Dynamic Host Read to Refresh Ratio Control in SSDs

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Non-volatile storage devices like SSDs face performance degradation due to read disturbance issues, where frequent read operations lead to data errors and require time-consuming read refresh operations, affecting overall performance, especially in applications with high read demands like online streaming services.

Innovation Solution

A method to dynamically manage the ratio of host read operations to read refresh operations in storage devices by controlling the number of read requests and updating this ratio based on accumulated read requests, read disturb thresholds, and block information, ensuring timely read refresh operations before data errors occur.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If read refresh operations are performed frequently to prevent read disturbance errors, then data reliability is improved, but read performance deteriorates due to time spent on refresh operations

Engineering Contradiction:
Improvedata reliabilityVSAvoidread performance
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent implements dynamic adjustment of the read refresh threshold based on the current read count of blocks. The threshold is not fixed but changes adaptively according to how many times blocks have been read, allowing the system to optimize between preventing read disturbance and maintaining read performance under varying workload conditions

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent changes the parameter of read refresh threshold dynamically. When the read count of a block exceeds a first threshold, the threshold is adjusted to a second threshold value, which is lower than the first. This parameter change allows the system to perform read refresh operations at appropriate moments without unnecessarily degrading read performance

Inventive Principle:
Principle #35Parameter changes

2Quantity of substance

If the number of memory chips and channels is increased to improve storage capacity and performance, then storage capacity and throughput are improved, but device complexity increases

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

Solution Approach 1:

The patent divides the storage device into multiple independent channels, each with its own read refresh queue and control logic. This segmentation allows parallel processing of read operations and read refresh operations across different channels, managing complexity through modular design while maintaining high throughput

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a controller as an intermediary that manages the complex coordination between multiple memory chips and channels. The controller implements the read refresh mechanism and threshold adjustment logic, abstracting the complexity from the overall system and enabling efficient management of multiple components

Inventive Principle:
Principle #24Intermediary (Mediator)

3Productivity

If read refresh operations are postponed to improve read performance, then read performance is improved, but data reliability deteriorates due to increased read disturbance

Engineering Contradiction:
Improveread performanceVSAvoiddata reliability
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent implements a feedback mechanism where the controller continuously monitors the read count of blocks and compares it against the dynamic threshold. When the read count exceeds the threshold, the controller triggers read refresh operations. This feedback loop ensures that read refresh is performed at the optimal moment to prevent read disturbance while minimizing performance impact

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent performs read refresh operations preliminarily, before read disturbance errors actually occur. By monitoring read counts and triggering refresh when thresholds are exceeded, the system proactively prevents data corruption rather than reacting to errors after they occur, maintaining both reliability and performance

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS11789659B2Method for dynamically managing host read operation and read refresh operation in a storage device, storage device, and storage medium
Publication Date: 2023.10.17 SK HYNIX INC
  • US11789659B2 patent drawing
  • US11789659B2 patent drawing
  • US11789659B2 patent drawing

AI summary

A method for dynamically managing host read operation and read refresh operation in a storage device, a storage device and a storage medium thereof are provided. The method includes: controlling, by a controller of the storage device, a ratio of the number of host read operation to the number of read refresh operation in the storage device to be in line with a first value and obtaining a total read request count which accumulates in the storage device; when a criterion for updating the ratio is satisfied, determining, by the controller, a second value for the ratio of the number of host read operation to the number of read refresh operation according to the total read request count and information of blocks to be refreshed in the storage device; and controlling, by the controller, the number of host read operation and the number of read refresh operation so that a ratio of the number of host read operation to the number of read refresh operation is in line with the second value.