Transaction Analysis Tool for Integrated Computing Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for estimating transaction processing performance in integrated computing systems are inefficient, requiring manual iterations and trial-and-error approaches to achieve adequate performance levels, especially under varying loading conditions, leading to time-consuming and suboptimal customization.
Innovation Solution
A transaction analysis tool that measures elapsed transaction times at different loading levels and uses a curve fitting algorithm to generate a function estimating expected performance, allowing designers to customize the system while maintaining performance objectives.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If manual iterations and trial-and-error methods are used to estimate transaction processing performance, then system customization can be achieved, but the process becomes time-consuming and inefficient
Solution Approach 1:
The patent applies preliminary action by pre-establishing a performance estimation model that uses historical transaction data and system configuration parameters to predict performance metrics before actual system deployment or customization. This allows designers to estimate transaction processing performance upfront, avoiding time-consuming manual iterations while still achieving accurate system customization.
2Reliability
If sufficient computing resources are purchased to meet varying business demands, then adequate capacity is ensured, but excess capacity represents wasted investment
Solution Approach 1:
The patent implements feedback by continuously monitoring actual system performance metrics and comparing them against predicted values from the estimation model. This feedback loop enables dynamic adjustment of resource allocation, allowing organizations to optimize their computing resource purchases based on actual usage patterns rather than over-provisioning for peak demands, thus eliminating wasted investment while maintaining reliability.
Solution Approach 2:
The patent applies parameter changes by using the performance estimation model to analyze how variations in system configuration parameters (such as number of processors, memory size, storage capacity) affect transaction processing performance. This allows organizations to identify the optimal parameter组合 that meets performance requirements at minimum cost, avoiding both over-investment and insufficient capacity.
3Loss of energy
If too few computing resources are purchased, then investment cost is reduced, but insufficient capacity damages business operations
Solution Approach 1:
By using the performance estimation model to predict system performance under different resource configurations before deployment, organizations can identify the minimum resource level that still meets performance requirements for continuous business operations. This preliminary analysis prevents under-provisioning while minimizing investment cost.
Data Source
AI summary
A transaction analysis system includes a computer-executable tool for obtaining first and second measured elapsed times to complete a transaction on at least one of the resources of an integrated computing system at first and second loading levels, respectively. The tool then generates, using a curve fitting algorithm, a function according to the first and second measured elapsed times to complete the transaction at the first and second loading level. The resulting function indicates an expected level of performance of the transaction at varying degrees of loading levels.