Management System Read-Ahead Performance Data Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In large and complex computer systems, predicting which performance data to display as a read-ahead target is challenging due to the vast amount of time-series data collected regularly or irregularly, making it difficult to accurately select relevant data for display based on access history.
Innovation Solution
Defining event types for which performance data display is requested in advance, the management system selects relevant components directly or indirectly coupled to the event component as read-ahead targets, and reads their performance data from storage to memory when necessary, allowing for partial or full data retrieval.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If performance data of all components is stored in memory for fast display, then display speed is improved, but memory capacity is insufficient
Solution Approach 1:
The system performs preliminary actions by detecting event types that may request performance data display in the future, and proactively reads relevant performance data from storage to memory before actual display requests occur. This anticipatory approach ensures data is readily available when needed, improving display speed without requiring all data to be permanently stored in memory.
Solution Approach 2:
Instead of uniformly storing all performance data in memory, the system selectively loads only the specific performance data relevant to detected event types and their associated components. This targeted approach optimizes memory usage by maintaining only the necessary data locally, while other data remains in storage.
2Quantity of substance
If performance data is read from storage to memory on demand, then memory capacity requirement is reduced, but display response time increases
Solution Approach 1:
The system detects event types in advance and proactively reads relevant performance data from storage to memory before actual display requests are made. This preliminary loading eliminates the time delay that would otherwise occur when data is read on demand, significantly reducing display response time while maintaining efficient memory usage.
3Speed
If read-ahead is performed for all performance data, then display speed is improved, but data selection accuracy decreases
Solution Approach 1:
The system performs preliminary detection of specific event types that may lead to performance data display requests, and only reads performance data relevant to those detected event types. This targeted preliminary action improves data selection accuracy by focusing on relevant data, while still achieving speed improvement through proactive loading of the correct data sets.
Solution Approach 2:
The system applies different handling strategies to different types of data based on their relevance to detected event types. By identifying and selectively loading only the performance data associated with detected events and their related components, the system maintains high data selection accuracy while improving display speed.
Data Source
AI summary
As a type of an event that occurs in a computer system, an event type for which display of performance data is to be requested in the future is defined in advance. A management system, upon detection of occurrence of an event that belongs to the event type for which it is defined that display of performance data is to be requested in the future, selects, as a read-ahead target, at least one of an event component which is a component in which the detected event has occurred, and a relevant component which is a component directly or indirectly coupled to the event component. When the performance data of the selected component is not stored in a memory, the management system reads the performance data of the selected component from a performance store to the memory.