License orchestrator to most efficiently distribute fee-based licenses

1. An apparatus comprising:

a processor; and

a memory that stores code executable by the processor to:

train a neural network using machine learning during a training phase to build a use history the use history comprising a per-use licensing cost of one or more licensable resources available to one or more computing devices and a cost of using existing capabilities of the one or more computing devices, the neural network using workload analytics from the one or more computing devices to build the use history of workloads comparable to a workload scheduled to be executed by a computing device of the one or more computing devices;

determine, using a workload scheduler, that the workload is scheduled to be executed by the computing device, the computing device comprising a licensable resource available for execution of the workload at a time the workload is scheduled for execution;

compare, using a cost comparator, a per-use licensing cost associated with using the licensable resource for execution of the workload with a cost of using existing capabilities of the computing device for execution of the workload, wherein the per-use licensing cost comprises a monetary cost of the per-use license at the time of execution of the workload plus a monetized cost associated with execution of the workload on the licensable resource at the time of execution of the workload and the cost of using existing capabilities comprises a monetized cost associated with execution of the workload with the existing capabilities at the time of execution of the workload;

automatically license, using a license distributor, the licensable resource without user input and use the licensable resource for a duration of execution of the workload, using a workload executor, at the scheduled time in response to determining that, at the time the workload is scheduled for execution, the per-use licensing cost of the licensable resource is less than using the existing capabilities of the computing device; and

use the existing capabilities of the computing device to execute the workload, using the workload executor, at the scheduled time without licensing the licensable resource in response to determining that, at the time the workload is scheduled for execution, using the existing capabilities of the computing device is less than the per-use licensing cost.

2. The apparatus of claim 1 , wherein the code is further executable by the processor to determine if a license for the licensable resource is available from a pool of licenses in response to determining that the workload is scheduled to be executed by the computing device.

3. The apparatus of claim 1 , wherein the per-use licensing cost includes the cost of the license.

4. The apparatus of claim 1 , wherein the licensable resource comprises a hardware resource.

5. The apparatus of claim 4 , wherein the licensable resource comprises one or more ports of a network switch, a graphical processing unit (“GPU”), an accelerator, an additional CPU, a field programmable gate array (“FPGA”), and/or remote access of a baseboard management controller (“BMC”).

6. The apparatus of claim 1 , wherein the licensable resource comprises a software feature of an application.

7. The apparatus of claim 1 , wherein the computing device is an edge computing device and/or a computing device at a location of the edge computing device and the license is available from a remote license orchestrator.

8. The apparatus of claim 7 , wherein the code executable by the processor to determine that the workload is scheduled to be executed by the computing device, compare the per-use licensing cost with the cost of using the existing capabilities of the computing device, and license and use the licensable resource is located in memory of the edge computing device.

9. A method comprising:

training a neural network using machine learning during a training phase to build a use history the use history comprising a per-use licensing cost of one or more licensable resources available to one or more computing devices and a cost of using existing capabilities of the one or more computing devices, the neural network using workload analytics from the one or more computing devices to build the use history of workloads comparable to a workload scheduled to be executed by a computing device of the one or more computing devices;

determining, using a workload scheduler, that the workload is scheduled to be executed by the computing device, the computing device comprising a licensable resource available for execution of the workload at a time the workload is scheduled for execution;

comparing, using a cost comparator, a per-use licensing cost associated with using the licensable resource for execution of the workload with a cost of using existing capabilities of the computing device for execution of the workload, wherein the per-use licensing cost comprises a monetary cost of the per-use license at the time of execution of the workload plus a monetized cost associated with execution of the workload on the licensable resource at the time of execution of the workload and the cost of using existing capabilities comprises a monetized cost associated with execution of the workload with the existing capabilities at the time of execution of the workload;

automatically licensing, using a license distributor, the licensable resource without user input and using the licensable resource for a duration of execution of the workload, using a workload executor, at the scheduled time in response to determining that, at the time the workload is scheduled for execution, the per-use licensing cost of the licensable resource is less than using the existing capabilities of the computing device; and

using the existing capabilities of the computing device to execute the workload using the workload executor, at the scheduled time without licensing the licensable resource in response to determining that, at the time the workload is scheduled for execution, using the existing capabilities of the computing device is less than the per-use licensing cost.

10. The method of claim 9 , further comprising determining when a license for the licensable resource is available from a pool of licenses in response to determining that the workload is scheduled to be executed by the computing device.

11. The method of claim 9 , wherein the licensable resource comprises a hardware resource and/or a software resource.

12. A program product comprising a computer readable storage medium comprising program code, the program code being configured to be executable by a processor to perform operations comprising:

training a neural network using machine learning during a training phase to build a use history the use history comprising a per-use licensing cost of one or more licensable resources available to one or more computing devices and a cost of using existing capabilities of the one or more computing devices, the neural network using workload analytics from the one or more computing devices to build the use history of workloads comparable to a workload scheduled to be executed by a computing device of the one or more computing devices;

determining, using a workload scheduler, that the workload is scheduled to be executed by the computing device, the computing device comprising a licensable resource available for execution of the workload at a time the workload is scheduled for execution;

comparing, using a cost comparator, a per-use licensing cost associated with using the licensable resource for execution of the workload with a cost of using existing capabilities of the computing device for execution of the workload, wherein the per-use licensing cost comprises a monetary cost of the per-use license at the time of execution of the workload plus a monetized cost associated with execution of the workload on the licensable resource at the time of execution of the workload and the cost of using existing capabilities comprises a monetized cost associated with execution of the workload with the existing capabilities at the time of execution of the workload;

automatically licensing, using a license distributor, the licensable resource without user input and using the licensable resource for a duration of execution of the workload, using a workload executor, at the scheduled time in response to determining that, at the time the workload is scheduled for execution, the per-use licensing cost of the licensable resource is less than using the existing capabilities of the computing device; and

using the existing capabilities of the computing device to execute the workload, using the workload executor, at the scheduled time without licensing the licensable resource in response to determining that, at the time the workload is scheduled for execution, using the existing capabilities of the computing device is less than the per-use licensing cost.

13. The program product of claim 12 , the code further configured to be executable by a processor to perform operations comprising determining if a license for the licensable resource is available from a pool of licenses in response to determining that the workload is scheduled to be executed by the computing device.