Modeling Carbon Emissions Of Workload Execution Environments Using Machine Learning
Workload placement based on carbon emissions, including: calculating, for each execution environment of a plurality of execution environments, a carbon emission cost associated with a workload; selecting, based on each carbon emission cost for the plurality of execution environments, a target execution environment; and executing the workload on the target execution environment.
1 . A method of workload placement based on carbon emissions, the method comprising:
selecting a target execution environment for execution of a workload based on carbon emission costs for at least one execution environment, of one or more execution environments, associated with projected workload behavior for one or more workloads; and
executing the workload on the target execution environment.
2 . The method of claim 1 further comprising:
providing one or more workload characteristics to a set of machine learning models trained to provide an estimated energy usage for executing the workload; and
selecting the target execution environment based on one or more outputs of the set of machine learning models.
3 . The method of claim 1 , wherein selecting the target execution environment further comprises:
providing emission sampling data for the at least one execution environment to an other set of machine learning models; and
selecting the target execution environment based on one or more outputs of the other set of machine learning models.
4 . The method of claim 1 , wherein selecting the target execution environment further comprises selecting the target execution environment based on one or more thresholds.
5 . The method of claim 1 , wherein carbon emission costs for the at least one execution environment are calculated based on a projected emission intensity associated with the at least one execution environment.
6 . The method of claim 1 , wherein carbon emission costs for the at least one execution environment are calculated based on one or more energy costs of one or more devices associated with the at least one execution environment.
7 . The method of claim 1 , further comprising deactivating, in the target execution environment, one or more devices in the target execution environment.
8 . The method of claim 1 , further comprising generating a report comprising one or more carbon usage metrics associated with the target execution environment.
9 . The method of claim 8 , wherein the one or more carbon usage metrics comprise one or more of: a carbon usage history, a carbon usage savings, a projected carbon usage, or a projected carbon usage savings.
10 . An apparatus for workload placement based on carbon emissions, the apparatus comprising a computer processor, a computer memory operatively coupled to the computer processor, the computer memory having disposed within its computer program instructions that, when executed by the computer processor, cause the apparatus to carry out the steps of:
selecting a target execution environment for execution of a workload based on carbon emission costs for at least one execution environment, of one or more execution environments, associated with projected workload behavior for one or more workloads; and
executing the workload on the target execution environment.
11 . The apparatus of claim 10 further comprising computer program instructions that, when executed by the computer processor, cause the apparatus to carry out the steps of:
providing a recommendation indicating the target execution environment; and
wherein executing the workload on the target execution environment comprises executing the workload on the target execution environment based on a response to the recommendation.
12 . The apparatus of claim 10 , wherein selecting the target execution environment further comprises selecting the target execution environment based on one or more thresholds.
13 . The apparatus of claim 10 , wherein selecting the target execution environment further comprises:
calculating, for each execution environment, a fitness score based on the carbon emission costs; and
selecting the target execution environment based on the fitness score.
14 . The apparatus of claim 10 , wherein the carbon emission costs for the at least one execution environment is calculated based on a projected emission intensity associated with the at least one execution environment.
15 . The apparatus of claim 10 , wherein the carbon emission costs for the at least one execution environment is calculated based on a projected workload behavior.
16 . The apparatus of claim 10 , wherein the carbon emission costs for the at least one execution environment is calculated based on one or more energy costs of one or more devices associated with the at least one execution environment.
17 . The apparatus of claim 16 further comprising computer program instructions that, when executed by the computer processor, cause the apparatus to carry out the steps of deactivating, in the target execution environment, one or more devices in the target execution environment.
18 . The apparatus of claim 10 further comprising computer program instructions that, when executed by the computer processor, cause the apparatus to carry out the steps of generating a report comprising one or more carbon usage metrics associated with the target execution environment.
19 . The apparatus of claim 18 , wherein the one or more carbon usage metrics comprise one or more of: a carbon usage history, a carbon usage savings, a projected carbon usage, or a projected carbon usage savings.
20 . The apparatus of claim 12 , further comprising computer program instructions that, when executed by the computer processor, cause the apparatus to carry out the steps of:
providing one or more workload characteristics to a set of machine learning models trained to provide an estimated energy usage for executing the workload; and
selecting the target execution environment based on one or more outputs of the set of machine learning models.