Automatic Profiling Report Generation for Distributed Systems

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Manual profiling in highly distributed systems, such as cloud applications with hundreds of nodes, is inefficient due to the need for manual analysis of profiling data across all nodes, which does not scale well and requires significant human interaction.

Innovation Solution

An automatic profiling report generation tool that allows users to specify analysis steps which can be executed automatically, using a client-server system with a front-end and back-end profiler, enabling the collection and analysis of profiling data without human intervention, and generating reports that highlight performance bottlenecks.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual profiling analysis is performed on all application nodes in highly distributed systems, then comprehensive performance analysis is achieved, but the time and effort required increases significantly and scalability is poor

Engineering Contradiction:
Improvecomprehensive performance analysisVSAvoidanalysis time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent extracts only the critical application nodes from the entire distributed system that require profiling attention. Instead of manually analyzing all nodes, the system automatically identifies and extracts the subset of nodes with performance issues, enabling focused analysis that maintains comprehensiveness where needed while dramatically reducing overall analysis time and effort.

Inventive Principle:
Principle #2Taking out (Extraction)

2Measurement precision

If manual profiling analysis is performed across all nodes, then complete system coverage is achieved, but the complexity of the process increases and requires significant human interaction

Engineering Contradiction:
Improvesystem coverageVSAvoidprocess complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system performs self-service by automatically identifying critical nodes and executing profiling analysis without requiring human interaction for each node. The automated system serves itself by selecting which nodes need analysis based on predefined criteria, eliminating the need for manual process management while maintaining complete system coverage through automated node selection and analysis execution.

Inventive Principle:
Principle #25Self-service

3Productivity

If profiling experts are used to reduce analysis time, then analysis speed improves, but the requirement for human interaction limits scalability in highly distributed systems

Engineering Contradiction:
Improveanalysis speedVSAvoidautomation level
Core Design Contradiction:
ProductivityVSExtent of automation

Solution Approach 1:

The patent replaces the mechanical system of human expert analysis with an automated computational system. Instead of relying on profiling experts to manually analyze each node, the system uses automated algorithms and tools to perform the analysis, substituting human cognitive work with machine-based processing. This maintains high analysis speed while dramatically improving automation level and scalability to highly distributed systems.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS9547537B2Automatic profiling report generation
Publication Date: 2017.01.17 SAP SE
  • US9547537B2 patent drawing
  • US9547537B2 patent drawing
  • US9547537B2 patent drawing

AI summary

A method for profiling an application on a virtual machine is provided. A series of analysis steps to be performed on profiled data can be created. The series of analysis steps can be saved as a report specification. A back-end profiler can then be caused to perform profiling on the application. Profiled data can be received from the back-end profiler. The profiled data can be stored as a model. The model can then be adapted based on the series of analysis steps from the report specification. Output data can be generated based on the adapted model. Finally, the output data is displayed to a user.