Data Center Cluster Layout for Dynamic Power and Cooling Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional data centers are inefficient and expensive due to overdesign to accommodate the highest conceivable power and cooling capacity needs, leading to significant resources going unused.
Innovation Solution
A fungible data center design with adjustable power and cooling capacity distribution among clusters and individual racks, allowing for dynamic allocation of resources without real-time physical reconfiguration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data centers are overdesigned to accommodate the highest conceivable power and cooling capacity needs, then reliability is improved, but loss of energy worsens due to significant resources going unused
Solution Approach 1:
The patent implements dynamic power and cooling allocation by dividing the data center into multiple independently coolable clusters with adjustable power demands. The system can dynamically increase or decrease the actual power demand of individual clusters based on real-time needs, transforming the static overdesigned infrastructure into a flexible system that adapts to varying loads without requiring physical reconfiguration.
2Ease of operation
If data centers are designed with fixed equal allocation of power and cooling to each rack, then ease of operation is improved, but adaptability worsens when power and cooling needs vary across different racks
Solution Approach 1:
The patent segments the data center into N independently coolable clusters, each with its own adjustable power demand parameters. This segmentation allows each cluster to be independently controlled and optimized based on its specific requirements, while the overall system maintains manageable complexity through standardized interfaces and centralized monitoring.
3Manufacturing precision
If data centers use traditional homogeneous rack configurations, then manufacturing precision is improved through standardization, but productivity worsens due to inefficient resource utilization
Solution Approach 1:
The patent applies local quality by allowing different clusters to have different actual power demands and cooling requirements while maintaining standardized rack physical structures. Each cluster can be optimized for its specific workload characteristics (e.g., high-performance computing, storage, networking) without requiring custom rack designs, thus achieving both manufacturing efficiency and operational optimization.
Data Source
AI summary
Provided, in one aspect, is a data center. The data center, in this aspect, includes a data center enclosure, the data center enclosure designed for a given supply of power (Ps). The data center, according to this aspect, further includes N independent coolable clusters of data center racks located within the data center enclosure, wherein N is at least two, and further wherein the N independent coolable clusters each have an ostensible power demand (Pos) approximately equal to Ps/N, and each of the N independent coolable clusters has a respective actual power demand (Pac) adjustable at, above or below the ostensible power demand (Pos).