BMC Thermal Control for Unmanaged Component Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing thermal management techniques for server hardware face challenges in accurately acquiring temperature data from unmanaged components due to the inability of the baseboard management controller (BMC) to directly access these components, leading to inefficient cooling, increased energy consumption, and operational costs, as well as downtime during component changes.
Innovation Solution
A schema is implemented to coordinate between the host system and the BMC, enabling the host agent to collect temperature data from unmanaged components and communicate it to the BMC, using identical schema copies in both systems to ensure accurate cooling unit adjustments based on internal sensor data, without requiring modifications to existing systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If proximity sensors are added to estimate temperatures of unmanaged components, then the BMC can receive temperature data from unmanaged components, but the measurement precision deteriorates because proximity sensor readings are not as accurate as internal sensor data
Solution Approach 1:
The patent introduces a host agent as an intermediary component that bridges the gap between unmanaged components and the BMC. The host agent directly accesses temperature data from unmanaged components through their internal sensors and communicates this accurate data to the BMC, eliminating the need for proximity sensors and their inherent measurement inaccuracies.
Solution Approach 2:
The patent segments the thermal management system into distinct functional components: managed components (directly accessible by BMC), unmanaged components (requiring host agent assistance), and the host agent (which facilitates data collection). This segmentation allows each component to operate optimally while working together to solve the data accessibility problem.
2Reliability
If manually configured control profiles are used to account for proximity sensor inaccuracies, then thermal management can be maintained, but the use of energy increases due to overcooling margins
Solution Approach 1:
The patent implements a feedback mechanism where the host agent continuously collects accurate temperature data from unmanaged components and transmits it to the BMC. This real-time feedback loop allows the BMC to receive precise temperature information and adjust cooling levels accordingly, eliminating the need for conservative overcooling margins and reducing energy waste.
3Loss of information
If proximity sensors are deployed to monitor unmanaged components, then thermal data can be collected, but the device complexity increases
Solution Approach 1:
The patent enables unmanaged components to essentially serve themselves by having their internal temperature sensors directly accessed by the host agent. This eliminates the need for additional proximity sensors and their associated complexity, as the system uses the components' own built-in sensing capabilities.
4Reliability
If firmware changes are made to accommodate component additions or removals, then thermal control can be maintained, but the loss of time increases due to system reboots and downtime
Solution Approach 1:
The patent implements a dynamic configuration system where the host agent can automatically detect and adapt to component additions or removals without requiring firmware changes. The system dynamically updates its understanding of which components are present and their thermal characteristics, allowing for hot-add and hot-remove operations without downtime.
Data Source
AI summary
The disclosed techniques provide enhanced thermal management control for unmanaged components of a host system via a baseboard management controller (“BMC”). The techniques disclosed herein utilize a schema that defines parameters for controlling the collection of temperature data from temperature sensors inside unmanaged components. The schema enables a host agent of the host system to coordinate with the BMC and communicate temperature data acquired from unmanaged components to the BMC. Once the BMC receives the temperature data, the BMC communicates the temperature data to a control profile to accurately control the cooling level of one or more cooling units in the host system.


