Dynamic Virtual Machine Container Resizing for Server Utilization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current server allocation tools rely on static resource partitioning, leading to poor resource utilization and inefficiency due to static entitlements that do not adapt to varying demands, resulting in suboptimal performance and potential request drops.
Innovation Solution
Implement a system for dynamically resizing virtual machine containers based on transaction mix information, desired performance levels, and actual resource consumption, using a transaction-mix-based queuing model to adjust resource entitlements dynamically, enabling automatic modification of resource allocation to match optimal utilization targets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If static resource partitioning is used to allocate servers to applications, then resource allocation is simple and stable, but resource utilization is poor and below 20%
Solution Approach 1:
The patent implements dynamic resource entitlement adjustment by continuously monitoring application performance metrics and transaction mix information. The system automatically modifies resource allocation entitlements in real-time based on actual demand, transitioning from static partitioning to dynamic adaptation. This resolves the contradiction by making the resource allocation system flexible and responsive to changing conditions while maintaining automated control.
Solution Approach 2:
The system employs feedback mechanisms by monitoring application performance metrics, transaction mix information, and resource consumption patterns. Based on this feedback, the system automatically adjusts resource entitlements to optimize utilization. The feedback loop enables the system to adapt to demand variations while maintaining stable and efficient resource allocation, resolving the contradiction between adaptability and stability.
2Reliability
If admission control is implemented to improve performance of admitted requests, then performance of admitted requests improves, but requests are either not admitted or dropped
Solution Approach 1:
The patent replaces static admission control with dynamic resource entitlement adjustment. Instead of rigidly rejecting or admitting requests based on fixed thresholds, the system continuously adapts resource allocation to match actual demand patterns. This dynamic approach maintains high performance for admitted requests while maximizing request throughput by avoiding unnecessary rejections, resolving the contradiction between reliability and productivity.
Solution Approach 2:
The system changes resource entitlement parameters dynamically based on monitored performance metrics and transaction mix information. By adjusting these parameters in real-time, the system optimizes both the performance of admitted requests and the overall request throughput, eliminating the need for conservative admission control that limits productivity.
3Productivity
If static entitlement of system resources is determined based on anticipated peak load, then resource allocation is simple, but it does not take full advantage of demands that vary over time
Solution Approach 1:
The patent implements a self-service resource management system that automatically monitors application performance, analyzes transaction mix information, and adjusts resource entitlements without manual intervention. The system serves itself by making autonomous decisions based on monitored data, improving resource utilization while avoiding the complexity of manual resource management. This resolves the contradiction by enabling automatic optimization without requiring complex human-operated systems.
Solution Approach 2:
The system uses feedback from performance metrics and transaction mix information to automatically adjust resource entitlements. This feedback-driven approach enables the system to adapt to varying demands over time without manual intervention, improving resource utilization while maintaining automated control that does not significantly increase system complexity.
Data Source
AI summary
Embodiments of the present invention pertain to dynamically resizing a virtual machine container. According to one embodiment, an optimal utilization is determined based on a desired performance for a multi-tiered application and transaction mix information that describes a mix of transactions that result from executing the multi-tiered application. Automatic modification of an amount of a physical resource a virtual machine container is entitled to for executing a particular application component associated with the multi-tiered application is enabled where the amount is automatically modified based on the optimal utilization.