Voltage Regulator Fault Data Capture via Synchronized Volatile Memory Caching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Determining the cause of voltage regulator failure in computer systems is difficult and time-consuming due to the complexity of reconstructing fault causes in retroactive analysis, as voltage regulators are prone to failure from various stressors and often operate in series or parallel configurations.
Innovation Solution
A system comprising multiple voltage regulators with controllers, volatile memory for temporary data caching, and non-volatile memory for storing fault data, featuring a common signal line that alerts other regulators to faults, allowing for synchronized data capture and storage of operating data prior to a fault event, facilitating analysis by management entities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If voltage regulators operate without synchronized fault data capture, then device complexity is reduced, but loss of information increases due to inability to reconstruct fault causes
Solution Approach 1:
The patent implements preliminary action by having voltage regulators continuously cache operating data in volatile memory during normal operation, and by pre-configuring the system with non-volatile memory and open drain outputs ready to capture and preserve fault data immediately when a fault occurs, eliminating the need for post-fault reconstruction
Solution Approach 2:
The patent applies copying by creating redundant copies of fault data: each voltage regulator copies its operating data from volatile memory to non-volatile memory upon detecting a fault, and additional copies are transmitted via open drain outputs to other regulators and management entities, ensuring data preservation even if the primary regulator fails
2Loss of time
If retroactive analysis is used to determine voltage regulator failure causes, then device complexity is minimized, but loss of time increases due to difficulty in reconstructing fault causes
Solution Approach 1:
The system performs preliminary action by continuously caching operating data in volatile memory during normal operation and automatically transferring it to non-volatile memory when a fault is detected, so that fault data is already prepared and preserved before analysis is needed, eliminating time-consuming retroactive reconstruction
Solution Approach 2:
The voltage regulators perform self-service by automatically detecting faults, capturing their own operating data, transferring it to non-volatile memory, and sending notifications to management entities without requiring external intervention or complex post-fault analysis systems
3Ease of operation
If fault data is not automatically captured and stored, then ease of operation is reduced, but productivity is improved by avoiding additional data processing overhead
Solution Approach 1:
The voltage regulators perform self-service by automatically detecting faults, capturing operating data from volatile memory, storing it in non-volatile memory, and notifying management entities without requiring external intervention, making fault data analysis easy while minimizing processing overhead through localized autonomous operation
Solution Approach 2:
The system implements feedback by having voltage regulators continuously monitor their own operating parameters, automatically detect when faults occur, and immediately trigger data capture and notification processes, providing real-time feedback that simplifies fault analysis without requiring continuous external monitoring
Data Source
AI summary
A method includes a plurality of voltage regulators distributing power to a plurality of components within a compute node, wherein each of the voltage regulators has a controller, volatile memory and non-volatile memory. The controller of each voltage regulator temporarily caches operating data for the voltage regulator in the volatile memory of the voltage regulator, wherein the controller temporarily caches the operating data collected from the voltage regulator over a sliding time period. When a first voltage regulator from among the plurality of voltage regulators experiences a fault event, the controller of each voltage regulator detects the fault event and automatically copies the cached operating data for the voltage regulator from the volatile memory of the voltage regulator to the non-volatile memory of the voltage regulator in response to detecting the fault event.