Open-Channel SSD Host Data Relocation for Retention
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional solid state drives (SSDs) lack the ability for hosts to directly manage garbage collection and wear leveling operations, limiting the flexibility and compatibility of firmware algorithms developed by one manufacturer for use on SSDs from other manufacturers, and face data retention issues due to factors like ambient temperature and access frequency.
Innovation Solution
The introduction of an open-channel SSD driver on the host allows direct management of non-volatile memory, including the use of a prediction function to assess data loss risk based on block characteristic parameters and relocate data to prevent retention failures by moving data to blank blocks when the risk exceeds a threshold.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a conventional SSD with flash translation layer is used, then the SSD can autonomously manage garbage collection and wear leveling, but the host cannot directly manage these operations and firmware algorithms are manufacturer-specific
Solution Approach 1:
The patent extracts the flash translation layer functionality from the SSD device and relocates it to the host system. The host now directly manages garbage collection and wear leveling operations by accessing the SSD's non-volatile memory blocks, thereby eliminating the need for manufacturer-specific firmware algorithms in the SSD while giving the host direct control over these critical operations.
Solution Approach 2:
By moving the flash translation layer to the host, the system achieves universality where a single host-based management approach can work with SSDs from different manufacturers. The host's garbage collection and wear leveling algorithms become applicable across multiple SSD devices, eliminating manufacturer-specific firmware constraints and enabling broader compatibility.
2Quantity of substance
If data is stored in non-volatile memory blocks, then storage capacity is utilized, but data retention failures occur due to ambient temperature and access frequency
Solution Approach 1:
The patent implements preliminary action by proactively monitoring block characteristic parameters (such as program/erase cycle counts and temperature exposure) and predicting data retention risks before actual data loss occurs. When prediction indicates high risk, the system pre-migrates data from vulnerable blocks to safer blocks, preventing retention failures before they happen rather than reacting after data loss.
Solution Approach 2:
The system establishes a feedback mechanism where block characteristic parameters are continuously monitored and fed back to the garbage collection and wear leveling algorithms. This feedback loop enables dynamic adjustment of data migration decisions based on real-time block health status, temperature conditions, and access patterns, thereby maintaining data retention reliability while utilizing storage capacity.
Data Source
AI summary
A data processing method for a computer system is provided. The computer system includes a host and an open-channel solid state drive. The open-channel solid state drive is connected with the host. The data processing method includes the following steps. Firstly, plural block characteristic parameters of a specified block in a non-volatile memory of the open-channel solid state drive are collected. Then, the plural block characteristic parameters are inputted into a prediction function, so that a prediction value is acquired. If the prediction value exceeds a threshold value, a data in the specified block of the non-volatile memory is moved to a blank block of the non-volatile memory by the host.