System and method for recommending configuration adjustments based on configurations at similarly capable information handling systems with lower carbon footprints
A usage profile based CO2 optimization system of an information handling system may comprise a processor to determine, using an ensemble machine-learning algorithm, rankings for user-disruptive static system configurations for a plurality of client information handling systems according to a level at which each user-disruptive static system configuration impacts carbon footprints of the client information handling systems, determine the first and second client information handling systems have matching values for a highest ranked user-disruptive static system configuration and mismatching values for an adjustable dynamic system configuration, determine the first client information handling system has a smaller carbon footprint than the second client information handling system, a network interface device to transmit a recommendation to the second client information handling system to adopt the value for the adjustable dynamic system configuration at the first client information handling system to decrease the carbon footprint of the second client information handling system.
1 . A usage profile based CO2 optimization system of an information handling system comprising:
a hardware processor, memory, and a power management unit to supply power to the hardware processor and the memory;
the hardware processor executing code instructions of an ensemble machine-learning algorithm of the usage profile based CO2 optimization system to determine rankings for each of a plurality of user-disruptive static system configurations determined to be disruptive for users to adjust for a plurality of client information handling systems according to a level at which each of the user-disruptive static system configurations impacts carbon footprints during operation in each user-disruptive static system configuration, wherein the rankings are used to determine similarity among functional capabilities of the user-disruptive static system configurations for the plurality of client information handling systems;
the hardware processor executing code instructions to receive first telemetry event logs in a first Javascript object notation (JSON) file of configuration and power consumption, processor operation, and network interface system operations to classify a first client information handling system within a CO2 optimization classification, based on a first value for a highest ranked user-disruptive static system configuration of the first client information handling system;
the hardware processor executing code instructions to receive second telemetry event logs in a second JSON file of configuration and power consumption, processor operation, and network interface system operations to classify a second client information handling systems within the CO2 optimization classification, based on a second value for the highest ranked user-disruptive static system configuration of the second client information handling system;
the hardware processor executing code instructions to determine from output of the ensemble machine-learning algorithm of the usage profile based CO2 optimization system that the first client information handling system has a smaller carbon footprint than the second client information handling system and identify a first adjustable dynamic system configuration for the first client information handling system does not match a second adjustable dynamic system configuration for the second client information handling system, where adjustments to the adjustable dynamic system configuration are not determined to be user-disruptive; and
a network interface device to transmit a recommendation to the second client information handling system to adopt the first adjustable dynamic system configuration to decrease the carbon footprint of the second client information handling system.
2 . The information handling system of claim 1 , wherein the user-disruptive static system configuration includes a hardware component type.
3 . The information handling system of claim 1 , wherein the user-disruptive static system configuration includes a measure of hardware component resource consumption by an executing software application.
4 . The information handling system of claim 1 , wherein the adjustable dynamic system configuration includes hardware policy settings.
5 . The information handling system of claim 1 , wherein the adjustable dynamic system configuration includes a performance mode for a hardware component.
6 . The information handling system of claim 1 , wherein the adjustable dynamic system configuration includes a power conservation setting for a hardware component.
7 . The information handling system of claim 1 , wherein the adjustable dynamic system configuration includes a limitation of hardware component resources consumed during execution of background software applications.
8 . A method of executing code instructions of a usage profile based CO2 optimization system at an information handling system for optimizing carbon footprint for a client information handling system based on a usage profile comprising:
determining, via a processor executing code instructions of, using an ensemble machine-learning algorithm for the usage profile based CO2 optimization system, rankings for each of a plurality of user-disruptive static system configurations determined to be disruptive for users to adjust for a plurality of client information handling systems according to a level at which each of the user-disruptive static system configurations impacts carbon footprints during operation in each user-disruptive static system configuration, wherein the rankings are used to determine similarity among functional capabilities of the user-disruptive static system configurations for the plurality of client information handling systems;
receiving first telemetry event logs in a first Javascript object notation (JSON) file of configuration and power consumption, processor operation, and network interface system operations as well as first location data for input into the ensemble machine-learning algorithm for classifying, via the hardware processor executing code instructions, a first client information handling system within a CO2 optimization classification, based on a first value for a highest ranked user-disruptive static system configuration of the first client information handling system;
receiving second telemetry event logs in a second JSON file of configuration and power consumption, processor operation, and network interface system operations as well as second location data for input into the ensemble machine-learning algorithm for classifying, via the hardware processor executing code instructions, a second client information handling systems within the CO2 optimization classification, based on a second value for the highest ranked user-disruptive static system configuration of the second client information handling system;
determining, via the hardware processor executing code instructions, from output of the ensemble machine-learning algorithm that the first client information handling system has a smaller carbon footprint than the second client information handling system;
identifying via the hardware processor executing code instructions, a first adjustable dynamic system configuration for the first client information handling system does not match a second adjustable dynamic system configuration for the second client information handling system, where adjustments to the adjustable dynamic system configuration are not determined to be user-disruptive; and
transmitting, via a network interface device, a recommendation to the second client information handling system to adopt the first adjustable dynamic system configuration to decrease the carbon footprint of the second client information handling system.
9 . The method of claim 8 , wherein the adjustable dynamic system configuration includes background software application usage.
10 . The method of claim 8 , wherein the adjustable dynamic system configuration includes software or firmware update settings.
11 . The method of claim 8 , wherein the adjustable dynamic system configuration includes a version of firmware installed.
12 . The method of claim 8 , wherein the user-disruptive static system configuration includes measured locations of the first and the second client information handling systems.
13 . The method of claim 8 , wherein the user-disruptive static system configuration includes measured health of a hardware component.
14 . The method of claim 8 , wherein the ensemble machine-learning algorithm is a stacking machine-learning algorithm executing a gradient descent method.
15 . A usage profile based CO2 optimization system of an information handling system comprising:
a hardware processor, memory, and a power management unit to supply power to the hardware processor and the memory;
the hardware processor executing code instructions of an ensemble machine-learning algorithm of the usage profile based CO2 optimization system to: determine, using rankings for each of a plurality of user-disruptive static system configurations determined to be disruptive for users to adjust for a plurality of client information handling systems according to a level at which each of the user-disruptive static system configurations impacts carbon footprints during operation in each user-disruptive static system configuration, wherein the rankings are used to determine similarity among functional capabilities of the user-disruptive static system configurations for the plurality of client information handling systems;
the hardware processor executing code instructions to receive first telemetry event logs in a first Javascript object notation (JSON) file of configuration and power consumption, processor operation, and network interface system operations to classify a first client information handling system within a CO2 optimization classification, based on a first value for a highest ranked user-disruptive static system configuration of the first client information handling system;
the hardware processor executing code instructions to receive second telemetry event logs in a second JSON file of configuration and power consumption, processor operation, and network interface system operations to classify a second client information handling systems within the CO2 optimization classification, based on a second value for the highest ranked user-disruptive static system configuration of the second client information handling system;
the hardware processor executing code instructions to determine from output of the ensemble machine-learning algorithm of the usage profile based CO2 optimization system that the first client information handling system has a smaller carbon footprint than the second client information handling system and to identify a first adjustable dynamic system configuration for the first client information handling system does not match a second adjustable dynamic system configuration for the second client information handling system; and
a network interface device to transmit a recommendation to the second client information handling system to adopt the first adjustable dynamic system configuration to decrease the carbon footprint of the second client information handling system.
16 . The information handling system of claim 15 , wherein the plurality of ensemble machine-learning algorithm includes a bootstrap aggregation (BOOST) machine-learning algorithm.
17 . The information handling system of claim 15 , wherein the plurality of ensemble machine-learning algorithms includes a stacking machine-learning algorithm.
18 . The information handling system of claim 15 , wherein the plurality of ensemble machine-learning algorithms includes a boosting machine-learning algorithm.
19 . The information handling system of claim 15 further comprising:
the processor to determine the rankings for each of the user-disruptive static system configurations by combining estimated rankings output by each of the plurality of ensemble machine-learning algorithms into an overall ranking for each of the user-disruptive static system configurations.
20 . The information handling system of claim 15 further comprising:
the processor to determine the rankings for each of the user-disruptive static system configurations by averaging estimated rankings output by each of the plurality of ensemble machine-learning algorithms into an overall ranking for each of the user-disruptive static system configurations.