Dynamic Thermal Load Balancing in Data Centers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Data centers face significant thermal stress due to localized 'hot spots' caused by uneven heat distribution, which existing hardware-based solutions struggle to address efficiently and economically, especially after provisioning, leading to potential damage and reduced resource efficiency.
Innovation Solution
Implementing a method that uses a data orchestrator to migrate computing workloads defined as virtual machines from areas of excessive heat to areas of lesser heat, based on temperature thresholds and sensor data, without altering cooling capacity or powering down resources, thereby optimizing thermal efficiency in real-time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Temperature
If hardware-based cooling methods are increased to address thermal stress, then temperature control is improved, but energy consumption and operational cost increase
Solution Approach 1:
The patent implements dynamic workload migration that continuously monitors temperature conditions and automatically relocates computing workloads between computing devices. This dynamic adjustment allows the system to respond to changing thermal conditions in real-time, reducing reliance on static hardware-based cooling solutions and thereby lowering energy consumption while maintaining effective temperature control.
Solution Approach 2:
The system changes the operational parameters by migrating workloads between computing devices based on temperature thresholds and thermal conditions. Instead of changing cooling capacity, the patent alters the distribution of computational tasks across devices, effectively using parameter changes in workload allocation to manage thermal stress and reduce cooling energy requirements.
2Temperature
If computing resources are powered down to reduce thermal stress, then temperature is reduced, but computing capacity and productivity decrease
Solution Approach 1:
The patent employs dynamic workload migration that moves computing tasks from thermally stressed devices to devices with adequate thermal headroom. This dynamic redistribution maintains full computing capacity across the data center by ensuring that workloads are continuously executed on available devices, eliminating the need to power down resources while still addressing thermal stress effectively.
Solution Approach 2:
The system creates copies of computing workloads and relocates them to different computing devices. By copying workloads rather than terminating them, the patent maintains computing capacity while redistributing thermal load, allowing the original device to cool down without losing productive capacity.
3Temperature
If workloads are migrated to address hot spots, then thermal efficiency is improved, but system complexity and migration overhead increase
Solution Approach 1:
The patent implements a feedback-driven workload migration system that monitors temperature conditions and automatically triggers migration only when thermal thresholds are exceeded. This feedback mechanism simplifies the system by making migration decisions based on clear, predefined thermal conditions rather than continuous complex optimization, reducing overall system complexity while maintaining thermal efficiency.
Solution Approach 2:
The system uses parameter-based decision making where workload migration is triggered by specific temperature threshold parameters. By establishing clear parameter thresholds for migration decisions, the patent reduces the complexity of determining when and where to migrate workloads, making the system easier to manage while achieving thermal efficiency goals.
Data Source
AI summary
A method for improving thermal efficiency in one or more data centers includes measuring a temperature at one or more computing devices having allocated thereto one or more computing workloads and determining whether the measured temperature exceeds a predetermined temperature threshold. If the predetermined temperature threshold is exceeded, sufficient computing workloads are migrated to one or more alternate computing devices to reduce the temperature at one or more of the computing devices to less than or equal to said predetermined temperature threshold. In accomplishing the method, there is provided a data orchestrator configured at least for receiving data concerning the measured temperature, determining whether the measured temperature exceeds the predetermined temperature threshold, and migrating one or more of the computing workloads to one or more alternate computing devices. Computer systems and computer programs available as a download or on a computer-readable medium for installation according to the invention are provided.