System and method for capacity forecasting in backup systems

1. A backup system for storing a copy of client data, comprising:

a persistent storage for storing the copy of the client data; and

a forecaster programmed to:

obtain storage use rates of the persistent storage over a prior time period;

select a first prediction period based on the prior time period;

partition the prior time period to obtain a plurality of partitioned time periods;

fuzzify the storage use rates over the prior time period to obtain fuzzified storage use rates over the prior time period;

apply a higher order fuzzy time series operator to the fuzzified storage use rates over the prior time period to obtain a plurality of higher order forecasting functions;

calculate a fuzzified forecast for the selected first prediction period using the higher order forecasting functions; and

defuzzify the fuzzified forecast to obtain a forecast for the first prediction period.

2. The backup system of claim 1 , wherein the forecaster is further programmed to:

make a determination that the forecast specified a used storage capacity that is larger than a predetermined storage capacity; and

in response to the determination, modify a backup policy associated with the client data to reduce a rate of storing the copy of the client data in the persistent storage.

3. The backup system of claim 2 , wherein modifying the backup policy comprises:

reducing a frequency of backup up the copy of the client data.

4. The backup system of claim 2 , wherein modifying the backup policy comprises:

selecting a portion of the client data; and

removing the selected portion from the client data specified for backup during future backups of the client data.

5. The backup system of claim 2 , wherein the predetermined storage capacity is a storage capacity of the persistent storage.

6. The backup system of claim 1 , wherein the forecaster is further programmed to:

make a determination that the forecast specified a used storage capacity that is larger than a predetermined storage capacity; and

in response to the determination, notify a user of the determination.

7. The backup system of claim 6 , wherein the predetermined storage capacity is half of the storage capacity of the persistent storage.

8. The backup system of claim 1 , wherein obtaining the storage use rates of the persistent storage over the prior time period comprises:

monitoring a used storage capacity of the persistent storage on a daily basis;

for each day of the prior time period, identifying a maximum used storage capacity during the respective day based on the monitoring; and

generating a time series relationship between each day of the prior time period and the respective identified maximum used storage capacity during the respective day.

9. The backup system of claim 8 , wherein fuzzyifying the storage use rates over the prior time period comprises:

obtaining a plurality of linguistic terms;

matching each identified maximum used storage capacity to a linguistic term of the plurality of linguistic terms; and

generating a time series relationship between each day of the prior time period and the linguistic term matched to the respective identified maximum used storage capacity during the respective day.

10. The backup system of claim 1 , wherein a forecast for the first prediction period specifies a used storage capacity over the first prediction period.

11. The backup system of claim 1 , wherein the forecaster is further programmed to:

make a first determination that the forecast specified a used storage capacity that is larger than a first predetermined storage capacity and less than a second predetermined storage capacity; and

in response to the first determination, notify a user of the determination.

12. The backup system of claim 11 , wherein the forecaster is further programmed to:

make a second determination that a second forecast associated with a second prediction time period specified a second used storage capacity that is larger than the second predetermined storage capacity; and

in response to the second determination, modify a backup policy associated with the client data to reduce a rate of storing the copy of the client data in the persistent storage,

wherein the second used storage capacity is larger than the used storage capacity.

13. A method for operating a backup system, comprising:

obtaining storage use rates of a persistent storage over a prior time period, wherein the persistent storage stores a copy of client data;

selecting a first prediction period based on the prior time period;

partitioning the prior time period to obtain a plurality of partitioned time periods;

fuzzifying the storage use rates over the prior time period to obtain fuzzified storage use rates over the prior time period;

applying a higher order fuzzy time series operator to the fuzzified storage use rates over the prior time period to obtain a plurality of higher order forecasting functions;

calculating a fuzzified forecast for the selected first prediction period using the high order forecasting functions; and

defuzzifying the fuzzified forecast to obtain a forecast for the first prediction period.

14. The method of claim 13 , further comprising:

making a determination that the forecast specified a used storage capacity that is larger than a predetermined storage capacity; and

in response to the determination, modifying a backup policy associated with the client data to reduce a rate of storing the copy of the client data in the persistent storage.

15. The method of claim 14 , wherein modifying the backup policy comprises:

selecting a portion of the client data; and

removing the selected portion from the client data specified for backup during future backups of the client data.

16. The method of claim 14 , wherein modifying the backup policy comprises:

reducing a frequency of backup up the copy of the client data.

17. A non-transitory computer readable medium comprising computer readable program code, which when executed by a computer processor enables the computer processor to perform a method for operating a backup system, the method comprising:

obtaining storage use rates of a persistent storage over a prior time period, wherein the persistent storage stores a copy of client data;

selecting a first prediction period based on the prior time period;

partitioning the prior time period to obtain a plurality of partitioned time periods;

fuzzifying the storage use rates over the prior time period to obtain fuzzified storage use rates over the prior time period;

applying a higher order fuzzy time series operator to the fuzzified storage use rates over the prior time period to obtain a plurality of higher order forecasting functions;

calculating a fuzzified forecast for the selected first prediction period using the high order forecasting functions; and

defuzzifying the fuzzified forecast to obtain a forecast for the first prediction period.

18. The non-transitory computer readable medium of claim 17 , wherein the method further comprises:

making a determination that the forecast specified a used storage capacity that is larger than a predetermined storage capacity; and

in response to the determination, modifying a backup policy associated with the client data to reduce a rate of storing the copy of the client data in the persistent storage.

19. The non-transitory computer readable medium of claim 18 , wherein modifying the backup policy comprises:

selecting a portion of the client data; and

removing the selected portion from the client data specified for backup during future backups of the client data.

20. The non-transitory computer readable medium of claim 18 , wherein modifying the backup policy comprises:

reducing a frequency of backup up the copy of the client data.