Recommendation and deployment engine and method for machine learning based processes in hybrid cloud environments
Methods and systems are provided for the deployment of machine learning based processes to public clouds. For example, a method for deploying a machine learning based process may include developing and training the machine learning based process to perform an activity, performing at least one of identifying and receiving an identification of a set of one or more public clouds that comply with a set of regulatory criteria used to regulate the activity, selecting a first public cloud of the set of one or more public clouds that complies with the set of regulatory criteria used to regulate the activity, and deploying the machine learning based process to the first public cloud of the set of one or more public clouds.
1. A method comprising:
developing and training, using a private cloud, a machine learning based process to perform an activity;
providing the trained machine learning based process as a service using the private cloud to a number of users of the private cloud;
identifying a recommended set of a plurality of candidate public clouds, wherein each of the plurality of candidate public clouds complies with a set of regulatory criteria used to regulate the activity and wherein the set of regulatory criteria represent regulations enacted by a government entity or a private association;
receiving a selection of a candidate public cloud of the recommended set; and
during a spike in demand on resources of the private cloud, performing a cloud bursting operation to deploy the trained machine learning based process to the selected candidate public cloud.
2. The method of claim 1 , wherein
the activity is a healthcare-related activity; and
the set of regulatory criteria include privacy regulations for individual patients.
3. The method of claim 1 , wherein the activity is fraud detection for a desired geographic region.
4. The method of claim 1 , further comprising:
identifying one or more candidate public clouds of the recommended set that reside within a desired geographic region, wherein
said receiving a selection includes limiting the selection to those candidate public clouds of the recommended set that reside within the desired geographic region.
5. The method of claim 1 , further comprising:
identifying one or more candidate public clouds of the recommended set that meet a minimum desired set of performance criteria, wherein
said receiving a selection includes limiting selection to those candidate public clouds of the recommended set that meet the minimum desired set of performance criteria.
6. The method of claim 1 , further comprising:
identifying one or more candidate public clouds of the recommended set that meet a desired budget constraint, wherein
said receiving a selection includes limiting the selection to those candidate public clouds that meet the desired budget constraint.
7. The method of claim 1 , further comprising:
identifying one or more candidate public clouds of the recommended set that is controlled by a desired entity, wherein
said receiving a selection includes limiting the selection to those candidate public clouds that are controlled by the desired entity.
8. The method of claim 1 , further comprising:
identifying one or more candidate public clouds of the recommended set that reside within a desired geographic region;
identifying one or more candidate public clouds of the recommended set that meet a minimum desired set of performance criteria; and
identifying one or more candidate public clouds of the recommended set that meet a desired budget constraint, wherein
said receiving a selection includes limiting the selection to those public clouds that reside within the desired geographic region, meet the minimum desired set of performance criteria, and meet the desired budget constraint.
9. The method of claim 8 , further comprising:
identifying one or more candidate public clouds of the recommended set that is controlled by a desired entity, wherein
said receiving a selection includes limiting the selection to those public clouds that are controlled by the desired entity.
10. The method of claim 8 , wherein
the activity is a healthcare-related activity; and
the set of regulatory criteria include privacy regulations for individual patients.
11. The method of claim 8 , wherein the activity is fraud detection for the desired geographic region.
12. The method of claim 8 , further comprising deploying the selected candidate public cloud such that the selected candidate public cloud complies with the set of regulatory criteria used to regulate the activity.
13. The method of claim 1 , further comprising designing and deploying the selected candidate public cloud such that the selected candidate public cloud complies with the set of regulatory criteria used to regulate the activity.
14. A non-transitory computer-readable medium comprising computer executable instructions stored thereon that when executed by a processor, cause the processor to:
identify a recommended set of a plurality of candidate public clouds, wherein each of the plurality of candidate public clouds complies with a set of regulatory criteria used to regulate an activity performed by a machine learning based process provided as a service using a private cloud to a number of users of the private cloud in which the machine learning based process was developed and trained and wherein the set of regulatory criteria represent regulations enacted by a government entity or a private association;
receive a selection of a candidate public cloud of the recommended set; and
during a spike in demand on resources of the private cloud, perform a cloud bursting operation to deploy the machine learning based process to the selected candidate public cloud.
15. The computer program product of claim 14 , wherein:
selection of the candidate public cloud of the recommended set further takes into consideration at least one of a desired geographic region, a minimum desired set of performance criteria, and a desired budget constraint.
16. The computer program product of claim 15 , wherein:
selection of the candidate public cloud of the recommended set further takes into consideration all of a desired geographic region, a minimum desired set of performance criteria, and a desired budget constraint.
17. A device comprising:
a processor and memory communicatively coupled to the processor, the memory containing instructions that cause the processor to:
develop and train, using a private cloud, a machine learning based process to perform an activity;
provide the trained machine learning based process as a service using the private cloud to a number of users of the private cloud;
identify a recommended set of a plurality of candidate public clouds, wherein each of the plurality of candidate public clouds complies with a set of regulatory criteria used to regulate the activity and wherein the set of regulatory criteria represent regulations enacted by a government entity or a private association;
receive a selection of a candidate public cloud of the recommended set; and
during a spike in demand on resources of the private cloud, perform a cloud bursting operation to deploy the trained machine learning based process to the selected candidate public cloud.