Custom Server Placement Controller for Data Center Infrastructure Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Data centers often face inefficiencies and bottlenecks due to mismatched server configurations and support infrastructure, as actual demand can differ from anticipated demand, leading to underutilization or overload of computing resources.
Innovation Solution
A near-real-time custom server system that uses robots to assemble and install servers with tailored configurations based on current demand, selecting slots with matching infrastructure support requirements, allowing for autonomous or coordinated assembly and placement to optimize resource usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If server configurations are selected based on anticipated demand, then infrastructure support can be designed and built in advance, but actual demand may differ from anticipated demand leading to inefficiencies and bottlenecks
Solution Approach 1:
The patent implements dynamic server configuration selection where the system continuously monitors actual server performance data and automatically adjusts configuration choices based on real-time demand patterns. This allows the infrastructure to transition from static, anticipation-based design to dynamic, demand-driven configuration, resolving the contradiction between advance infrastructure preparation and actual utilization needs.
Solution Approach 2:
The system incorporates feedback mechanisms that collect actual server usage data, performance metrics, and demand information, then use this feedback to refine future configuration selections. This closed-loop approach ensures that infrastructure decisions are based on actual performance rather than predictions, eliminating inefficiencies while maintaining the ability to design infrastructure in advance.
2Adaptability or versatility
If facility operators query customers about planned applications to select server configurations, then server configurations can be optimized for anticipated needs, but actual use may vary from planned use and customers may prefer not to share detailed information
Solution Approach 1:
The system enables servers to self-configure by automatically analyzing their own performance data, resource consumption patterns, and operational requirements to determine optimal configurations. This eliminates the need for customers to disclose detailed application information while still achieving configuration optimization, as the servers themselves provide the necessary intelligence for configuration selection.
Solution Approach 2:
The patent replaces manual information-gathering processes (querying customers directly) with automated systems that infer configuration needs through analysis of server performance metrics and usage patterns. This substitution maintains configuration optimization capabilities while preserving customer privacy by eliminating direct information requests.
3Productivity
If profiling techniques are used to determine applications being executed on servers, then server configurations can be optimized based on actual usage, but such techniques may require access to customer applications and reduce customer privacy or raise security issues
Solution Approach 1:
The system extracts only the necessary configuration information from server operations, separating configuration optimization from deep application analysis. By focusing extraction on high-level performance metrics and resource usage patterns rather than detailed application data, the system achieves configuration optimization without requiring access to customer applications, thus preserving privacy and security.
Solution Approach 2:
The patent employs temporary, disposable analysis mechanisms that process server data locally and discard results after configuration optimization is complete. This approach minimizes the footprint of profiling techniques, ensuring that no persistent storage of sensitive application data occurs, thereby reducing security risks while maintaining optimization capabilities.
4Ease of manufacture
If support infrastructure is designed and built based on anticipated demand for servers, then infrastructure can be prepared in advance, but installed server configurations may differ from anticipated configurations resulting in over supply or under supply of support infrastructure
Solution Approach 1:
The system performs preliminary infrastructure preparation based on projected demand patterns, but incorporates flexible, adaptive components that can be dynamically adjusted as actual server configurations are installed. This allows the infrastructure to be pre-configured for anticipated needs while maintaining the ability to adapt to actual configurations, eliminating both over-supply and under-supply issues.
Solution Approach 2:
The patent implements variable infrastructure parameters that can be adjusted based on actual server configuration data. Instead of fixed infrastructure designs, the system uses adjustable parameters (such as power capacity, cooling requirements, network bandwidth) that automatically adapt to match actual server needs, ensuring precise infrastructure support while maintaining advance preparation capabilities.
Data Source
AI summary
A server placement controller determines a placement location for a custom server based on infrastructure support system requirements of the custom server and based on infrastructure support system capacities at respective unoccupied slots of a server mounting structure of a data center. In some embodiments, a server placement controller may utilize a hierarchical optimization process to select a placement location for a custom server, wherein the selected placement location meets the requirements of the custom server while also optimizing use of one or more infrastructure support systems, such as power infrastructure support system, a networking infrastructure support system, a cooling infrastructure support system, or other infrastructure support systems of a data center.