RAIM Memory ECC Distribution for Lower Overhead Fault Tolerance

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional Redundant Array of Independent/Inexpensive Memory (RAIM) designs require excessive resources and overhead for fault-tolerant and error correction functionalities, leading to high power consumption and cost, especially in environments with high integration density or harsh conditions like high altitudes and astronomical radiation.

Innovation Solution

The proposed memory system reduces the number of ECC memory devices by distributing ECC symbols across multiple channels and units of data and ECC DRAM chips, eliminating the need for dedicated ECC DIMMs and optimizing the ratio of data to ECC DRAMs, thereby reducing memory overhead and improving resource efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If conventional RAIM designs use dedicated ECC DIMMs and additional ECC memory devices to provide fault-tolerant and error correction functionalities, then reliability and fault-tolerance are improved, but device complexity and resource overhead increase

Engineering Contradiction:
Improvefault-toleranceVSAvoidmemory overhead
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent combines data and ECC memory devices into integrated memory modules, eliminating dedicated ECC DIMMs. Each memory module contains both data memory devices and ECC memory devices that work together as a unified system, reducing overall device complexity while maintaining fault-tolerance capabilities

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The ECC memory devices serve multiple functions: they provide error correction for data in the same memory module, enable RAIM functionalities across multiple channels, and support various RAS features. This multi-functionality reduces the need for separate dedicated ECC components

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

2Reliability

If conventional RAIM designs allocate additional ECC memory devices per channel to support RAIM functionalities, then error correction capability is improved, but power consumption increases

Engineering Contradiction:
Improveerror correction capabilityVSAvoidpower consumption
Core Design Contradiction:
ReliabilityVSUse of energy by moving object

Solution Approach 1:

The patent segments ECC functionality into distributed ECC memory devices across multiple memory modules and channels. Instead of concentrating ECC resources in dedicated DIMMs, the ECC capability is segmented and distributed throughout the memory system, allowing for more efficient power utilization

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system uses a predetermined number of ECC symbols that may be fewer than traditional designs would allocate per channel, yet still provides sufficient error correction capability. This partial action approach reduces power consumption while maintaining adequate error correction for the application requirements

Inventive Principle:
Principle #16Partial or excessive action

3Reliability

If conventional RAIM designs use more ECC memory devices to ensure robustness in harsh environments, then reliability under harsh conditions is improved, but cost increases

Engineering Contradiction:
Improverobustness in harsh environmentsVSAvoidnumber of DRAM chips
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

The patent changes the organizational parameters of memory devices, transitioning from dedicated ECC DIMMs to integrated memory modules with specific ratios of data to ECC devices. This parameter change optimizes the balance between reliability and quantity of components

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The system uses an asymmetric distribution of ECC symbols across channels, with a predetermined number of ECC symbols allocated to support RAIM functionalities. This asymmetric allocation optimizes component usage while maintaining robustness against errors in harsh environments

Inventive Principle:
Principle #4Asymmetry

4Reliability

If conventional RAIM designs allocate dedicated ECC DIMMs for each channel to support RAIM functionalities, then fault-tolerance is improved, but memory bandwidth efficiency decreases

Engineering Contradiction:
Improvefault-toleranceVSAvoidmemory bandwidth efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

By merging data and ECC memory devices into unified memory modules with integrated interfaces, the system eliminates the overhead of separate dedicated ECC DIMM communications. This consolidation improves memory bandwidth efficiency while maintaining fault-tolerance through the combined operation of data and ECC devices within each module

Inventive Principle:
Principle #5Merging (Combining)

Data Source

PatentUS10824508B2High efficiency redundant array of independent memory
Publication Date: 2020.11.03 INTERNATIONAL BUSINESS MACHINE CORPORATION
  • US10824508B2 patent drawing
  • US10824508B2 patent drawing
  • US10824508B2 patent drawing

AI summary

A memory system includes memory modules having a number of sets of memory devices including data memory devices for data and error correction code (ECC). The ECC memory devices carry ECC symbols for the memory modules. A host receives and decodes the ECC symbols and executes error correction operations. The host and the memory modules are coupled by a number of channels.