Load Estimation via Execution Address Frequency Distribution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods struggle to accurately measure the load of applications that cannot be assessed through CPU utilization in the operating system, particularly in virtualized environments where CPU utilization is consistently high or indistinguishable from idle states.
Innovation Solution
A system that collects and analyzes execution address samples from a CPU to differentiate between load levels by utilizing frequency distribution patterns, allowing for accurate load estimation even when CPU utilization is consistently high or indistinguishable from idle states.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If CPU utilization measurement is used to assess application load, then the measurement method is simple and widely available, but it fails to accurately distinguish load levels in virtualized environments where CPU utilization is consistently high
Solution Approach 1:
The patent segments the measurement approach by dividing execution addresses into multiple blocks and analyzing the frequency distribution pattern across these blocks. Instead of relying on a single CPU utilization metric, the system segments the address space into blocks and evaluates the distribution characteristics within each block to determine load levels, thereby achieving accurate load measurement in virtualized environments.
2Measurement precision
If execution address sampling is performed to differentiate load levels, then load estimation accuracy is improved, but the complexity of data collection and analysis increases
Solution Approach 1:
The patent extracts only the essential feature needed for load differentiation - the frequency distribution pattern of execution addresses across blocks. Instead of analyzing all execution details, the system extracts and analyzes specifically the occupancy rates of address blocks, filtering out unnecessary information and focusing on the discriminatory pattern that indicates load levels.
Solution Approach 2:
The patent creates a simplified representation (copy) of the execution status by recording frequency distribution patterns of address blocks. This copied pattern data serves as a surrogate for complex execution analysis, enabling load level determination through comparison of distribution patterns rather than direct analysis of raw execution traces.
3Measurement precision
If frequency distribution patterns of execution addresses are analyzed, then accurate load differentiation is achieved, but the processing overhead and computational resources increase
Solution Approach 1:
The patent applies partial action by analyzing only a representative subset of execution addresses organized in blocks, rather than processing every execution event. By sampling and analyzing the frequency distribution across blocked address ranges, the system achieves sufficient load differentiation accuracy with reduced computational overhead compared to comprehensive analysis.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A disclosed load estimation method includes: collecting run information of a processor being executing a predetermined program; specifying execution status of the processor based on the collected run information; and estimating a load of the predetermined program based on a result of comparison between the execution status of the processor and execution characteristics of the processor. Each of the execution characteristics is stored in association with a load level of the predetermined program.