Server Load Allocation Based on Environmental Suitability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Data centers face overheating issues due to imbalanced processing loads and inefficiencies in cooling systems, which can lead to server shutdowns and economic losses, despite the use of specialized cooling units like CRACs.
Innovation Solution
A method is implemented to allocate processing loads to spaces within the data center based on their suitability for handling additional thermal and computational loads, using a computer server management system that integrates with building automation systems to monitor environmental conditions and distribute workload across air conditioning units, ensuring optimal cooling and reducing heat stress.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If processing loads are allocated without considering environmental conditions, then computational productivity is improved, but overheating risk increases leading to server shutdowns
Solution Approach 1:
The patent applies local quality by evaluating environmental conditions (temperature, humidity, cooling capacity) of specific spaces within the data center and allocating workloads accordingly. Different spaces receive different levels of computational load based on their local cooling suitability, ensuring that hot spots are avoided while maximizing overall productivity.
Solution Approach 2:
The system continuously monitors environmental conditions in data center spaces and uses this feedback to dynamically adjust workload allocation. The suitability index is recalculated based on real-time temperature and cooling capacity data, creating a closed-loop control system that prevents overheating while maintaining high computational throughput.
2Power
If CRAC units are used to enhance cooling power, then cooling capacity is improved, but imbalanced loading and local conditions still cause overheating risks
Solution Approach 1:
The patent makes the cooling system dynamic by continuously evaluating the suitability index based on real-time environmental conditions and adjusting workload allocation accordingly. Rather than static cooling capacity assignment, the system dynamically routes computational loads to spaces with adequate cooling headroom, optimizing the utilization of CRAC units and preventing local overheating.
Solution Approach 2:
The system changes the parameter of workload allocation by introducing the suitability index, which is calculated from multiple environmental parameters (temperature, humidity, cooling capacity). This transforms the cooling management from a fixed capacity approach to a flexible, parameter-based allocation strategy that adapts to changing conditions.
3Ease of operation
If workloads are distributed evenly across all servers, then computational load balancing is improved, but thermal hot spots develop in inadequately cooled areas
Solution Approach 1:
The patent replaces uniform workload distribution with location-aware allocation by evaluating the cooling suitability of each space. Workloads are directed to specific servers based on their local environmental conditions, creating a non-uniform but thermally optimized distribution pattern that prevents hot spots while maintaining overall system balance.
Solution Approach 2:
The system performs preliminary evaluation of space suitability before allocating workloads. By calculating the suitability index in advance based on environmental conditions and cooling capacity, the system proactively prevents thermal hot spots rather than reacting to temperature issues after they occur.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method includes generating a space information value for each of a plurality of spaces based on at least one environmental condition measurement for the corresponding space. Each space includes one or more computing devices. The space information value includes information regarding the relative suitability of a corresponding space for accepting computing load. The method also includes determining an allocation of additional computing load based on the space information values.