Storage resource capacity prediction utilizing a plurality of time series forecasting models

1. A method comprising:

monitoring one or more storage systems to obtain historical storage resource utilization data characterizing resource utilization for a given set of storage resources of the one or more storage systems for a first period of time;

generating a plurality of different model-specific storage resource capacity predictions of resource utilization for the given set of storage resources for a second period of time subsequent to the first period of time, the plurality of different model-specific storage resource capacity predictions being generated by applying the historical storage resource utilization data as input to respective ones of a plurality of different time series capacity prediction forecasting models that take into account one or more different factors, including at least one of a seasonality factor and a trend factor;

evaluating the plurality of different model-specific storage resource capacity predictions utilizing one or more empirical conditions, the one or more empirical conditions comprising at least one of noise and uncertainty;

selecting a subset of the plurality of model-specific storage resource capacity predictions having one or more designated characteristics, the one or more designated characteristics comprising at least one of a level of noise below a designated noise threshold and a level of uncertainty below a designated uncertainty threshold;

determining an overall storage resource capacity prediction based at least in part on a combination of the selected subset of the model-specific storage resource capacity predictions; and

modifying a provisioning of storage resources of the one or more storage systems based at least in part on the overall storage resource capacity prediction;

wherein the method is performed by at least one processing device comprising a processor coupled to a memory.

2. The method of claim 1 wherein the given set of storage resources comprises a storage pool allocated for a given user, the storage pool comprising a first set of storage resources allocated from a first one of the one or more storage systems and a second set of storage resources allocated from a second one of the one or more storage systems.

3. The method of claim 1 wherein the given set of storage resources comprises a portion of a storage pool allocated for a given user, the portion of the storage pool comprising storage resources allocated from a single one of the one or more storage systems.

4. The method of claim 1 wherein the plurality of time series capacity prediction forecasting models comprises:

at least a first time series capacity prediction forecasting model that takes into account a first type of seasonality and trend factors; and

at least a second time series capacity prediction forecasting model that takes into account a second type of seasonality and trend factors.

5. The method of claim 1 wherein the first time series capacity prediction forecasting model comprises at least one of a Seasonality Trend using Loess (STL) model and an Autoregressive Integrated Moving Average (ARIMA)-based model.

6. The method of claim 1 wherein generating the plurality of model-specific storage resource capacity predictions further comprises applying pre-processing by passing the historical storage resource utilization data through one or more filters.

7. The method of claim 6 wherein the one or more filters comprises at least one of:

a first filter configured to impute missing data in the historical storage resource utilization data;

a second filter configured to remove additive outliers in the historical storage resource utilization data; and

a third filter configured to remove level change outliers in the historical storage resource utilization data.

8. The method of claim 1 wherein the model-specific storage resource capacity predictions comprise respective forecasts each with an associated cone of uncertainty defined by a forecast storage resource utilization value, an upper bound on the forecast utilization value, and a lower bound on the forecast storage resource utilization value.

9. The method of claim 8 wherein selecting the subset of the model-specific storage resource capacity predictions having the one or more designated characteristics comprises selecting respective ones of the model-specific storage resource capacity predictions having associated cones of uncertainty with differences between the upper bound and the lower bound less than the designated uncertainty threshold.

10. The method of claim 1 wherein selecting the subset of the model-specific storage resource capacity predictions having the one or more designated characteristics comprises selecting respective ones of the model-specific storage resource capacity predictions having associated noise less than the designated noise threshold.

11. The method of claim 1 wherein determining the overall storage resource capacity prediction comprises calculating the overall storage resource capacity prediction as a weighted average of the selected subset of the model-specific storage resource capacity predictions, wherein weights for the selected subset of the model-specific storage resource capacity predictions are based at least in part on historical performance of respective ones of the time series capacity prediction forecasting models used to generate the selected subset of the model-specific storage resource capacity predictions.

12. The method of claim 1 further comprising determining a given one of a plurality of different time ranges when available storage resources of the one or more storage systems is expected to fall below a designated capacity threshold based on the overall storage resource capacity prediction, and wherein modifying the provisioning of the storage resources of the one or more storage systems is based at least in part on the given time range when the available storage resources of the one or more storage systems is expected to fall below the designated capacity threshold.

13. The method of claim 12 wherein when the given time range is within a designated time from a current time, modifying the provisioning of the storage resources of the one or more storage systems comprises increasing storage resources of the one or more storage systems.

14. The method of claim 12 wherein when the given time range is greater than a designated time from a current time, modifying the provisioning of the storage resources of the one or more storage systems comprises generating an alert indicating an expected time when the available storage resources of the one or more storage systems is expected to fall below the designated capacity threshold.

15. A computer program product comprising a non-transitory processor-readable storage medium having stored therein program code of one or more software programs, wherein the program code when executed by at least one processing device causes the at least one processing device:

to monitor one or more storage systems to obtain historical storage resource utilization data characterizing resource utilization for a given set of storage resources of the one or more storage systems for a first period of time;

to generate a plurality of different model-specific storage resource capacity predictions of resource utilization for the given set of storage resources for a second period of time subsequent to the first period of time, the plurality of different model-specific storage resource capacity predictions being generated by applying the historical storage resource utilization data as input to respective ones of a plurality of different time series capacity prediction forecasting models that take into account one or more different factors, including at least one of a seasonality factor and a trend factor;

to evaluate the plurality of different model-specific storage resource capacity predictions utilizing one or more empirical conditions, the one or more empirical conditions comprising at least one of noise and uncertainty;

to select a subset of the plurality of model-specific storage resource capacity predictions having one or more designated characteristics, the one or more designated characteristics comprising at least one of a level of noise below a designated noise threshold and a level of uncertainty below a designated uncertainty threshold;

to determine an overall storage resource capacity prediction based at least in part on a combination of the selected subset of the model-specific storage resource capacity predictions; and

to modify a provisioning of storage resources of the one or more storage systems based at least in part on the overall storage resource capacity prediction.

16. The computer program product of claim 15 wherein:

the model-specific storage resource capacity predictions comprise respective forecasts each with an associated cone of uncertainty defined by a forecast storage resource utilization value, an upper bound on the forecast utilization value, and a lower bound on the forecast storage resource utilization value; and

selecting the subset of the model-specific storage resource capacity predictions having the one or more designated characteristics comprises selecting respective ones of the model-specific storage resource capacity predictions having associated cones of uncertainty with differences between the upper bound and the lower bound less than the designated uncertainty threshold.

17. The computer program product of claim 15 wherein selecting the subset of the model-specific storage resource capacity predictions having the one or more designated characteristics comprises selecting respective ones of the model-specific storage resource capacity predictions having associated noise less than the designated noise threshold.

18. An apparatus comprising:

at least one processing device comprising a processor coupled to a memory;

the at least one processing device being configured:

to monitor one or more storage systems to obtain historical storage resource utilization data characterizing resource utilization for a given set of storage resources of the one or more storage systems for a first period of time;

to generate a plurality of different model-specific storage resource capacity predictions of resource utilization for the given set of storage resources for a second period of time subsequent to the first period of time, the plurality of different model-specific storage resource capacity predictions being generated by applying the historical storage resource utilization data as input to respective ones of a plurality of different time series capacity prediction forecasting models that take into account one or more different factors, including at least one of a seasonality factor and a trend factor;

to evaluate the plurality of different model-specific storage resource capacity predictions utilizing one or more empirical conditions, the one or more empirical conditions comprising at least one of noise and uncertainty;

to select a subset of the plurality of model-specific storage resource capacity predictions having one or more designated characteristics, the one or more designated characteristics comprising at least one of a level of noise below a designated noise threshold and a level of uncertainty below a designated uncertainty threshold;

to determine an overall storage resource capacity prediction based at least in part on a combination of the selected subset of the model-specific storage resource capacity predictions; and

to modify a provisioning of storage resources of the one or more storage systems based at least in part on the overall storage resource capacity prediction.

19. The apparatus of claim 18 wherein:

the model-specific storage resource capacity predictions comprise respective forecasts each with an associated cone of uncertainty defined by a forecast storage resource utilization value, an upper bound on the forecast utilization value, and a lower bound on the forecast storage resource utilization value; and

selecting the subset of the model-specific storage resource capacity predictions having the one or more designated characteristics comprises selecting respective ones of the model-specific storage resource capacity predictions having associated cones of uncertainty with differences between the upper bound and the lower bound less than the designated uncertainty threshold.

20. The apparatus of claim 18 wherein selecting the subset of the model-specific storage resource capacity predictions having the one or more designated characteristics comprises selecting respective ones of the model-specific storage resource capacity predictions having associated noise less than the designated noise threshold.