Unified Power Model for Cloud Workload Energy Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for calculating energy consumption in Cloud computing systems are inadequate due to their reliance on static energy consumption models that fail to account for volatile workload demands and changing system configurations, leading to inaccurate estimates.
Innovation Solution
A unified power consumption model is developed and stored in a database, allowing for accurate estimation of workload energy consumption by identifying the appropriate model based on current system characteristics and adjusting parameters through monitoring.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If static energy consumption models are used for each hardware and software component, then the modeling process is simplified and decoupled from infrastructure changes, but the accuracy of energy consumption estimates deteriorates due to inability to account for volatile workload demands and changing system configurations
Solution Approach 1:
The patent combines multiple static energy consumption models for individual hardware and software components into a single unified power consumption model. This unified model integrates the energy characteristics of CPU, memory, storage, and software layers, allowing accurate estimation of total system energy consumption while maintaining model decoupling from infrastructure changes.
Solution Approach 2:
The patent introduces dynamic parameters and monitoring mechanisms that allow the unified power consumption model to adapt to volatile workload demands and changing system configurations. The model incorporates real-time workload characteristics and system state information to dynamically adjust energy consumption estimates, transforming static component models into a dynamic system-level model.
2Ease of manufacture
If individual static energy consumption models are maintained for each component, then the system is easier to implement initially, but the adaptability to changing workloads and system configurations deteriorates
Solution Approach 1:
The patent creates a universal unified power consumption model that serves multiple functions: it can estimate energy consumption for different workload types (transient, overtime, benchmark), adapt to various system configurations, and provide both short-term and long-term energy predictions. This single model replaces multiple component-specific models while maintaining ease of implementation through a standardized approach.
Solution Approach 2:
The patent implements feedback mechanisms that monitor actual system performance and workload characteristics, then use this information to adjust the unified power consumption model parameters. This feedback loop enables the model to adapt to changing workloads and system configurations while maintaining accurate energy consumption estimates throughout the system's operational lifecycle.
Data Source
AI summary
Computer-implemented methods for estimating the energy consumption of a workload in a Cloud computing system are provided. Aspects include receiving a request for an estimated energy consumption of the workload and obtaining characteristics of the Cloud computing system executing the workload. Aspects also include identifying and employing a unified power consumption model from a power model database based on the characteristics and calculating the estimated energy consumption of the workload based on the unified power consumption model.