Modular Cooling Delivery Subsystem for Segregated Data Center PoDs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional data center infrastructure struggles to efficiently and cost-effectively support new computing architectures with segregated compute and storage point of deliveries (PoDs), particularly in managing thermal loads due to increasing heat generation from high-performance electronics, and lacks adaptability to varying hardware configurations.
Innovation Solution
The design incorporates flexible cooling delivery subsystems using a combination of liquid and air cooling, with independently controlled cooling capacity based on air temperature feedback loops, modular cooling units, and fluid distribution piping to manage thermal loads across different PoD configurations, treating the entire data center as a cold aisle and allowing for scalable and adaptable thermal management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional data center infrastructure is used, then initial deployment is simple, but it cannot efficiently support new computing architectures with segregated compute and storage PoDs
Solution Approach 1:
The data center is divided into multiple Point of Delivery (PoD) modules, each containing segregated compute and storage racks. Each PoD is a self-contained unit with its own cooling delivery subsystem, allowing independent configuration and deployment. This modular segmentation enables the infrastructure to adapt to different computing architectures without requiring complete system redesign.
Solution Approach 2:
The cooling delivery subsystem is designed with universal components that can serve multiple PoD configurations. The liquid cooling manifolds and piping systems are configured to accommodate various rack arrangements and power densities. This universal design allows the same infrastructure framework to support different compute/storage segregation ratios and hardware configurations.
2Productivity
If high-performance electronics components are increased, then computing performance is improved, but heat generation and thermal management difficulty increase
Solution Approach 1:
The patent implements liquid cooling systems using hydraulic principles to remove heat from high-performance electronics. Cooling liquid circulates through manifolds and tubing positioned near heat-generating components in compute and storage racks. This hydraulic cooling approach efficiently transfers heat from dense electronics configurations without the limitations of air cooling, enabling sustained high computing performance.
Solution Approach 2:
The cooling system is configured with local quality variations to match heat generation patterns. Different PoDs and rack positions receive customized cooling capacity based on their specific thermal loads. The liquid cooling manifold system allows selective adjustment of cooling flow to different zones, providing optimal thermal management for high-performance components while energy efficiency.
3Adaptability or versatility
If cooling infrastructure is made flexible to adapt to changing hardware configurations, then adaptability is improved, but system complexity and difficulty of implementation increase
Solution Approach 1:
The cooling delivery subsystem incorporates dynamic adjustment capabilities through independently controllable cooling units. Each PoD's cooling capacity can be adjusted based on real-time thermal requirements and hardware configurations. This dynamic control allows the cooling infrastructure to adapt to changing computing workloads and hardware arrangements without manual reconfiguration of the entire system.
Solution Approach 2:
The cooling system is segmented into independently controllable units at the PoD level, with further subdivision into rack-level or even component-level control zones. This segmentation allows flexible adaptation to different hardware configurations by activating or adjusting only the necessary cooling zones, rather than managing the entire cooling system as a single complex unit.
4Productivity
If compute and storage PoDs are segregated, then computing architecture efficiency is improved, but thermal management and facility infrastructure requirements become more complex
Solution Approach 1:
The facility infrastructure is segmented to match the segregated PoD architecture. Each PoD has dedicated cooling delivery subsystems and power distribution units, allowing independent management of compute and storage thermal loads. This segmentation simplifies the overall infrastructure by creating modular, self-contained units rather than requiring a monolithic complex system.
Solution Approach 2:
The facility infrastructure employs universal designs that can serve both compute and storage PoDs with the same basic systems. The liquid cooling manifold design, power distribution architecture, and facility layout are configured to accommodate various compute/storage segregation ratios, reducing the need for specialized infrastructure for each PoD type.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enables uniform thermal management across varying compute and storage PoDs, maintaining optimal temperatures and extending the life of electronic components, while allowing for easy scalability and maintenance, thus enhancing efficiency, reliability, and adaptability to changing IT requirements.
Implementation Method 1
a first loop with inlet and outlet ports having a first pair of liquid connectors coupled to liquid supply/return lines that are integrated to the electronic rack to form a primary loop
Implementation Method 2
one or more cooling fans associated with the server blades to provide air cooling
Data Source
AI summary
Disclosed are layouts, arrangements, and configurations of room airflow, server racks and cooling delivery subsystems that support data centers having system architectures with segregated compute and storage point of deliveries (PoDs). A cooling row is arranged in the rear side of racks housing a PoD to supply the PoD with the cooling delivery subsystem, which includes a recirculating loop supplying cooling fluid to the PoD and returning warm fluid to the cooling source. Modular cooling units of the cooling row includes cooling coils used to cool the airflow leaving the racks. The cooling capacity of the cooling liquid supplied to the cooling delivery subsystem for a PoD may be independently adjusted and controlled in a feedback loop based on the temperature of the air of the PoD after it has been cooled by the cooling coils, such as the outlet air temperature from the cooling row housing the PoD.