Workload Scheduling for Data Center Cost Reduction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Data centers face high operational costs and energy strain due to power consumption and heat management, especially with the increasing demand for geographically dispersed server clusters and reliance on redundant systems and air conditioning.

Innovation Solution

A workload scheduling system that identifies the cheapest available data center to perform tasks by analyzing renewable energy forecasts, ambient temperatures, and utility costs across geographically dispersed data centers, optimizing load distribution to minimize computational costs.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If server redundancy and UPS systems are implemented to ensure continuous operation, then reliability is improved, but device complexity and cost increase

Engineering Contradiction:

Improvecontinuous operationVSAvoidsystem duplication

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent creates virtual copies of computing workloads that can be migrated between physical data centers. Instead of duplicating entire physical systems (servers, UPS, cooling), the invention copies computational tasks and moves them between locations, achieving redundancy through virtualization rather than physical duplication.

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent introduces a workload scheduling system as an intermediary layer between computing tasks and physical data centers. This mediator manages task distribution, migration, and placement across multiple data centers, coordinating resources to ensure continuous operation without requiring each site to have complete redundant infrastructure.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Temperature

If large air conditioning systems are deployed to manage heat from servers and UPS systems, then temperature control is improved, but energy consumption increases

Engineering Contradiction:

Improveenvironmental controlVSAvoidpower consumption

Core Design Contradiction:

TemperatureVSUse of energy by moving object

Solution Approach 1:

The patent dynamically distributes workloads across multiple data centers based on real-time conditions including temperature and energy availability. Instead of statically over-provisioning cooling systems, the system adapts task placement to current environmental conditions, moving computationally intensive tasks to locations with better thermal conditions or renewable energy availability.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent changes the operational parameters of data centers by utilizing ambient temperature variations across different geographic locations and times. The system leverages natural temperature differences to reduce cooling requirements, placing workloads in environments where passive cooling or reduced active cooling is sufficient.

Inventive Principle:

Principle #35Parameter changes

3Adaptability or versatility

If data centers are geographically dispersed to provide cloud and remote computing services, then adaptability is improved, but coordination complexity and energy management difficulty increase

Engineering Contradiction:

Improveservice distributionVSAvoidload management

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent creates a universal workload scheduling system that can manage diverse computing tasks across heterogeneous data center environments. The scheduling framework provides a common interface and management layer that works across different geographic locations, hardware configurations, and energy sources, simplifying the coordination of dispersed resources through a unified system.

Inventive Principle:

Principle #6Universality (Multi-functionality)

4Productivity

If more data centers are built to meet growing demand, then productivity is improved, but energy resources are strained during heavy power usage periods

Engineering Contradiction:

Improvecomputational capacityVSAvoidenergy resource strain

Core Design Contradiction:

ProductivityVSUse of energy by moving object

Solution Approach 1:

The patent performs preliminary assessment of renewable energy availability and environmental conditions before placing workloads in data centers. By forecasting energy availability and pre-positioning tasks in locations with upcoming renewable energy generation or favorable conditions, the system avoids strain on energy resources during peak demand periods.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent maintains continuous utilization of available computing resources across the data center network by dynamically balancing workloads. Instead of allowing idle capacity in some locations while overloading others, the system continuously optimizes task distribution to keep resources productive while matching energy consumption with renewable generation patterns.

Inventive Principle:

Principle #20Continuity of useful action

Data Source

PatentUS9654414B2Scheduling cost efficient datacenter load distribution

Publication Date: 2017.05.16 MAPLEBEAR INC

AI summary

A method for scheduling cost efficient data center load distribution is described. The method includes receiving a task to be performed by computing resources within a set of data centers. The method further includes determining, all available data centers to perform the task. The method further includes determining lowest computing cost task schedule from available data centers. The method further includes scheduling the task to be completed at an available data center with the lowest cost computing.