Multivariable Controller for Data Center IT and Facilities Coordination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Modern data centers face significant power consumption challenges due to the large number of computing servers, which leads to high electricity costs and inefficiencies in cooling demands, as conventional control systems separate IT and facilities management, resulting in suboptimal resource utilization.
Innovation Solution
A multivariable controller is implemented to coordinate the control of computing devices and building infrastructure, leveraging information from thermal management, HVAC, and power systems to adjust server modes and workload distribution, using model predictive control techniques to minimize power consumption while ensuring adequate computing resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional separate control systems are used for IT and facilities management, then system simplicity is maintained, but resource utilization efficiency deteriorates
Solution Approach 1:
The patent merges previously separate IT control systems and facilities control systems into a unified multivariable control platform. This integration allows coordinated optimization of computing devices, thermal management, HVAC, and power systems, directly resolving the contradiction by improving resource utilization efficiency through combined control while accepting increased system complexity as a necessary trade-off for achieving superior overall performance and energy efficiency.
2Productivity
If the number of computing servers is increased to meet performance demands, then computing capacity is improved, but power consumption increases
Solution Approach 1:
The patent implements dynamic control that continuously adjusts server operating modes, workload distribution, and facility system operations based on real-time conditions and predictive models. This dynamic optimization enables the system to maintain required computing capacity while minimizing power consumption by adapting resource allocation to actual demand patterns rather than operating at fixed high-capacity states.
Solution Approach 2:
The multivariable control system changes operational parameters across multiple systems including server CPU frequencies, server states (active/standby), HVAC settings, and thermal management parameters. By optimizing these parameters collectively through model predictive control, the system achieves high computing capacity with reduced power consumption compared to traditional static parameter settings.
3Productivity
If server workload is increased to maximize utilization, then productivity is improved, but cooling demands increase
Solution Approach 1:
The system uses predictive models to forecast future workload patterns and thermal conditions, allowing preemptive adjustment of server workloads and cooling system operations. By taking preliminary actions based on predictions rather than reactive responses to current conditions, the system optimizes the balance between server utilization and cooling energy consumption, preventing excessive cooling demands before they occur.
Solution Approach 2:
The multivariable control system implements closed-loop feedback that continuously monitors server workload, thermal conditions, and cooling system performance. This feedback enables real-time optimization of workload distribution and cooling operations, adjusting server utilization levels and cooling intensity to maintain productivity while minimizing cooling energy consumption based on actual system state.
Data Source
AI summary
A method includes obtaining first information associated with control of multiple computing devices, where the first information relates to possible changes to processing tasks performed by the computing devices. The method also includes obtaining second information associated with building infrastructure operations performed by one or more building systems of one or more buildings that house the computing devices. The method further includes identifying one or more changes to one or more of the computing devices using the first and second information. In addition, the method includes outputting third information identifying the one or more changes.