Datacenter Workload Placement for Predictive Server Idling

Solution Overview

Problem

Datacenters face inefficiencies in energy consumption due to varying workload demands, leading to increased carbon footprint and operational costs, particularly during non-peak hours when server utilization is low.

Innovation Solution

A predictive feedback control loop-based server management framework that dynamically adjusts workload distribution among servers using a compute workload demand forecast, bin-packing active servers for higher utilization and idling underutilized servers into idle states, optimizing energy efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If servers are kept running to meet potential workload demand, then service availability is maintained, but energy consumption increases during non-peak hours

Engineering Contradiction:

Improveservice availabilityVSAvoidenergy consumption

Core Design Contradiction:

ReliabilityVSUse of energy by moving object

Solution Approach 1:

The patent implements dynamic server state management by transitioning servers between active and idle states based on real-time workload demand. The system monitors workload queues and dynamically adjusts server operational states, keeping servers active during peak demand and transitioning them to idle states during non-peak hours, thereby optimizing the balance between service availability and energy consumption

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system performs preliminary actions by pre-warming idle servers and pre-allocating resources before peak workload periods occur. The workload forecast component predicts future demand and triggers proactive server activation, allowing servers to be prepared and ready before actual workload arrives, thus maintaining service availability while enabling energy savings during low-demand periods

Inventive Principle:

Principle #10Preliminary action

2Productivity

If more servers are deployed to handle peak workload, then processing capacity increases, but energy waste occurs during non-peak hours

Engineering Contradiction:

Improveprocessing capacityVSAvoidenergy waste

Core Design Contradiction:

ProductivityVSLoss of energy

Solution Approach 1:

The system applies partial action by activating only the necessary number of servers required to handle the current workload demand rather than keeping all servers continuously active. The intelligent energy algorithm calculates the optimal number of active servers based on real-time workload metrics, enabling the datacenter to maintain adequate processing capacity while minimizing the number of running servers during low-demand periods, thus reducing energy waste

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The patent implements dynamic server state management by transitioning servers between active and idle states based on real-time workload demand. The system monitors workload queues and dynamically adjusts server operational states, keeping servers active during peak demand and transitioning them to idle states during non-peak hours, thereby optimizing the balance between service availability and energy consumption

Inventive Principle:

Principle #15Dynamics

3Speed

If servers remain in active state for quick workload response, then response time is reduced, but energy efficiency decreases

Engineering Contradiction:

Improveworkload response timeVSAvoidenergy efficiency

Core Design Contradiction:

SpeedVSUse of energy by moving object

Solution Approach 1:

The system performs preliminary actions by pre-warming idle servers and pre-allocating resources before peak workload periods occur. The workload forecast component predicts future demand and triggers proactive server activation, allowing servers to be prepared and ready before actual workload arrives, thus maintaining service availability while enabling energy savings during low-demand periods

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent implements dynamic server state management by transitioning servers between active and idle states based on real-time workload demand. The system monitors workload queues and dynamically adjusts server operational states, keeping servers active during peak demand and transitioning them to idle states during non-peak hours, thereby optimizing the balance between service availability and energy consumption

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS12495089B2Efficient datacenter energy management based on compute capacity and fleet management

Publication Date: 2025.12.09 GOOGLE LLC

AI summary

The technology is generally directed to a management framework that uses a predictive feedback control loop to reduce energy consumption of a datacenter. The framework determines how to place a series of jobs or workloads on the available pool of machines in a datacenter. For example, the framework may consider the current workload profile and the workload demand forecast of the datacenter to determine an updated workload profile. The updated workload profile may include a redistribution of the workloads or jobs onto a first subset of the machines such that a second subset of the machines can enter an idle state.