System and method for approximating replication completion time
One embodiment provides a computer implemented method of estimating replication completion time. The method includes creating a historical dataset of prior replication data; determining a set of replication parameters to consider; inputting the historical dataset and the set of replication parameters to a replication completion time estimator module; generating a replication completion time prediction based on the historical dataset and the set of replication parameters; and generating a confidence prediction corresponding to the replication completion time prediction.
1. A computer implemented method of estimating replication completion time, the method comprising:
creating a historical dataset of prior replication data;
determining a set of replication parameters to consider based on input by a user;
inputting the historical dataset and the set of replication parameters to a replication completion time estimator module;
generating a plurality of replication completion time predictions based on the historical dataset and the set of replication parameters, each replication completion time prediction corresponding to a particular potential replication destination and a particular replication type based on the input by the user;
generating a confidence prediction corresponding to each replication completion time prediction, the confidence prediction comprising a training set and a calibration set of disjoint training data for a regression model, wherein residuals of the regression model is used to quantify uncertainty in future perditions of the confidence prediction;
executing a resulting particular replication type at a resulting particular replication destination based on the confidence prediction corresponding to the replication completion time predictions for each potential replication destination and replication type; and
when executing the resulting particular replication type at the resulting particular replication destination, notifying a completion message of the execution of the resulting particular replication type at the resulting particular replication destination to the user.
2. The computer implemented method as in claim 1 , wherein the confidence prediction includes a lower boundary and an upper boundary corresponding to the replication completion time prediction.
3. The computer implemented method as in claim 2 , wherein the confidence prediction includes a percentage calculation that an actual completion time will fall within the lower boundary and the upper boundary.
4. The computer implemented method as in claim 2 , further comprising:
rejecting a replication completion time prediction if it falls outside the lower boundary or the upper boundary.
5. The computer implemented method as in claim 1 , further comprising:
generating and displaying a graphical representation of the replication completion time prediction and the confidence prediction using a graphical user interface.
6. A system for estimating replication completion time comprising:
a storage system management console;
a historical database configured to store a historical dataset of prior replication data; and
a replication completion time estimator module configured to:
receive the historical dataset and a set of replication parameters based on input by a user;
generate a plurality of replication completion time predictions based on the historical dataset and the set of replication parameters, each replication completion time prediction corresponding to a particular potential replication destination and a particular replication type based on the input by the user;
generate a confidence prediction corresponding to each replication completion time prediction, the confidence prediction comprising a training set and a calibration set of disjoint training data for a regression model, wherein residuals of the regression model is used to quantify uncertainty in future perditions of the confidence prediction;
execute a resulting particular replication type at a resulting particular replication destination based on the confidence prediction corresponding to the replication completion time predictions for each potential replication destination and replication type; and
when executing the resulting particular replication type at the resulting particular replication destination, notify a completion message to the user.
7. The system as in claim 6 , wherein the confidence prediction includes a lower boundary and an upper boundary corresponding to the replication completion time prediction.
8. The system as in claim 7 , wherein the confidence prediction includes a percentage calculation that an actual completion time will fall within the lower boundary and the upper boundary.
9. The system as in claim 7 , wherein the replication completion time estimator module is further configured to:
reject a replication completion time prediction if it falls outside the lower boundary or the upper boundary.
10. The system as in claim 6 , further comprising:
a graphical user interface configured to generate and display a graphical representation of the replication completion time prediction and the confidence prediction.
11. A non-transitory computer-readable medium storing instructions which, when executed by one or more processors, cause the one or more processors to perform a method of estimating replication completion time, the method comprising:
creating a historical dataset of prior replication data;
determining a set of replication parameters to consider based on input by a user;
inputting the historical dataset and the set of replication parameters to a replication completion time estimator module;
generating a plurality of replication completion time predictions based on the historical dataset and the set of replication parameters, each replication completion time prediction corresponding to a particular potential replication destination and a particular replication type based on the input by the user;
generating a confidence prediction corresponding to each replication completion time prediction, the confidence prediction comprising a training set and a calibration set of disjoint training data for a regression model, wherein residuals of the regression model is used to quantify uncertainty in future perditions of the confidence prediction;
executing a resulting particular replication type at a resulting particular replication destination based on the confidence prediction corresponding to the replication completion time predictions for each potential replication destination and replication type; and
when executing the resulting particular replication type at the resulting particular replication destination, notifying a completion message to the user.
12. The non-transitory computer-readable medium as in claim 11 , wherein the confidence prediction includes a lower boundary and an upper boundary corresponding to the replication completion time prediction.
13. The method as in claim 1 , wherein generating the plurality of replication completion time predictions includes using a conformal quantile regressor.
14. The method as in claim 5 , wherein executing the particular replication type at a particular replication destination includes receiving a selection via the graphical user interface of the particular replication type and replication destination from user.
15. The system as in claim 6 , wherein generating the plurality of replication completion time predictions includes using a conformal quantile regressor.
16. The system as in claim 10 , wherein executing the particular replication type at a particular replication destination includes receiving a selection via the graphical user interface of the particular replication type and replication destination from the user.
17. The non-transitory computer-readable medium as in claim 11 , wherein the method further comprises:
generating and displaying a graphical representation of the replication completion time prediction and the confidence prediction via a graphical user interface; and
receiving a selection via the graphical user interface of the particular replication type and replication destination from the user.
18. The non-transitory computer-readable medium as in claim 11 , wherein generating the plurality of replication completion time predictions includes using a conformal quantile regressor.