Automated Server Ranking for Power-Constrained Service Placement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data center management systems lack an efficient method for automated service profile placement across servers with varying power characteristics, leading to increased power consumption and operational costs, as well as manual errors in server selection due to the complexity of handling multiple servers and priorities.
Innovation Solution
A system and method for automated service profile placement that ranks servers based on power characteristics, including power group caps, server capabilities, and PSU capacity, using a unified computing system manager to select the most suitable server for placement, considering both power and non-power related constraints, to optimize power allocation and reduce manual errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual server selection is used for service profile placement, then flexibility in handling complex constraints is maintained, but errors increase and efficiency decreases
Solution Approach 1:
The system performs automated self-assessment of server power characteristics, capabilities, and constraints. The server pool automatically ranks itself based on power group caps, PSU capacities, and other parameters, eliminating manual evaluation while maintaining accuracy through structured automated decision-making
Solution Approach 2:
The patent replaces manual mechanical selection processes with an automated computational ranking system. The system uses algorithms to calculate scores based on power characteristics, capability metrics, and constraint satisfaction, substituting human judgment with deterministic computational evaluation
2Use of energy by stationary object
If servers are manually selected without automated ranking, then system complexity is reduced, but power consumption optimization is lost
Solution Approach 1:
The system segments the server pool into power groups based on power characteristics and caps. Each server is evaluated independently on multiple dimensions (power consumption, capability, constraints) and ranked separately, allowing systematic optimization without overwhelming complexity
Solution Approach 2:
The patent transforms qualitative server characteristics into quantitative parameters that can be systematically compared. Power consumption, capability scores, and constraint satisfaction levels are converted into numerical values that feed into the ranking algorithm, enabling automated optimization
3Productivity
If automated placement without power-based ranking is used, then processing speed is maintained, but power allocation efficiency deteriorates
Solution Approach 1:
The system performs preliminary ranking of all servers in the pool before actual service profile placement occurs. Power characteristics, capabilities, and constraints are pre-evaluated and scored, so that when placement is needed, the system can quickly select from pre-ranked options without sacrificing optimization
Solution Approach 2:
The ranking system operates continuously or periodically to maintain an up-to-date ordered list of servers based on current power characteristics and constraints. This continuous maintenance of the ranking allows rapid placement decisions while ensuring power optimization is always applied
Data Source
AI summary
A method includes ranking a plurality of servers in a network environment according to power characteristics, and choosing a server from the plurality of servers for placement of a service profile according to at least a priority specified in the service profile and the server's rank. The ranking includes identifying power groups into which the plurality of servers are partitioned, identifying respective power group caps of the power groups, determining respective power ranges and respective power supply multipliers of the plurality of servers, calculating a score of each server in the plurality of servers, and assigning ranks to the scores. Choosing the server includes partitioning the servers into priority sets according to respective ranks of servers, and searching the priority sets beginning with a priority set having the priority specified in the service profile until a suitable server is found.