Cloud Power Optimization via Workload Placement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current resource management platforms in cloud computing environments prioritize performance over power usage, leading to inefficient datacenter power management and increased environmental impact, as they fail to effectively balance workload distribution and consider power consumption across physical hosts and virtual machines.
Innovation Solution
The implementation of a power optimization system that determines background and active power usage of physical hosts and virtual machines, balances workload distribution based on power profiles, and utilizes thermal hotspot proximity and utility rate structures to schedule workloads on hosts with high compute per watt efficiency, integrating with existing performance-based algorithms to achieve power efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If resource management platforms prioritize workload performance and consolidation, then computing efficiency and resource utilization improve, but datacenter power consumption increases
Solution Approach 1:
The patent applies local quality by creating power profiles for different physical hosts that capture their specific power consumption characteristics. Instead of treating all hosts uniformly, the system assigns unique power profiles that reflect individual host properties, enabling differentiated workload placement decisions that optimize power efficiency at the local host level while maintaining overall system productivity.
Solution Approach 2:
The patent changes the parameter considerations in workload management from purely performance-based metrics to include power consumption parameters. By introducing power profiles with multiple parameters (background power, active power, compute per watt) and using these parameters in placement decisions, the system transforms the optimization objective to balance both performance and power efficiency.
2Quantity of substance
If workloads are concentrated on fewer hosts to improve resource utilization, then infrastructure costs decrease, but thermal hotspots and power inefficiency increase
Solution Approach 1:
The patent addresses thermal hotspot issues by incorporating location-aware power profiles that consider the physical distribution of workloads across datacenter infrastructure. The system evaluates power consumption and thermal characteristics at the individual host level, enabling placement decisions that distribute heat generation more evenly across the datacenter floor plan, thereby reducing localized thermal hotspots while maintaining resource consolidation benefits.
Solution Approach 2:
The patent introduces a new dimension to workload management by considering the physical spatial dimension in addition to logical resource allocation. Power profiles include location-based characteristics that account for thermal and power distribution across the datacenter physical infrastructure. This dimensional addition enables the system to optimize not just for resource consolidation but also for thermal management and power efficiency in the physical space.
3Ease of manufacture
If enterprises use cloud services to reduce infrastructure costs, then capital expenditure decreases, but indirect power cost awareness and optimization opportunities are lost
Solution Approach 1:
The patent implements feedback mechanisms that provide visibility into power consumption metrics and costs associated with cloud workloads. By monitoring and reporting power usage, cost attributes are fed back to enterprises, enabling them to make informed decisions about workload placement and optimization even in cloud environments. This feedback loop restores power cost awareness that would otherwise be lost in indirect cloud pricing models.
Solution Approach 2:
The patent introduces an intermediary layer between cloud infrastructure and enterprise decision-making that captures and translates power consumption data into actionable cost information. Power profiles and cost attributes act as intermediaries that bridge the gap between physical power consumption and enterprise cost accounting, enabling optimization opportunities to be identified and acted upon despite the indirect nature of cloud pricing.
Data Source
AI summary
A power optimization system may include a cloud management server coupled to a plurality of clusters via a network, a resource management module residing in the cloud management server, and a cloud power optimizer module residing in the resource management module. Each cluster may include a plurality of physical hosts with at least one virtual machine (VM) running on each physical host. During operation, the cloud power optimizer module may determine background and active power usages of each physical host in the plurality of clusters. Further, the cloud power optimizer module may determine power usage of each VM based on the determined background and active power usages of each physical host. Furthermore, the cloud power optimizer module may continuously balance a distribution of workload on the plurality of physical hosts based on the determined power usage of each VM.