Contradiction:
Improvedata integrityVSAvoiddata recovery time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent applies preliminary action by continuously mirroring data from operational primary memory dies to the spare memory die before actual failure occurs. This pre-positioning of data eliminates the need for time-consuming post-failure data reconstruction, as the spare die already contains the necessary data to replace the failed die immediately.

Inventive Principle:
Principle #10Preliminary action

2Device complexity

If traditional die sparing is used without predictive maintenance, then system complexity is reduced, but reliability decreases due to longer downtime during failure replacement

Engineering Contradiction:
Improvememory system complexityVSAvoidsystem availability
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The system performs preliminary data mirroring operations during normal operation to prepare the spare memory die in advance. This preliminary action enables immediate failover without requiring complex real-time data reconstruction operations, thus maintaining simplicity while improving reliability.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The memory controller continuously monitors the operational status and performance metrics of primary memory dies to identify which die is most likely to fail next. This feedback mechanism enables intelligent selection of the target die for data mirroring, optimizing the reliability improvement without adding excessive complexity to the system.

Inventive Principle:
Principle #23Feedback

3Reliability

If data mirroring is performed continuously to all primary memory dies, then data accuracy improves, but use of energy increases due to repeated data transfer operations

Engineering Contradiction:
Improvedata accuracyVSAvoidenergy consumption
Core Design Contradiction:
ReliabilityVSUse of energy by moving object

Solution Approach 1:

Instead of uniformly mirroring data to all primary memory dies, the system applies local quality by selectively identifying and mirroring only the specific primary die that exhibits the highest failure probability based on monitored performance metrics. This targeted approach maintains data accuracy for the critical die while minimizing unnecessary data transfer operations and energy consumption across other healthy dies.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The system performs partial action by mirroring data only to the extent necessary - specifically to the one spare memory die designated for potential failure replacement, rather than creating multiple mirrors or continuously replicating to all dies. This partial approach provides sufficient data accuracy improvement while controlling energy usage.

Inventive Principle:
Principle #16Partial or excessive action

4Reliability

If error correction codes are applied to enhance data integrity, then reliability improves, but device complexity increases due to additional encoding and decoding operations

Engineering Contradiction:
Improvedata integrityVSAvoiderror correction complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent combines preliminary action with error correction by pre-mirroring data to the spare die before failure occurs, which eliminates the need for complex real-time error reconstruction operations. The ECC codes are applied to the data during this pre-positioning process, ensuring integrity is maintained without requiring complex post-failure correction operations.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS10048877B2Predictive memory maintenance
Publication Date: 2018.08.14 INTEL CORP
  • US10048877B2 patent drawing
  • US10048877B2 patent drawing
  • US10048877B2 patent drawing

AI summary

Predictive memory maintenance in accordance with one aspect of the present description, can anticipate a failure of a selected primary memory die of an array, and pre-load a spare memory die with the data of the selected primary memory die deemed to have a likelihood of failure, prior to any actual failure of the selected memory die. In the event that the selected primary memory die does subsequently fail, the spare memory die pre-loaded with the data of the selected primary memory die can readily take the place of the failed primary memory die with a pre-existing copy of the data of the failed primary memory die. Other aspects are described herein.