Analytics Task Throttling via Dynamic Queue Repositioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Complex enterprise networks pose challenges for IT administrators in monitoring and managing computing and storage resources, leading to difficulties in identifying the root cause of performance issues in mission-critical applications.
Innovation Solution
An application-centric infrastructure management system that dynamically assigns device tiers based on application importance, prioritizes analytic tasks, and adjusts polling intervals to ensure efficient resource utilization and effective monitoring.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If continuous data analysis is performed to monitor enterprise network applications, then problem detection capability is improved, but system resource utilization deteriorates
Solution Approach 1:
The patent implements dynamic throttling of analytics processing based on system conditions. The analytics service adjusts its processing rate in real-time, increasing monitoring intensity when problems are detected and reducing intensity when the system is stable, thereby maintaining reliable problem detection while optimizing resource utilization throughout the monitoring period.
Solution Approach 2:
The system employs periodic analytics processing with variable intervals. Instead of continuous processing, the analytics service operates in cycles, adjusting the period between analysis operations based on system state. This periodic approach maintains adequate monitoring coverage while significantly reducing overall resource consumption compared to continuous analysis.
2Adaptability or versatility
If multiple analytic tasks are executed simultaneously to monitor various applications, then monitoring coverage is improved, but analytic service stability deteriorates
Solution Approach 1:
The patent segments the analytics service into multiple independent worker processes or threads, each handling specific analytic tasks. This segmentation allows the system to monitor multiple applications simultaneously while isolating failures to individual segments, preventing cascading failures that would compromise overall service stability.
Solution Approach 2:
The system introduces a task queue and scheduler as intermediary components between task submission and execution. This intermediary layer manages the workload distribution, controls the rate at which analytic tasks are processed, and provides buffering capacity to handle peaks in monitoring demand without overwhelming the analytics service infrastructure.
3Use of energy by moving object
If high polling intervals are used to reduce resource consumption, then resource utilization is improved, but problem detection speed deteriorates
Solution Approach 1:
The patent implements dynamic adjustment of polling intervals based on system conditions and application criticality. For critical applications or when anomalies are detected, the polling interval automatically decreases to provide faster detection. For stable, non-critical systems, the interval increases to conserve resources. This dynamic approach optimizes the trade-off between resource utilization and detection speed in real-time.
Solution Approach 2:
The system applies different polling intervals to different applications and system components based on their specific requirements. Critical mission-sensitive applications receive more frequent polling with shorter intervals, while non-critical applications use longer intervals. This localized differentiation allows the system to maintain fast problem detection where needed while achieving overall resource efficiency across the entire monitoring infrastructure.
Data Source
AI summary
A method comprising: receiving an analytic task which includes a priority indicator associated with the analytic task, determining a position of the analytic task in a task queue, the task queue arranged in an order according to their priority indicators, selecting the analytic task based on the order of the task queue, sending the analytic task to an analytics service to determine if the analytics service has sufficient available resources to perform the analytic task, receiving an indication that the analytics service does not have sufficient available resources to perform the analytic task, repositioning the analytic task within the task queue, selecting the analytic task from the task queue based on the order of the task queue and a new position of the analytic task, sending the analytic task to the analytics service, and retrieving another analytic task to send to the analytics service.