Computing Cluster Performance-Data Acquisition With Adaptive Frequencies
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High Performance Computing (HPC) systems face challenges in processing overheads of performance indicator data during monitoring and analysis, leading to inaccuracies in performance analysis results due to low acquisition frequencies, which can cause loss or distortion of data.
Innovation Solution
A computing cluster system with a management and control node and multiple computing nodes, where each node deploys acquisitors to acquire performance indicators at varying frequencies. A primary node adjusts these frequencies based on change information, synchronizing adjustments across nodes to ensure accurate data acquisition while reducing processing overheads.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a relatively low acquisition frequency is used to reduce the amount of performance indicator data, then processing overheads are reduced, but accuracy of performance analysis results deteriorates due to loss or distortion of data
Solution Approach 1:
The patent implements dynamic acquisition frequency adjustment where the system transitions from a static low acquisition frequency to a dynamic frequency that adapts based on performance indicator changes. The management and control node monitors performance indicators and adjusts the acquisition frequency in real-time, increasing frequency when changes are detected and reducing it when stable, thereby resolving the contradiction between reducing processing overhead and maintaining analysis accuracy
Solution Approach 2:
The system changes the parameter of acquisition frequency from a fixed low value to a variable value that adjusts based on performance indicator variations. By monitoring changes in performance indicators and dynamically modifying the acquisition frequency parameter, the system achieves both reduced processing overhead during stable periods and maintained accuracy during changing conditions
2Measurement precision
If a high acquisition frequency is used to ensure accuracy of performance analysis results, then measurement precision is improved, but processing overheads increase due to large amount of data
Solution Approach 1:
The system dynamically adjusts the acquisition frequency parameter based on performance indicator changes, using high frequency only when necessary to maintain accuracy while using low frequency during stable periods to reduce processing overhead
Solution Approach 2:
The acquisition frequency transitions from a static high value to a dynamic value that adapts to system conditions, being high only when performance indicators change and low when stable, thereby maintaining accuracy while reducing overall processing overhead
3Device complexity
If the same acquisition frequency is used across all computing nodes, then device complexity is reduced, but adaptability deteriorates as nodes cannot respond to their specific performance changes
Solution Approach 1:
The management and control node implements a feedback mechanism where it collects performance indicator data from computing nodes, analyzes changes, and sends adjustment instructions back to the nodes. This centralized feedback loop enables adaptive frequency adjustment across nodes without requiring complex local decision-making logic, maintaining low device complexity while achieving high adaptability
Solution Approach 2:
The management and control node acts as an intermediary between computing nodes and acquisition frequencies. Instead of nodes independently managing their own frequencies (complex), the intermediary monitors all nodes and coordinates frequency adjustments, simplifying individual node complexity while enabling system-wide adaptability
Data Source
AI summary
Embodiments of the present application provide a computing cluster, and a data acquisition method and apparatus for same, and a storage medium. In the embodiments of the present application, under a computing cluster scenario, acquisition frequencies of performance indicator data is adaptively changed based on change information of the performance indicator data, so that acquisition accuracy can be ensured to guarantee the accuracy of performance analysis based on the performance indicator data and decision-making based on an analysis result, and acquisition and processing overheads of the performance indicator data can also be reduced. In a process of adaptively changing the acquisition frequencies, regarding at least two computing nodes executing a same working task, a primary node among the at least two computing nodes is responsible for adaptive change processing of the acquisition frequencies and synchronizing it to an else computing node when a change is required, and the else computing node is not responsible for the adaptive change processing of the acquisition frequencies, so that the processing burden of the else computing node can be reduced.