Virtual Machine Scaling Threshold Adaptation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing virtual machine scaling technologies face inefficiencies due to static threshold values for key performance indicators (KPIs), leading to unnecessary auto-scaling operations and oscillations, especially in network function virtualization environments with varying hardware and infrastructure conditions.
Innovation Solution
Adapting threshold values for virtual machine scaling based on both system and external key performance indicators, dynamically adjusting upper and lower thresholds to prevent opposite scaling actions and ensure efficient resource allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If static threshold values are used for scaling decisions, then the system is simple to configure and operate, but unnecessary auto-scaling operations and oscillations occur
Solution Approach 1:
The patent implements dynamic threshold adjustment by introducing a learning mechanism that adapts threshold values based on historical KPI data and scaling outcomes. The system transitions from static, manually-configured thresholds to dynamic thresholds that automatically adjust based on observed system behavior and external conditions, thereby reducing unnecessary scaling operations while maintaining operational simplicity.
Solution Approach 2:
The patent employs feedback mechanisms by monitoring the outcomes of scaling operations and using this information to adjust future threshold values. The system evaluates whether scaling actions were necessary by analyzing post-scaling KPI trends and external conditions, then feeds this information back into the threshold adjustment logic to prevent future unnecessary scaling operations.
2Stability of the object's composition
If conservative threshold values are used to avoid unnecessary scaling, then scaling oscillations are reduced, but resource allocation efficiency decreases
Solution Approach 1:
The patent changes the parameter of threshold values from fixed conservative values to adaptive values that are modified based on learned patterns from historical data. The system adjusts threshold parameters dynamically, increasing them when conditions indicate scaling is likely to be beneficial, and decreasing them when scaling would be unnecessary, thereby optimizing both stability and resource efficiency.
Solution Approach 2:
The patent introduces a flexible threshold adjustment mechanism that acts as an adaptive layer between the rigid static thresholds and the dynamic scaling decisions. This flexible mechanism allows thresholds to bend and adapt to changing conditions rather than remaining rigidly fixed, enabling the system to maintain stability while improving resource allocation efficiency.
3Ease of manufacture
If threshold values are manually configured and remain static, then the configuration is straightforward, but the system cannot adapt to changed network or infrastructure conditions
Solution Approach 1:
The patent implements self-service by enabling the system to automatically adjust its own threshold values without requiring manual reconfiguration. The learning mechanism autonomously monitors system behavior, analyzes external conditions, and modifies threshold parameters as needed, allowing the system to adapt to changing network and infrastructure conditions while maintaining straightforward initial configuration.
Solution Approach 2:
The patent applies preliminary action by pre-configuring the system with initial threshold values and a learning mechanism that prepares the system for future adaptations. The system proactively learns from historical data and external conditions in advance, building up knowledge that enables it to adapt quickly and effectively when conditions change, rather than reacting passively to changes.
Data Source
AI summary
Methods, apparatus, and articles of manufacture are disclosed to trigger a scaling action for scaling an application having a set of one or more virtual machines (VMs). Virtualized Network Functions (VNF) are scaled by adding or removing resources to/from existing VMs. In an example method for triggering a scaling action for scaling an application having a set of one or more VMs, a threshold value is adapted based on an evaluation of a monitored system key performance indicator and a monitored external key performance indicator. The threshold value is used for triggering the scaling action. The scaling action is validated based on the monitored external key performance indicator.