International Business Machines Corporation

Rajesh K Jeyapaul

Alfredo V Mendoza

Sivapatham Muthaiah

Deploying energy-rated deep neural network models on energy-scored edge devices is provided. An overall energy efficiency rating is assigned to a deep neural network model based on utilizing software optimization and hardware accelerators during training of the deep neural network model. Energy scores are assigned to respective edge devices in an edge inference computing environment based on properties of each respective edge device. Particular edge devices are selected that have a corresponding energy score within a defined edge device energy score range for the overall energy efficiency rating that corresponds to the deep neural network model. The deep neural network model is deployed to the particular edge devices that have a corresponding energy score within the defined edge device energy score range for the overall energy efficiency rating that corresponds to the deep neural network model.

Deployment and management of energy efficient deep neural network models on edge inference computing devices

A computer-implemented method, system, and/or computer program product controls uploading applications to a mobile device. A server receives a power descriptor from a mobile device, where the power descriptor identifies an amount of power available to the mobile device from a battery on the mobile device. The server receives a set of application downloading rules for the mobile device, where the set of application downloading rules comprises a medium rule, and where the medium rule prioritizes downloading a second application over a first application in response to the amount of power available to the mobile device from the battery being below a predetermined value. The server receives a request to upload the first application and the second application to the mobile device and a rule selection message. The server uploads to the mobile device the first application after uploading the second application based on the selected medium rule.

Selectively uploading applications to a mobile device based on power consumption

A computer-implemented method, system, and/or computer program product controls downloading of an application to a mobile device based on power consumption of the application. A server identifies a first power consumption level requirement of a first application and a second power consumption level requirement of a second application, a first priority rating of the first application and a second priority rating of the second application, a power descriptor that identifies an amount of power available to a mobile device, and a set of application downloading rules. The set of application downloading rules includes a soft rule, a hard rule, and a medium rule. The server then selectively uploads, to the mobile device, the first application or the second application based on a selected rule.

A computer-implemented method, system, and/or computer program product controls downloading of an application to a mobile device based on power consumption of the application. A server receives a first power nameplate for a first application and a second power nameplate for a second application, a first priority rating of the first application and a second priority rating of the second application, a power descriptor that identifies an amount of power available to a mobile device, and a set of application downloading rules. The set of application downloading rules includes a soft rule, a hard rule, and a medium rule. The server then selectively uploads, to the mobile device, the first application or the second application based on a selected rule.

A clustered computing data processing system can be configured for inventory management in a highly available cluster. The system can include a cluster of computers managed by a cluster agent and providing high availability for a hosted application in the cluster. The system also can include a systems management tool executing in memory of a host computer system and coupled to the cluster agent and maintaining an inventory of data pertaining to the computers in the cluster. The system yet further can include an Internet protocol (IP) alias discover module coupled to the tool. The module can include program code executing in the memory of the host computer system so as to receive notification from the cluster agent of a change to addressing data for the computers in the cluster and, responsive to the notification, to update the inventory to reflect the change to the addressing data.

Inventory updating of an internet protocol (IP) alias within a highly available computing cluster

Computer-implemented methods for workload management and related computer program products are disclosed. One method comprises receiving corrosion rate signals from a first sensor associated with a first compute node, determining a first corrosion level for the first compute node as a function of the corrosion rate signal received from the first sensor, and automatically removing a first workload from the first compute node in response to determining that the first compute node has a first corrosion level that is greater than a setpoint level of corrosion.

Growing community of inventors

Bengaluru, India

Rajesh K Jeyapaul