Storage Device Power Management via Priority Component Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In multi-component devices like storage devices with integrated FPGA and SSD, achieving performance balance while adhering to a limited power budget is challenging due to complex relationships between workloads and power consumption, requiring additional algorithms and sensors for power throttling.
Innovation Solution
A storage device and method that includes a power sensor and global controller to determine priority components operating with fixed performance and non-priority components with variable performance, using power control information from a host device to manage power distribution and performance across components.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by stationary object
If power throttling operation is performed by use of power sensors for respective components, then power consumption can be monitored, but additional algorithm is required and exact powers of components are difficult to predict
Solution Approach 1:
The patent segments the monitoring function by providing separate power sensors for each component (FPGA and SSD) rather than using a single aggregate sensor. This allows independent power measurement for each component, enabling precise power management without requiring complex predictive algorithms. Each sensor directly measures its component's power consumption, simplifying the control logic.
2Adaptability or versatility
If components operate independently, then each component can function autonomously, but exact relationships between workloads and powers are required for performance balance
Solution Approach 1:
The patent implements feedback control by continuously monitoring the actual power consumption of each component via power sensors and using this real-time information to adjust component operations. The controller receives power consumption data from sensors and dynamically balances performance across components based on actual power usage, eliminating the need for pre-established workload-power relationships or complex predictive models.
3Use of energy by stationary object
If performance balance is implemented in multi-component device, then power budget can be maintained, but complex structure and additional algorithms are required
Solution Approach 1:
The patent enables self-service power management where each component is equipped with its own power sensor that directly measures its power consumption. This self-measurement capability allows components to autonomously provide power usage information to the controller, which then automatically balances performance across components. This eliminates the need for complex external monitoring systems or sophisticated algorithms, as the system self-regulates based on direct feedback from component-level sensors.
Data Source
AI summary
A storage device includes a solid state drive (SSD), a field programmable gate array (FPGA), a power sensor and a global controller. The SSD stores data and receives power through a power rail connected to a host device. The FPGA processes data read from the SSD or data to be stored in the SSD and receives power through the power rail. The power sensor is connected to the power rail and generates a measured power value corresponding to a total power consumed by the SSD and the FPGA by measuring the total power. The global controller determines one of the SSD and the FPGA as a priority component operating with a fixed performance and determines the other of the SSD and the FPGA as a non-priority component operating with a variable performance in a priority mode based on power control information provided from the host device.