Allocation of shared computing resources using a classifier chain

1. A method, comprising:

obtaining an application to be executed in a shared computing environment having a plurality of resources with a plurality of combinations of one or more hardware types;

obtaining, using at least one processing device, a plurality of discriminative features for the application to be executed;

obtaining, using the at least one processing device, at least one trained machine learning classifier chain, wherein the at least one trained machine learning classifier chain is trained using historical data and comprises a plurality of classifiers, wherein the plurality of classifiers comprises a classifier for each of a plurality of hardware types, wherein the at least one trained machine learning classifier chain is trained using a set of discriminative features obtained for a plurality of executed applications and a corresponding combination of hardware types allocated to each executed application and wherein each of the classifiers in the plurality of classifiers in the chain is trained using the set of discriminative features obtained for the plurality of executed applications and the combination of hardware types assigned by each of zero or more prior classifiers in the plurality of classifiers in the chain; and

generating, using the at least one processing device and the at least one trained machine learning classifier chain, a prediction, using the discriminative features, of the combination of hardware types needed to satisfy one or more service level agreement requirements for the application to be executed in the shared computing environment.

2. The method of claim 1 , wherein the at least one trained machine learning classifier chain is further trained using a corresponding measurement of one or more metrics of said service level agreement requirements obtained for each executed application.

3. The method of claim 2 , wherein the set of discriminative features obtained for each of the plurality of executed applications is extracted from source code for each executed application.

4. The method of claim 2 , wherein the corresponding combination of hardware types allocated to each executed application indicates the particular one or more hardware types allocated to each executed application.

5. The method of claim 1 , wherein the at least one trained machine learning classifier chain models correlations among the combinations of the one or more hardware types.

6. The method of claim 1 , wherein the generating step determines a classification for each of the classifiers in the plurality of classifiers in the classifier chain using the set of discriminative features obtained for the application and the combination of hardware types assigned by each of zero or more prior classifiers in the plurality of classifiers in the chain.

7. A computer program product, comprising a non-transitory machine-readable storage medium having encoded therein executable code of one or more software programs, wherein the one or more software programs when executed by at least one processing device perform the following steps:

obtaining an application to be executed in a shared computing environment having a plurality of resources with a plurality of combinations of one or more hardware types;

obtaining a plurality of discriminative features for the application to be executed;

obtaining at least one trained machine learning classifier chain, wherein the at least one trained machine learning classifier chain is trained using historical data and comprises a plurality of classifiers, wherein the plurality of classifiers comprises a classifier for each of a plurality of hardware types, wherein the at least one trained machine learning classifier chain is trained using a set of discriminative features obtained for a plurality of executed applications and a corresponding combination of hardware types allocated to each executed application and wherein each of the classifiers in the plurality of classifiers in the chain is trained using the set of discriminative features obtained for the plurality of executed applications and the combination of hardware types assigned by each of zero or more prior classifiers in the plurality of classifiers in the chain; and

generating, using the at least one processing device and the at least one trained machine learning classifier chain, a prediction, using the discriminative features, of the combination of hardware types needed to satisfy one or more service level agreement requirements for the application to be executed in the shared computing environment.

8. The computer program product of claim 7 , wherein the at least one trained machine learning classifier chain is further trained using a corresponding measurement of one or more metrics of said service level agreement requirements obtained for each executed application.

9. The computer program product of claim 8 , wherein the set of discriminative features obtained for each of the plurality of executed applications is extracted from source code for each executed application.

10. The computer program product of claim 8 , wherein the corresponding combination of hardware types allocated to each executed application indicates the particular one or more hardware types allocated to each executed application.

11. The computer program product of claim 7 , wherein the generating step determines a classification for each of the classifiers in the plurality of classifiers in the classifier chain using the set of discriminative features obtained for the application and the combination of hardware types assigned by each of zero or more prior classifiers in the plurality of classifiers in the chain.

12. An apparatus, comprising:

a memory; and

at least one processing device, coupled to the memory, operative to implement the following steps:

obtaining an application to be executed in a shared computing environment having a plurality of resources with a plurality of combinations of one or more hardware types;

obtaining a plurality of discriminative features for the application to be executed;

obtaining at least one trained machine learning classifier chain, wherein the at least one trained machine learning classifier chain is trained using historical data and comprises a plurality of classifiers, wherein the plurality of classifiers comprises a classifier for each of a plurality of hardware types, wherein the at least one trained machine learning classifier chain is trained using a set of discriminative features obtained for a plurality of executed applications and a corresponding combination of hardware types allocated to each executed application and wherein each of the classifiers in the plurality of classifiers in the chain is trained using the set of discriminative features obtained for the plurality of executed applications and the combination of hardware types assigned by each of zero or more prior classifiers in the plurality of classifiers in the chain; and

generating, using the at least one processing device and the at least one trained machine learning classifier chain, a prediction, using the discriminative features, of the combination of hardware types needed to satisfy one or more service level agreement requirements for the application to be executed in the shared computing environment.

13. The apparatus of claim 12 , wherein the at least one trained machine learning classifier chain is further trained using a corresponding measurement of one or more metrics of said service level agreement requirements obtained for each executed application.

14. The apparatus of claim 13 , wherein the set of discriminative features obtained for each of the plurality of executed applications is extracted from source code for each executed application.

15. The apparatus of claim 13 , wherein the corresponding combination of hardware types allocated to each executed application indicates the particular one or more hardware types allocated to each executed application.

16. The apparatus of claim 12 , wherein the at least one trained machine learning classifier chain models correlations among the combinations of the one or more hardware types.

17. The apparatus of claim 12 , wherein the generating step determines a classification for each of the classifiers in the plurality of classifiers in the classifier chain using the set of discriminative features obtained for the application and the combination of hardware types assigned by each of zero or more prior classifiers in the plurality of classifiers in the chain.

18. The method of claim 1 , wherein the set of discriminative features obtained for the plurality of executed applications comprises characteristics of each executed application and wherein the corresponding combination of hardware types allocated to each executed application comprises an amount of compute resources allocated to each executed application.

19. The computer program product of claim 8 , wherein the set of discriminative features obtained for the plurality of executed applications comprises characteristics of each executed application and wherein the corresponding combination of hardware types allocated to each executed application comprises an amount of compute resources allocated to each executed application.

20. The apparatus of claim 12 , wherein the set of discriminative features obtained for the plurality of executed applications comprises characteristics of each executed application and wherein the corresponding combination of hardware types allocated to each executed application comprises an amount of compute resources allocated to each executed application.