Workload Allocation via Node Activity Threshold Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In multi-task processing systems, efficiently allocating workloads among computing devices to minimize computing time and optimize resource usage remains a challenge, as existing methods often require transferring information from all nodes, leading to increased resource consumption and reduced efficiency.
Innovation Solution
A method that involves acquiring a state graph of devices running a workload, updating it based on active node values compared to a threshold, and allocating subsequent workloads only to devices with active nodes above the threshold, utilizing a policy model based on reinforcement learning to optimize workload distribution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If information from all nodes is transferred for workload allocation, then allocation accuracy is improved, but resource consumption increases
Solution Approach 1:
The patent extracts only the necessary information (active values of nodes above threshold) from the complete node set, rather than transferring all node information. This selective extraction reduces communication overhead and resource consumption while maintaining sufficient accuracy for workload allocation decisions.
Solution Approach 2:
The patent applies partial action by processing only a subset of nodes (those with active values exceeding the threshold) rather than all nodes. This partial processing approach reduces computational load and resource consumption while still providing adequate information for effective workload allocation.
2Reliability
If all node information is processed, then allocation completeness is improved, but computing efficiency decreases
Solution Approach 1:
The patent extracts only the relevant subset of nodes (those with active values above threshold) for processing, eliminating unnecessary computational work on low-activity nodes while retaining sufficient information for reliable workload allocation decisions.
Solution Approach 2:
The patent processes only a partial set of nodes rather than all nodes, improving computing efficiency by reducing the processing scope. The threshold-based filtering ensures that the partial processing still captures the most significant nodes for accurate allocation.
3Loss of energy
If threshold filtering is applied to nodes, then resource savings are improved, but information loss increases
Solution Approach 1:
The patent changes the parameter threshold for node selection, using an adaptive threshold mechanism to determine which nodes are significant enough to process. By dynamically adjusting the threshold based on system state, the patent maximizes resource savings while minimizing information loss about important nodes.
Solution Approach 2:
The patent employs feedback mechanisms where the threshold and selection criteria are continuously refined based on observed system performance and allocation outcomes. This feedback loop ensures that the threshold filtering preserves necessary information while achieving resource savings.
Data Source
AI summary
Embodiments of the present disclosure relate to a method, an electronic device, and a computer program product for allocating a workload. The method includes acquiring a first state graph of a plurality of devices that run a first workload at a first time point. The method further includes updating the first state graph based on a comparison between an active value of at least one node in the first state graph and a predetermined threshold. The method further includes determining a first load state of the plurality of devices at the first time point based on an updated first state graph. The method further includes allocating a second workload to the plurality of devices at a second time point based on the first load state, where active values of nodes in the updated first state graph are greater than the predetermined threshold.