Modular SSD Architecture with Segmented Controllers

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing solid-state drives (SSDs) face challenges with large capacity configurations due to increased signal connection points, complex management of numerous memory devices, and inflexible block-oriented non-volatile memory allocation, leading to inefficiencies and reliability issues.

Innovation Solution

The implementation of a modular SSD architecture with a simplified controller and point-to-point module interfaces, allowing for dedicated resource allocation and management functions to be offloaded to individual memory modules, enabling flexible user data block lengths and improved bandwidth utilization.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If the controller manages a large number of individual memory devices to increase storage capacity, then the storage capacity is improved, but the device complexity and number of signal connection points increase proportionally

Engineering Contradiction:
Improvestorage capacityVSAvoidcontroller complexity
Core Design Contradiction:
Quantity of substanceVSDevice complexity

Solution Approach 1:

The patent divides the memory system into multiple independent memory devices, each with its own controller. This segmentation allows the main controller to manage fewer devices while achieving high storage capacity through parallel configuration. Each memory device operates semi-autonomously, reducing the burden on the main controller and limiting the propagation of complexity.

Inventive Principle:
Principle #1Segmentation

2Quantity of substance

If the controller manages a large number of individual memory devices to increase storage capacity, then the storage capacity is improved, but the number of signal connection points increases proportionally

Engineering Contradiction:
Improvestorage capacityVSAvoidsignal connection points
Core Design Contradiction:
Quantity of substanceVSLength of stationary object

Solution Approach 1:

By segmenting the system into multiple memory devices with individual controllers, the patent reduces the number of signal connection points required at the main controller. Each memory device handles its own control signals locally, eliminating the need for extensive point-to-point connections from the main controller to each individual memory cell.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces intermediate controllers at the memory device level that act as mediators between the main controller and the memory cells. These intermediate controllers aggregate control functions, reducing the signal connection burden on the main controller while maintaining efficient control over large numbers of memory devices.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Extent of automation

If the interface carries traffic for management functions such as wear leveling, then the management functions are performed, but the interface between controller and memory devices is burdened

Engineering Contradiction:
Improvemanagement functionsVSAvoidinterface burden
Core Design Contradiction:
Extent of automationVSDevice complexity

Solution Approach 1:

The patent segments management functions by implementing wear leveling and other control functions at the individual memory device level rather than centralized in the main controller. This distribution allows management functions to operate autonomously at each device, eliminating the need for constant controller intervention and reducing interface traffic burden.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Each memory device in the patent is equipped with autonomous capabilities to perform management functions such as wear leveling independently. This self-service approach allows memory devices to maintain and optimize themselves without requiring continuous controller involvement, significantly reducing the burden on the controller-memory interface.

Inventive Principle:
Principle #25Self-service

4Device complexity

If block-oriented memory uses fixed page size allocation, then the memory structure is simplified, but the adaptability to different application requirements is reduced

Engineering Contradiction:
Improvememory structureVSAvoidallocation flexibility
Core Design Contradiction:
Device complexityVSAdaptability or versatility

Solution Approach 1:

The patent implements dynamic block allocation within the fixed page structure, allowing the logical organization of data to adapt to different application requirements while maintaining the physical page-based memory structure. This dynamic allocation enables variable block sizes and flexible data organization without changing the underlying memory architecture.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent changes the parameter of block allocation flexibility by allowing variable block sizes and different allocation strategies within the fixed page framework. This enables the memory system to adapt to different application requirements (such as different block sizes for file systems or databases) while maintaining the simplicity of the physical page-based structure.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP2671162B1Control arrangements and methods for accessing block oriented nonvolatile memory
Publication Date: 2019.02.27 MICRON TECHNOLOGY INC
  • EP2671162B1 patent drawingFigure 1
  • EP2671162B1 patent drawingFigure 2
  • EP2671162B1 patent drawingFigure 3

AI summary

A read/write arrangement is described for use in accessing at least one nonvolatile memory device in read/write operations with the memory device being made up of a plurality of memory cells which memory cells are organized as a set of pages that are physically and sequentially addressable with each page having a page length such that a page boundary is defined between successive ones of the pages in the set. The read/write arrangement includes a control arrangement that is configured to store and access a group of data blocks that is associated with a given write operation in a successive series of pages of the memory such that at least an initial page in the series is filled and each block includes a block length that is different than the page length.