Resource prediction for cloud computing

1. A method for predicting an allocation of processing resources provided by a cloud computing module, the method comprising:

detecting input parameters containing information about at least a data set to be processed by the cloud computing module and a predefined processing task to be carried out on the data set, wherein the detecting is based on limiting the input para meters to border conditions determined by a plurality of different models provided in a model database, and wherein each model of the plurality of different models provides a relationship between at least one of predefined processing tasks and a predicted allocation of the processing resources;

based on a determination that a model of the plurality of different models used to predict the allocation of the processing resources for the detected input parameters, is not available in the model database:

determining, based on a plurality of available processing resource configurations provided by the cloud computing module, that not enough meta data elements are available to generate the model, and in response, initiating generation of a test data set, wherein the test data set is defined by one or more parameters of an available processing resource configuration of the plurality of available processing resource configurations, wherein the plurality of available processing resource configurations indicates possible combinations of the processing resources and the predefined processing tasks available to the cloud computing module;

processing the test data set using the processing resources and the predefined processing task in order to generate a meta data element for the test data set;

determining a time frame needed to carry out the predefined processing task on the test data set;

generating the model based on the time frame and the processing resources allocated to carry out the predefined processing task on the test data set; and

predicting the allocation of the processing resources based on the generated model; and

based on a determination that a model of the plurality of different models used to predict the allocation of the processing resources for the detected input parameters, is available in the model database:

selecting the model from the plurality of different models provided in the model database; and

predicting the allocation of the processing resources based on both the selected model and the detected input parameters.

2. The method of claim 1 , wherein the predicting the allocation of the processing resources based on both the selected model and the detected input parameters comprises suggesting at least one of a processing resource configuration and a time needed by the processing resources of the suggested at least one processing resource configuration to carry out the predefined processing task on the data set, taking into account the detected input parameters and the plurality of available processing resource configurations provided by the cloud computing module.

3. The method of claim 2 :

further comprising providing a historical database containing information about historical processing events, each historical processing event comprising information about:

which of the predefined processing tasks was carried out on a historical data set;

which processing resources were allocated for processing the historical data set; and

a further time frame needed for the processing of the historical data set;

wherein the information about each historical processing event is stored in a corresponding meta data element.

4. The method of claim 2 , further comprising storing the allocated processing resources and the further time frame needed to carry out the predefined processing task as a meta data element in the historical database.

5. The method of claim 2 :

wherein the detected input parameters contain an additional input parameter, the additional input parameter containing information about one of the plurality of available processing resource configurations and the time needed by the suggested at least one processing resource configuration to carry out the predefined processing task, and

wherein the suggesting comprises suggesting another processing resource configuration and the time needed to carry out the predefined processing task taking into account the detected input parameters, the additional input parameter, and the plurality of available processing resource configurations.

6. The method of claim 1 , further comprising:

generating the plurality of different models provided in the model database; and

generating a classification function for a given model of the plurality of different models, wherein the classification function describes a relationship between the predefined processing task corresponding to the given model and the predicted allocation of the processing resources, based on both the selected model and the detected input parameters, corresponding to the given model based on information about how much processing resources were allocated for carrying out different processing tasks on historical data sets.

7. The method of claim 6 , further comprising checking in intervals whether the plurality of different models provided in the model database is enough for the plurality of available processing resource configurations and/or have to be updated.

8. The method of claim 2 :

wherein the processing resources and the predefined processing tasks are identified, and

wherein the plurality of available processing resource configurations is determined based on the processing resources and the predefined processing tasks.

9. The method of claim 1 , further comprising:

comparing the predicted allocation of the processing resources, based on both the selected model and the detected input parameters, to an actual allocation of the processing resources when the predefined processing task is carried out on the data set; and

initiating, in response to determining that the predicted allocation of the processing resources based on both the selected model and the detected input parameters differs from the actual allocation of the processing resources by more than a predefined threshold, an amendment of the model used for the prediction.

10. A system configured to predict an allocation of processing resources provided by a cloud computing module, the system comprising:

a processor; and

memory containing instructions executable by the processor whereby the processor is configured to:

detect input parameters containing information about at least a data set to be processed by the cloud computing module and a predefined processing task to be carried out on the data set, wherein the detection is based on limiting the input para meters to border conditions determined by a plurality of different models provided in a model database, and wherein each model of the plurality of different models provides a relationship between at least one of predefined processing tasks and a predicted allocation of the processing resources;

based on a determination that a model of the plurality of different models used to predict the allocation of the processing resources for the detected input parameters, is not available in the model database, the processor is configured to:

determine, based on a plurality of available processing resource configurations provided by the cloud computing module, that not enough meta data elements are available to generate the model, and in response, initiate generation of a test data set, wherein the test data set is defined by one or more parameters of an available processing resource configuration of the plurality of available processing resource configurations, wherein the plurality of available processing resource configurations indicates possible combinations of the processing resources and the predefined processing tasks available to the cloud computing module;

process the test data set using the processing resources and the predefined processing task in order to generate a meta data element for the test data set;

determine a time frame needed to carry out the predefined processing task on the test data set;

generate the model based on the time frame and the processing resources allocated to carry out the predefined processing task on the test data set; and

based on a determination that a model of the plurality of different models used to predict the allocation of the processing resources for the detected input parameters, is available in the model database, the processor is configured to:

select, from the model database containing the plurality of different models, the model; and

predict the allocation of the processing resources using the selected model and based on the detected input parameters.

11. The system of claim 10 , wherein:

the instructions are such that the processor is further configured to suggest at least one of a processing resource configuration and a time needed by the processing resources of the suggested at least one processing resource configuration to carry out the predefined processing task on the data set, taking into account the detected input parameters and the plurality of available processing resource configurations provided by the cloud computing module.

12. The system of claim 11 :

further comprising a historical database containing information about historical processing events, each historical processing event comprising information about:

which of the predefined processing tasks was carried out on a historical data set;

which processing resources were allocated for processing the historical data set; and

a further time frame needed for the processing of the historical data set,

wherein the information about each historical processing event is stored in a corresponding meta data element.

13. The system of claim 12 , wherein the instructions are such that the processor is further configured to store the allocated processing resources and the further time frame needed to carry out the predefined processing task as a meta data element in the historical database.

14. The system according to claim 11 , wherein:

the detected input parameters contain an additional input parameter, the additional input parameter containing information about one of the plurality of available processing resource configurations and the time needed by the suggested at least one processing resource configuration to carry out the predefined processing task; and

the instructions are such that the processor is further configured to suggest another processing resource configuration and the time needed to carry out the predefined processing task taking into account the detected input parameters, the additional input parameter, and the plurality of available processing resource configurations.

15. The system of claim 10 , wherein:

the instructions are such that the processor is further configured to generate the plurality of different models provided in the model database; and

to generate a given model of the plurality of different models, the instructions are such that the processor is further configured to generate a classification function describing a relationship between the predefined processing task corresponding to the given model and the predicted allocation of the processing resources, based on both the selected model and the detected input parameters, corresponding to the given model based on information about how much processing resources were allocated for carrying out different processing tasks on historical data sets.

16. The system of claim 15 , wherein the instructions are such that the processor is further configured to check in intervals whether the plurality of different models provided in the model database is enough for the plurality of available processing resource configurations and/or have to be updated.

17. The system of claim 15 , wherein the instructions are such that the processor is further configured to:

identify the processing resources and the predefined processing tasks; and

determine the plurality of available processing resource configurations based on the processing resources and the predefined processing tasks.

18. The system of claim 10 , wherein the instructions are such that the processor is further configured to:

compare the predicted allocation of the processing resources, based on both the selected model and the detected input parameters, to an actual allocation of the processing resources when the predefined processing task is carried out on the data set; and

responsive to determining that the predicted allocation of the processing resources based on both the selected model and the detected input parameters differs from the actual allocation of the processing resources by more than a predefined threshold, trigger an amendment of the model used for the prediction.