Method, electronic device, and computer program product for model training and duration prediction

1 . A method, comprising:

acquiring a first set of parameter values related to a first snapshot of a data object, the first snapshot being deleted from a storage system through a first deletion operation;

acquiring a first duration during which the first deletion operation is performed;

generating a prediction model based on at least the first set of parameter values and the first duration, the prediction model being used for determining a predicted duration required for deleting each of one or more additional snapshots from the storage system, wherein the first set of parameter values comprises at least a number of snapshots being deleted when the first snapshot is deleted and a size of the data object when the first snapshot is created, and wherein the prediction model comprises a multiple-input machine learning model comprising at least a first input configured to receive a snapshot number parameter value comprising a number of snapshots being deleted when a given one of the additional snapshots is deleted and a second input configured to receive a data object size parameter value comprising a size of the data object when the given additional snapshot is created, the multiple-input machine learning model being configured to generate, at an output thereof, the predicted duration as a weighted function of at least the snapshot number parameter value and the data object size parameter value, utilizing weight parameters that are determined at least in part in conjunction with training of the multiple-input machine learning model; and

controlling deletion of the one or more additional snapshots from the storage system based on one or more respective predicted durations generated by the prediction model for one or more additional sets of parameter values.

2 . The method according to claim 1 , wherein the first set of parameter values further comprises at least one of the following: central processing unit (CPU) usage, and a storage duration of the first snapshot.

3 . The method according to claim 1 , wherein generating the prediction model comprises:

obtaining parameters of the prediction model by training a linear regression model using the first set of parameter values and the first duration.

4 . The method according to claim 3 , further comprising:

acquiring a second set of parameter values related to a second snapshot of the data object, the second snapshot being deleted from the storage system through a second deletion operation;

acquiring a second duration during which the second deletion operation is performed;

applying the second set of parameter values to the trained linear regression model to determine a predicted duration for the second deletion operation; and

determining the availability of the trained linear regression model based on the predicted duration and the second duration.

5 . The method according to claim 4 , wherein determining the availability of the trained linear regression model comprises:

determining a fitting degree for the trained linear regression model based on the predicted duration and the second duration; and

if it is determined that the fitting degree is equal to or higher than a threshold degree, determining the trained linear regression model as an available prediction model.

6 . The method according to claim 5 , further comprising:

if it is determined that the fitting degree is less than the threshold degree, continuing training the linear regression model by using parameter values of the deleted snapshot and a duration of the corresponding deletion operation.

7 . A computer program product comprising a non-volatile computer-readable medium storing machine-executable instructions, wherein the machine-executable instructions, when executed by a machine, cause the machine to perform steps of the method according to claim 1 .

8 . A method, comprising:

receiving a request for deleting a target snapshot of a data object in a storage system, the request comprising an identification of the target snapshot;

acquiring a set of parameter values related to the target snapshot based on the identification;

determining, based on the set of parameter values, a duration required for deleting the target snapshot, wherein the set of parameter values comprises at least a number of snapshots being deleted when the target snapshot is deleted and a size of the data object when the target snapshot is created; and

controlling deletion of the target snapshot from the storage system based on the determined duration;

wherein determining a duration required for deleting the target snapshot comprises determining the duration by applying the set of parameter values to a prediction model, and wherein the prediction model comprises a multiple-input machine learning model comprising at least a first input configured to receive a snapshot number parameter value comprising the number of snapshots being deleted when the target snapshot is deleted and a second input configured to receive a data object size parameter value comprising the size of the data object when the target snapshot is created, the multiple-input machine learning model being configured to generate, at an output thereof, the duration as a weighted function of at least the snapshot number parameter value and the data object size parameter value, utilizing weight parameters that are determined at least in part in conjunction with training of the multiple-input machine learning model.

9 . The method according to claim 8 , wherein the set of parameter values further comprises at least one of the following: central processing unit (CPU) usage, and a storage duration of the target snapshot.

10 . A computer program product comprising a non-volatile computer-readable medium storing machine-executable instructions, wherein the machine-executable instructions, when executed by a machine, cause the machine to perform steps of the method according to claim 8 .

11 . An electronic device, comprising:

at least one processor; and

memory coupled to the at least one processor and having instructions stored thereon, wherein the instructions, when executed by the at least one processor, cause the electronic device to perform actions including:

acquiring a first set of parameter values related to a first snapshot of a data object, the first snapshot being deleted from a storage system through a first deletion operation;

acquiring a first duration during which the first deletion operation is performed;

generating a prediction model based on at least the first set of parameter values and the first duration, the prediction model being used for determining a predicted duration required for deleting each of one or more additional snapshots from the storage system, wherein the first set of parameter values comprises at least a number of snapshots being deleted when the first snapshot is deleted and a size of the data object when the first snapshot is created, and wherein the prediction model comprises a multiple-input machine learning model comprising at least a first input configured to receive a snapshot number parameter value comprising a number of snapshots being deleted when a given one of the additional snapshots is deleted and a second input configured to receive a data object size parameter value comprising a size of the data object when the given additional snapshot is created, the multiple-input machine learning model being configured to generate, at an output thereof, the predicted duration as a weighted function of at least the snapshot number parameter value and the data object size parameter value, utilizing weight parameters that are determined at least in part in conjunction with training of the multiple-input machine learning model; and

controlling deletion of the one or more additional snapshots from the storage system based on one or more respective predicted durations generated by the prediction model for one or more additional sets of parameter values.

12 . The electronic device according to claim 11 , wherein the first set of parameter values further comprises at least one of the following: central processing unit (CPU) usage, and a storage duration of the first snapshot.

13 . The electronic device according to claim 11 , wherein generating the prediction model comprises:

obtaining parameters of the prediction model by training a linear regression model using the first set of parameter values and the first duration.

14 . The electronic device according to claim 13 , wherein the actions further comprise:

acquiring a second set of parameter values related to a second snapshot of the data object, the second snapshot being deleted from the storage system through a second deletion operation;

acquiring a second duration during which the second deletion operation is performed;

applying the second set of parameter values to the trained linear regression model to determine a predicted duration for the second deletion operation; and

determining the availability of the trained linear regression model based on the predicted duration and the second duration.

15 . The electronic device according to claim 14 , wherein determining the availability of the trained linear regression model comprises:

determining a fitting degree for the trained linear regression model based on the predicted duration and the second duration; and

if it is determined that the fitting degree is equal to or higher than a threshold degree, determining the trained linear regression model as an available prediction model.

16 . The electronic device according to claim 15 , wherein the actions further comprise:

if it is determined that the fitting degree is less than the threshold degree, continuing training the linear regression model by using parameter values of the deleted snapshot and a duration of the corresponding deletion operation.

17 . The electronic device according to claim 11 , wherein the instructions, when executed by the at least one processor, further cause the electronic device to perform actions including:

receiving a request for deleting a target snapshot of a data object in the storage system, the request comprising an identification of the target snapshot;

acquiring a target set of parameter values related to the target snapshot based on the identification; and

determining, based on the target set of parameter values, a duration required for deleting the target snapshot.

18 . The electronic device according to claim 17 , wherein determining a duration required for deleting the target snapshot comprises:

determining the duration by applying the target set of parameter values to the prediction model.

19 . The electronic device according to claim 17 , wherein the target set of parameter values comprises at least one of the following: central processing unit (CPU) usage, the number of snapshots being deleted when the target snapshot is deleted, a storage duration of the target snapshot, and a size of the data object when the target snapshot is created.