Auto-scaling Thresholds in Elastic Computing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Improperly defined auto-scaling parameters in elastic computing environments lead to unnecessary and ineffective auto-scaling operations, increasing operational costs and degrading performance.
Innovation Solution
A method for configuring auto-scaling parameters in elastic computing environments, involving the calculation of optimal values using specified inequalities to reduce the number of auto-scaling operations and minimize performance degradation, with a system that assists users in defining feasible configurations and alerts when auto-scaling operations are not attainable.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If auto-scaling parameters are set to be sensitive to performance metric changes, then the computing environment can respond quickly to load changes, but unnecessary auto-scaling operations increase operational costs
Solution Approach 1:
The patent implements dynamic threshold adjustment where the sensitivity of auto-scaling triggers is not fixed but adapts based on current system state. The performance metric thresholds are calculated dynamically considering current resource utilization, workload characteristics, and historical patterns, allowing the system to be more responsive during critical periods while being more conservative during stable periods, thus balancing response speed with operational cost reduction
Solution Approach 2:
The system changes the parameters used for auto-scaling decisions by calculating optimal threshold values based on multiple factors including current resource utilization, workload type, and performance targets. Instead of using fixed thresholds, the system adjusts the parameter values dynamically to match actual system conditions, reducing unnecessary scaling operations while maintaining adequate response to genuine load changes
2Manufacturing precision
If auto-scaling operations are performed frequently to maintain desired resource levels, then resource allocation accuracy improves, but system performance degrades due to oscillations
Solution Approach 1:
The patent applies cushioning by establishing buffer zones around target resource levels. Instead of triggering immediate scaling operations at every threshold crossing, the system creates hysteresis bands that prevent oscillations. When performance metrics approach thresholds, the system anticipates future states and adjusts thresholds or scaling increments in advance, cushioning against the harmful effects of frequent oscillations while maintaining adequate resource allocation accuracy
Solution Approach 2:
The system implements partial action by adjusting auto-scaling parameters incrementally rather than making full-scale adjustments at each threshold crossing. The patent calculates optimal scaling increments that are sufficient to move toward target resource levels without overshooting, thereby reducing the frequency of oscillations while still achieving the desired resource allocation accuracy over time
3Device complexity
If fixed auto-scaling thresholds are used, then system complexity is reduced, but the system cannot adapt to varying workload patterns
Solution Approach 1:
The patent implements self-service by enabling the auto-scaling system to automatically calculate and adjust its own thresholds based on observed workload patterns and performance metrics. The system monitors historical data, identifies recurring patterns, and autonomously adapts threshold values without requiring manual reconfiguration or complex external control systems, thereby maintaining relatively simple system architecture while achieving high adaptability to varying workload patterns
Data Source
AI summary
Embodiments of the present invention provide systems, methods, and computer program products for configuring auto-scaling parameters of a computing environment, as well as alerting a user when auto-scaling operations are not attainable given current operating configurations.