Dynamic Read Request Threshold for Storage Controllers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing storage device systems use static limits for outstanding read requests, which are inefficient for varying latency connections, leading to unnecessary memory reservation, complex reordering circuitry, and potential timeouts, especially in low latency scenarios.
Innovation Solution
A storage device controller dynamically adjusts the maximum outstanding read requests threshold based on the host-device connection protocol and measured latency, optimizing read request scheduling by setting a dynamic read request threshold.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a static limit on outstanding read requests is used to account for high latency connections, then the data transfer bus can be fully utilized in high latency scenarios, but memory usage increases unnecessarily and reordering circuitry becomes more complicated in low latency connections
Solution Approach 1:
The patent implements a dynamic read request threshold that automatically adjusts based on measured latency conditions. The controller monitors latency and modifies the threshold accordingly, transitioning from a static high-value setting to a dynamic adaptive setting. This resolves the contradiction by allowing high bus utilization during high latency periods while reducing the threshold during low latency periods, thereby simplifying reordering circuitry requirements when they are not needed.
Solution Approach 2:
The patent changes the parameter of the read request threshold from a fixed static value to a dynamic value that varies based on latency measurements. By adjusting this parameter according to actual connection conditions, the system achieves optimal bus utilization without permanently maintaining the complexity associated with high thresholds, thus resolving the contradiction between productivity and device complexity.
2Reliability
If a static limit on outstanding read requests is used to account for high latency connections, then sufficient memory is reserved for all scenarios, but memory usage increases unnecessarily in low latency connections
Solution Approach 1:
The patent applies dynamics by making the read request threshold adaptive rather than static. The controller continuously monitors latency and adjusts the threshold in real-time, allowing memory allocation to scale with actual needs. This ensures sufficient memory is reserved during high latency connections while preventing unnecessary memory consumption during low latency connections, resolving the contradiction between reliability and quantity of substance.
Solution Approach 2:
The system performs self-service by automatically monitoring its own latency conditions and adjusting the read request threshold without external intervention. This self-adjusting mechanism ensures memory allocation is always appropriate for current conditions, eliminating the need for over-provisioning memory to handle worst-case scenarios, thus resolving the contradiction between reliability and memory usage.
3Productivity
If a high number of outstanding read requests is maintained to ensure full bus utilization, then data transfer efficiency improves in high latency scenarios, but unwanted timeouts occur in low latency connections
Solution Approach 1:
The patent implements feedback by having the controller monitor latency conditions and use this information to adjust the read request threshold. This closed-loop control prevents timeouts by reducing the threshold when latency is low (preventing excessive outstanding requests) and increases it when latency is high (maintaining efficiency). The feedback mechanism resolves the contradiction between productivity and reliability by adapting to actual connection conditions.
Solution Approach 2:
The dynamic adjustment of the read request threshold based on measured latency allows the system to optimize for data transfer efficiency during high latency periods while preventing timeouts during low latency periods. By making the threshold flexible rather than fixed, the system achieves both high productivity and high reliability across varying connection conditions.
Data Source
AI summary
The present disclosure generally relates to a method and device for improving data read request scheduling in storage devices. When a connection between a host computer and a storage device is established, a controller of the storage device sets a maximum outstanding read requests threshold for the storage device, based on the host-device connection protocol. The maximum outstanding read requests threshold can be dynamically adjusted in response to changes in the host-device connection protocol. Further, a latency tracker of the controller may be used to measure latency between the host computer and the storage device and adjust the maximum outstanding read requests threshold accordingly.