Management Computer Node Association Analysis for System Performance Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
It is challenging to estimate the performance limits of a computer system and operations system due to their independent nature and varying configurations, making it difficult to scale out and handle system changes, especially when performance is not uniform or when clusters are formed with low-cost servers.
Innovation Solution
A system management method that analyzes the configuration of the computer system to specify important nodes, changes resource allocation, calculates the strength of associations among task nodes, and determines the impact range of load changes to estimate performance limits through a stress test.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the scale of the data center is expanded by scaling out with more computer resources, then the capacity and performance of the system is improved, but it becomes difficult to estimate the performance limits and determine the appropriate units for increasing resources
Solution Approach 1:
The patent applies preliminary action by performing stress tests before actual scaling operations to establish performance baselines and limits. The management computer executes stress tests on the operations system to measure performance characteristics under various load conditions, allowing the provider to estimate performance limits before committing to scaling decisions. This proactive measurement approach enables informed planning of resource expansion without blindly adding resources.
2Adaptability or versatility
If the configuration of the operations system is customized for each user, then the adaptability and service quality are improved, but the complexity of managing and estimating performance across different configurations increases
Solution Approach 1:
The patent implements feedback mechanisms where the management computer continuously monitors and measures the performance of operations systems with different user-specific configurations. By executing stress tests and collecting performance data from various configured systems, the management computer builds a knowledge base that feeds back into future scaling decisions. This feedback loop allows the system to learn from actual performance measurements of different configurations, reducing management complexity over time while maintaining adaptability.
3Measurement precision
If stress tests are performed on the entire operations system, then comprehensive performance data is obtained, but the time and resources required for testing increase significantly
Solution Approach 1:
The patent applies segmentation by dividing the operations system into individual task nodes for separate stress testing. Instead of testing the entire system as a monolithic unit, the management computer executes stress tests on each task node independently or in controlled combinations. This segmented approach reduces the overall testing time and resource requirements while still providing comprehensive performance data across different system configurations and load scenarios.
Data Source
AI summary
A system management method for a management computer coupled to a computer system, the computer system including a plurality of computers, an operations system being built thereon the computer system, the operations system including a plurality of task nodes each having allocated thereto computer resources, the system management method including: a step of analyzing a configuration of the computer system for specifying a important node, which is an important task node in the operations system; a step of changing an allocation amount of the computer resources allocated to the important node for measuring a load of the operations system; a step of calculating a first weighting representing a strength of associations among the plurality of task nodes based on a measurement result of the load; and a step of specifying a range impacted by a change in the load of the important node based on the calculated first weighting.