Real-Time Bottleneck Analysis via UDP Block Storage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional performance monitoring techniques fail to analyze bottlenecks in real-time due to memory burdens, overhead, and limitations in collecting and storing large amounts of performance information, leading to ineffective identification of bottleneck causes and times.
Innovation Solution
A real-time bottleneck analysis method that involves installing a bottleneck analysis component on an application server to collect and store performance information in blocks using UDP, allowing for immediate identification and calculation of bottleneck generation points and times without external expert assistance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If performance information is stored until transaction termination in APM scheme, then performance analysis can be conducted, but memory burden increases and only limited call information can be collected
Solution Approach 1:
The patent segments performance information storage by creating a separate storage space specifically for call information within the stack trace data structure. This allows call information to be preserved without proportionally increasing overall memory burden, as only relevant call stack data is retained rather than all performance metrics until transaction termination.
Solution Approach 2:
The patent extracts and prioritizes call information from the broader performance monitoring data. By focusing specifically on capturing call stack traces and bottlenecks, the system achieves comprehensive call information collection without storing unnecessary performance data that would consume excessive memory resources.
2Loss of information
If all function performance information is collected in profiler scheme, then complete performance data is obtained, but large overhead is caused and integrated performance monitoring is impossible
Solution Approach 1:
The patent extracts only the essential performance information needed for bottleneck analysis - specifically call stack traces and performance metrics at critical points - rather than collecting all function performance data. This selective extraction reduces system overhead while maintaining the ability to identify and analyze performance bottlenecks effectively.
Solution Approach 2:
The patent creates a universal bottleneck analysis mechanism that can monitor multiple applications and services through a common architecture. The bottleneck analysis component can be deployed across different systems to collect and analyze performance data uniformly, enabling integrated performance monitoring without requiring separate profiling mechanisms for each application.
3Measurement precision
If dump analysis scheme is used to collect running function call stack, then current stack state is captured, but actual bottleneck determination relies on estimation and bottleneck time cannot be measured
Solution Approach 1:
The patent implements preliminary action by continuously capturing and storing call stack information and performance metrics in real-time before bottlenecks occur. This allows the system to have pre-recorded data ready for immediate analysis when bottlenecks are detected, eliminating the need for estimation and enabling precise measurement of bottleneck duration and timing.
Solution Approach 2:
The patent establishes a feedback mechanism where performance data is continuously collected, analyzed, and used to identify bottlenecks with precise timing information. The system provides feedback on bottleneck detection, measurement, and resolution, enabling accurate determination of both bottleneck identity and temporal characteristics rather than relying on snapshot-based estimation.
4Speed
If real-time bottleneck analysis is implemented, then immediate performance information is provided, but large amount of performance information must be collected and stored
Solution Approach 1:
The patent extracts and focuses on collecting only the performance information essential for real-time bottleneck analysis - primarily call stack traces, function execution times, and bottleneck indicators - rather than storing all possible performance metrics. This selective data collection enables real-time analysis capability while keeping the volume of stored information manageable.
Solution Approach 2:
The patent changes the parameters of data collection by adjusting what performance metrics are captured and at what frequency. The system prioritizes collecting data with high relevance to bottleneck detection while reducing collection of less critical metrics, optimizing the balance between real-time analysis capability and data storage requirements through parameter optimization.
Data Source
AI summary
The present invention relates to a method for automatically analyzing a bottleneck in real time and an apparatus for performing the method. The method for automatically analyzing a bottleneck in real time may comprise the steps of: an application server receiving a bottleneck analysis component; and the application server installing the bottleneck analysis component, wherein the bottleneck analysis component may add a call code for a performance information collector to an application installed on the application server that is to be monitored, wherein the bottleneck analysis component may call the performance information collector according to execution of a service function of the application, requested by a client, to generate service performance information for analyzing a bottleneck phenomenon.