Server Cluster Selection Using Failure-Risk Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies fail to consider the varying failure risks of physical servers due to factors like operation time and temperature, leading to potential selection of servers with higher failure risks as placement destinations for virtual machines or containers.
Innovation Solution
A server group selection system that classifies physical servers into clusters based on failure factors using parameters such as disk write count, operation time, and temperature, and selects clusters with lower failure probabilities as placement destinations for virtual machines or containers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If physical servers are selected for placement without considering failure factors, then placement simplicity is maintained, but system reliability deteriorates due to potential selection of servers with higher failure risks
Solution Approach 1:
The patent segments physical servers into multiple clusters based on failure factors such as operation time, temperature, and disk write counts. Each cluster represents servers with similar failure risk profiles, allowing the placement system to select from multiple segmented groups rather than treating all servers uniformly. This segmentation enables reliability-based selection without requiring complex individual server evaluation.
Solution Approach 2:
The patent performs preliminary classification of physical servers into clusters based on failure factors before the actual placement decision is made. By pre-grouping servers according to their reliability characteristics, the system eliminates the need for complex real-time evaluation during placement, thus improving reliability while maintaining placement simplicity.
2Measurement precision
If servers are classified into multiple clusters based on failure factors, then placement precision is improved by avoiding high-risk servers, but calculation complexity increases due to additional classification processing
Solution Approach 1:
The patent divides the server population into distinct clusters based on failure factors, enabling precise placement decisions by selecting from predefined groups. This segmentation approach improves placement precision by ensuring virtual machines are placed on servers with appropriate risk profiles, while the clustering algorithm manages calculation complexity through efficient grouping.
Solution Approach 2:
The patent changes the evaluation parameters from simple resource availability to include failure factors such as operation time, temperature, and disk write counts. By incorporating these additional parameters into the clustering process, the system achieves more precise placement decisions that account for server reliability, while the parameter-based approach maintains computational efficiency.
3Reliability
If failure factors are considered in server selection, then availability is improved by reducing simultaneous failure risks, but information processing requirements increase due to collection and analysis of additional server parameters
Solution Approach 1:
The patent collects and analyzes failure factor information from physical servers in advance, before placement decisions are made. By performing this information processing preliminarily and storing the results in cluster classifications, the system improves availability through informed selection without requiring intensive real-time data processing during placement operations.
Solution Approach 2:
The patent creates simplified representations of server states through clustering, where each cluster serves as a copy or summary of servers with similar failure characteristics. This copying approach reduces the volume of information that needs to be processed during placement, as the system works with cluster-level summaries rather than individual server details.
Data Source
AI summary
A system (S) includes: a classification unit (21) that classifies a plurality of physical servers into a plurality of clusters based on parameters collected from the physical servers; and a destination selection unit (22) that selects, from the plurality of clusters classified by the classification unit (21), any of clusters capable of ensuring resources required for placing virtual machines or containers to be added as a placement destination. Accordingly, the system (S) can determine a server group as the placement destination based on the failure factors of the physical servers.