OPTIMIZING POWER CONSUMPTION BY DYNAMIC WORKLOAD ADJUSTMENT

Patent №

US 9,501,115

Granted

2016-11-22

Filed 2013

Owner

INTERNATIONAL BUSINESS MACHINES CORPORATION

AI components

3

kr · evo · hardware

Assignment

Recorded

Dataset

AIPD

2023_r1 edition

Application

13905432

A method and system for optimizing power consumption of a data center by dynamic workload adjustment. At least one candidate workload solution for the data center is generated. Each candidate workload solution represents a respective application map that specifies a respective workload distribution among application programs of the data center. Workload of the data center is dynamically adjusted from a current workload distribution to an optimal workload solution. The optimal workload solution is a candidate workload solution of the at least one candidate workload solution having a lowest sum of a respective power cost and a respective migration cost. Dynamically adjusting the workload of the data center includes: estimating a respective overall cost of each candidate workload solution, selecting the optimal workload solution that has a lowest overall cost as determined from the estimating, and transferring the optimal workload solution to devices of a computer system for deployment.

Knowledge representationEvolutionary computationAI hardwareG06F 1/329G06F 1/26G06F 1/3203G06F 9/5088G06F 9/5094Y02D 10/00

AI classification

Knowledge representation0.91
AI hardware0.88
Evolutionary computation0.74
Planning0.18
Natural language0.00
Machine learning0.00
Speech0.00
Vision0.00

Ownership

INTERNATIONAL BUSINESS MACHINES CORPORATION

assignment · 305120850

Assignors

MANDAGERE, NAGAPRAMOD, MURTHY, SESHASHAYEE S., UTTAMCHANDANI, SANDEEP M., ZHOU, PIN

On an employer assignment, the assignors are typically the inventors.

From the same owner

© 2026 NYSGPT2525 LLC