Dual-State Leaky Bucket Memory Performance Shaping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Memory devices, such as Flash memory devices, face performance issues due to excessive time spent on internal operations in response to host-side requests, leading to reduced overall performance, and existing solutions introduce side effects that complicate management.
Innovation Solution
A method and apparatus utilizing a dual-state leaky bucket model to monitor and control performance metrics of access control groups, dynamically adjusting performance quotas by configuring drain rates based on fill levels, ensuring efficient data access performance shaping without introducing significant side effects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the memory device processes host-side requests with comprehensive internal operations, then data access completeness is improved, but the time spent on internal operations increases excessively
Solution Approach 1:
The patent extracts and separates critical performance metrics (read latency, write latency, read IOPS, write IOPS) from the comprehensive internal operations. By monitoring these specific metrics independently and applying the dual-state leaky bucket model selectively to each access control group, the system ensures data access completeness while preventing excessive time consumption in non-critical operations.
Solution Approach 2:
The patent dynamically changes performance parameters (drain rates) based on monitored metrics and dual-state conditions. By adjusting drain rates according to whether the system is in state 0 or state 1, the memory device optimizes the balance between completing data access operations and minimizing internal operation time, thereby resolving the contradiction between reliability and time loss.
2Productivity
If existing solutions are implemented to correct performance issues, then performance metrics are improved, but side effects are introduced
Solution Approach 1:
The patent introduces a dual-state leaky bucket model as an intermediary mechanism between host commands and memory operations. This model acts as a mediator that monitors performance metrics and dynamically adjusts access control group behaviors through drain rate configuration, improving performance metrics while avoiding the side effects of direct intervention methods used in existing solutions.
Solution Approach 2:
The system enables self-service by allowing access control groups to automatically adjust their own performance parameters based on monitored metrics and dual-state conditions. Each access control group independently responds to performance shaping requirements without requiring external intervention, thereby improving productivity while minimizing harmful side effects from centralized control adjustments.
3Productivity
If performance quotas are dynamically adjusted for access control groups, then data access performance is improved, but system complexity increases
Solution Approach 1:
The patent implements dynamic performance quotas through the dual-state leaky bucket model, where drain rates are adjusted in real-time based on monitored performance metrics and state transitions. This dynamic approach improves data access performance by adapting to changing conditions while managing system complexity through automated state-based control logic that replaces static, pre-configured performance parameters.
Data Source
AI summary
A method for performing data access performance shaping of a memory device and associated apparatus are provided. The method may include: receiving a plurality of host commands from a host device, for performing data access on the NV memory according to the plurality of host commands; and monitoring the plurality of host commands to control respective performance metrics of a plurality of access control groups of the memory device with a dual-state leaky bucket (LB) model, wherein regarding any access control group, for example: determining at least one first performance metric according to at least one first command to be a first LB fill level of a dual-state LB; in response to the first LB fill level being below a state threshold, determining the dual-state LB to be in a first predetermined state, and configuring the dual-state LB to have a first predetermined drain rate, for dynamically adjusting performance quota.