Edge Agent Scheduling for Performance Metric Calculation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Computing devices face challenges in efficiently processing performance metrics, as collecting and processing these metrics from a large quantity of devices can consume significant resources, leading to potential performance degradation and increased burden on computing resources.
Innovation Solution
Implementing an agent on edge devices to perform performance metric calculations during determined downtime, utilizing low-activity periods to execute calculations without affecting user experience, by activating and deactivating the agent based on activity models and thresholds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If performance metrics are collected and processed from a large quantity of computing devices by central computing resources, then comprehensive performance data can be obtained, but resource consumption increases significantly and may cause performance degradation
Solution Approach 1:
The patent divides the centralized performance metric processing into distributed edge computing tasks. Each computing device runs a local agent that performs metric calculations independently, segmenting the workload from central computing resources. This reduces resource consumption at the center while maintaining comprehensive data collection across all devices.
Solution Approach 2:
The patent implements preliminary action by having computing devices collect and pre-process performance metrics locally before transmitting results to central resources. The agent on each device performs calculations in advance during downtime, so that when data is sent to the center, the heavy computational burden has already been reduced, minimizing resource consumption during peak operations.
2Productivity
If performance metric calculations are performed continuously on computing devices, then up-to-date performance data is available, but user experience deteriorates due to resource utilization during active periods
Solution Approach 1:
The patent implements periodic action by scheduling performance metric calculations to occur only during determined downtime intervals when the computing device is not actively being used. The agent monitors device activity and triggers calculations during idle periods, maintaining up-to-date performance data without interfering with user operations during active periods.
Solution Approach 2:
The patent applies dynamics by making the agent's execution state adaptive rather than static. The agent dynamically transitions between active and inactive states based on real-time device usage conditions. When the device is in use, the agent remains inactive to preserve user experience; when downtime is detected, the agent activates to perform calculations, thus balancing productivity with ease of operation.
3Loss of energy
If an agent is deployed on computing devices to perform calculations, then central computing resources are relieved of processing burden, but device complexity increases
Solution Approach 1:
The patent implements self-service by enabling computing devices to autonomously perform their own performance metric calculations through locally deployed agents. Each device's agent independently monitors its own performance, collects metrics, and transmits results without requiring central resource intervention for calculation. This self-service capability reduces central resource utilization while the modular agent design keeps individual device complexity manageable.
Data Source
AI summary
In some examples, a computing device can include a processor resource and a non-transitory memory resource storing machine-readable instructions stored thereon that, when executed, cause the processor resource to: generate a model of activity for the computing device, determine a time period for performing a calculation based on the model, wherein the calculation utilizes performance metrics associated with the computing device, activate an agent at a start time of the time period to perform the calculation, send, by the agent, a result of the calculation to a remote computing device, and deactivate the agent in response to sending the result.