Cross-Layer Power Management via SLA Violation Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Data centers face challenges in power conservation and service level agreement (SLA) compliance due to continuous CPU load adjustments, leading to increased power consumption and management costs, as existing solutions primarily focus on CPU utilization rather than service level violations.
Innovation Solution
Implementing a cross-layer power management system using management brokers (M-brokers) and management channels (M-channels) to coordinate power control across hardware, virtualization, and application layers, responding to service level violations by adjusting dynamic frequency and voltage scaling, VM migration, and other actuators within a defined power budget.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If CPU utilization monitoring is used to prevent SLA violations, then service level compliance is improved, but power consumption increases due to continuous CPU load adjustments and application migrations
Solution Approach 1:
The patent changes the monitoring parameter from CPU utilization to service level violation detection. Instead of continuously monitoring CPU metrics and adjusting loads, the system detects actual SLA violations and triggers targeted responses, reducing unnecessary power-consuming operations while maintaining compliance.
Solution Approach 2:
The patent extracts the core problem indicator from CPU utilization metrics to actual SLA violation events. By removing the intermediate step of CPU monitoring and directly detecting service level breaches, the system eliminates unnecessary reactive adjustments that consume power without providing benefit.
2Reliability
If applications are continuously moved between CPUs to prevent SLA violations, then service level compliance is maintained, but system stability deteriorates
Solution Approach 1:
The patent implements feedback by detecting SLA violations and triggering targeted application migrations only when necessary. This closed-loop approach prevents continuous oscillating migrations by responding to actual performance degradation events rather than preemptively reacting to CPU utilization thresholds, thereby stabilizing the system.
Solution Approach 2:
The system takes preliminary action by establishing SLA monitoring and violation detection mechanisms in advance. When violations occur, pre-defined migration policies are executed, avoiding the need for continuous reactive adjustments and reducing system instability caused by frequent load balancing operations.
3Loss of energy
If proactive response to service level violations is implemented, then power conservation is improved, but system complexity increases due to multi-layer coordination
Solution Approach 1:
The patent introduces an intermediary component that coordinates between the SLA monitoring layer and the power management layer. This mediator translates service level violations into targeted power-saving actions across hardware, virtualization, and application layers, managing the complexity internally while presenting a simplified proactive response interface.
Solution Approach 2:
The patent segments the power management system into distinct layers (hardware, virtualization, application) with specialized components for SLA detection and coordinated response. This modular segmentation allows each layer to handle specific tasks independently, managing overall system complexity while enabling comprehensive proactive power conservation.
Data Source
AI summary
A method for cross-layer power management in a multi-layer system includes determining whether there is a service level violation for an application running on a hardware platform. Power consumption of the hardware platform is controlled in response to the service level violation.