Balanced Die Set Execution in Data Storage Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In solid state drives (SSDs), different owners or hosts experience varying levels of service due to non-uniform data access command performance across die sets, leading to issues like data access latency, error rates, and command execution time inconsistencies, especially as the SSD approaches full capacity, where write amplification increases and overprovisioning decreases.
Innovation Solution
A quality of service module is employed to measure performance metrics across die sets, generating proactive and reactive strategies to alter data access commands and allocate resources, ensuring consistent performance by prioritizing tasks and adjusting power and buffer resources to maintain balanced data access execution across all die sets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If data access commands are executed across multiple die sets in an SSD, then storage capacity and parallelism are improved, but performance non-uniformity and access time inconsistency worsen
Solution Approach 1:
The patent applies local quality by dynamically adjusting data access commands based on the specific performance characteristics of each die set. The quality of service module monitors performance metrics for individual die sets and modifies access patterns to compensate for local variations in speed, latency, and reliability, ensuring each die set operates at its optimal performance level rather than treating all die sets uniformly
Solution Approach 2:
The system implements dynamics by continuously monitoring performance metrics and adapting data access commands in real-time. The quality of service module dynamically adjusts command allocation, timing, and routing based on current die set performance states, allowing the system to respond to changing conditions and maintain performance uniformity across all die sets as they are filled to different levels
2Productivity
If die sets are filled to different levels to optimize performance, then write amplification is reduced, but access time non-uniformity and service quality inconsistency worsen
Solution Approach 1:
The patent implements feedback by continuously monitoring performance metrics from each die set and using this information to adjust data access commands. The quality of service module receives feedback on access times, latency, and completion status, then modifies subsequent command allocation to compensate for performance variations, creating a closed-loop control system that maintains consistent service quality across die sets filled to different levels
Solution Approach 2:
The system applies parameter changes by modifying data access command parameters such as timing, allocation priority, and routing based on the fill level and performance characteristics of each die set. This allows the system to optimize write efficiency for partially filled die sets while compensating for access time variations through dynamic parameter adjustment
3Quantity of substance
If data is distributed across multiple die sets, then storage overprovisioning is reduced, but command execution performance non-uniformity worsens
Solution Approach 1:
The patent applies preliminary action by proactively monitoring die set performance metrics and predicting potential performance issues before they affect data access. The quality of service module takes preliminary actions to balance data distribution and adjust access patterns across die sets, preventing performance non-uniformity from developing rather than reacting to it after it occurs
Data Source
AI summary
A data storage system can arrange semiconductor memory into a plurality of die sets where performance metrics of execution of a first data access command to a first die set and of a second data access command to a second die set are measured. A proactive strategy is generated to maintain consistent data access command execution performance with a quality of service module based on the measured performance metrics and a third data access command is altered, as directed by the proactive strategy, to prevent a predicted non-uniformity of data access command performance between the first die set and the second die set.