Dynamic optimization of backup policy
Embodiments of the present disclosure provide a method of backup management, an electronic device and a computer program product. The method comprises: determining a plurality of candidate backup policies for a plurality of clients of a data backup system, determining an expected load balance degree with respect to time for the data backup system to perform data backups for the plurality of clients using the plurality of candidate backup policies, determining an actual load balance degree with respect to time for the data backup system while the data backup system is performing the data backups for the plurality of clients using a plurality of current backup policies, and selecting a plurality of backup policies to be used for the plurality of clients respectively, based on a comparison of the expected load balance degree and the actual load balance degree.
1. A method of backup management, comprising:
determining a plurality of candidate backup policies for a plurality of clients of a data backup system;
determining an expected load balance degree for the data backup system to perform data backups for the plurality of clients using the plurality of candidate backup policies, wherein determining the expected load balance degree comprises:
simulating a plurality of concurrent backup jobs of the data backup system at a plurality of time points during use of the plurality of candidate backup policies; and
determining an expected standard deviation of the plurality of concurrent backup jobs, wherein the expected load balance degree is based on the expected standard deviation;
determining an actual load balance degree for the data backup system while the data backup system is performing the data backups for the plurality of clients using a plurality of current backup policies; and
selecting, between the plurality of candidate backup polices and the plurality of current backup policies, a plurality of backup policies to be used for the plurality of clients, based on a comparison of the expected load balance degree and the actual load balance degree.
2. The method of claim 1 , wherein determining the plurality of candidate backup policies comprises:
for each of a plurality of available backup policies, determining a degree of suitability of the available backup policy for a client of the plurality of clients, to obtain a set of degrees of suitability of the plurality of available backup policies for the client; and
selecting the plurality of candidate backup policies based at least on the sets of degrees of suitability for the client.
3. The method of claim 2 , wherein determining the degree of suitability comprises:
for the client:
predicting a backup execution time length for the client based on backup execution historical data of the client; and
determining a matching degree between the backup execution time length and a length of a backup time window of an available backup policy of the plurality of available backup policies.
4. The method of claim 2 , wherein determining the degree of suitability comprises:
predicting a backup execution time length for the client based on backup execution historical data of the client;
determining a backup execution time period of using an available backup policy of the plurality of available backup policies for the client based on a backup start time indicated by the available backup policy and the backup execution time length for the client;
determining, from the plurality of available backup policies, a number of available backup policies with a backup start time within the backup execution time period; and
modify the degree of suitability based on the number.
5. The method of claim 2 , wherein selecting the plurality of candidate backup policies comprises:
obtaining the plurality of candidate backup policies using an augmenting path algorithm of a bipartite graph model.
6. The method of claim 1 , wherein determining the actual load balance degree comprises:
identifying a plurality of concurrent backup jobs of the data backup system at a plurality of historical time points during use of the plurality of current backup policies based on backup execution historical data of the data backup system; and
determining an actual standard deviation of the plurality of concurrent backup jobs, wherein the actual load balance degree is based on the actual standard deviation.
7. An electronic device, comprising:
at least one processor; and
at least one memory storing computer program instructions, the at least one memory and the computer program instructions being configured, with the at least one processor, to cause the electronic device to:
determine a plurality of candidate backup policies for a plurality of clients of a data backup system;
determine an expected load balance degree for the data backup system to perform data backups for the plurality of clients using the plurality of candidate backup policies, wherein determining the expected load balance degree comprises:
simulating a plurality of concurrent backup jobs of the data backup system at a plurality of time points during use of the plurality of candidate backup policies; and
determining an expected standard deviation of the plurality of concurrent backup jobs, wherein the expected load balance degree is based on the expected standard deviation;
determine an actual load balance degree for the data backup system while the data backup system is performing the data backups for the plurality of clients using a plurality of current backup policies; and
select, between the plurality of candidate backup polices and the plurality of current backup policies, a plurality of backup policies to be used for the plurality of clients, based on a comparison of the expected load balance degree and the actual load balance degree.
8. The electronic device of claim 7 , wherein the at least one memory and the computer program instructions are configured, with the at least one processor, to cause the electronic device to determine the plurality of candidate backup policies by:
for each of a plurality of available backup policies, determining a degree of suitability of the available backup policy for a client of the plurality of clients, to obtain a set of degrees of suitability of the plurality of available backup policies for the client; and
selecting the plurality of candidate backup policies based at least on the sets of degrees of suitability for the client.
9. The electronic device of claim 8 , wherein the at least one memory and the computer program instructions are configured, with the at least one processor, to cause the electronic device to determine the degree of suitability by:
for the client:
predicting a backup execution time length for the client based on backup execution historical data of the client; and
determining a matching degree between the backup execution time length and a length of a backup time window of an available backup policy of the plurality of available backup policies.
10. The electronic device of claim 8 , wherein the at least one memory and the computer program instructions are configured, with the at least one processor, to cause the electronic device to determine the degree of suitability by:
predicting a backup execution time length for the client based on backup execution historical data of the client;
determining a backup execution time period of using an available backup policy of the plurality of available backup policies for the client based on a backup start time indicated by the available backup policy and the backup execution time length for the client;
determining, from the plurality of available backup policies, a number of available backup policies with a backup start time within the backup execution time period; and
modify the degree of suitability to be negatively correlated with the number.
11. The electronic device of claim 8 , wherein the at least one memory and the computer program instructions are configured, with the at least one processor, to cause the electronic device to select the plurality of candidate backup policies by:
obtaining the plurality of candidate backup policies using an augmenting path algorithm of a bipartite graph model.
12. The electronic device of claim 7 , wherein the at least one memory and the computer program instructions are configured, with the at least one processor, to cause the electronic device to determine the actual load balance degree by:
identifying a plurality of concurrent backup jobs of the data backup system at a plurality of historical time points during use of the plurality of current backup policies based on backup execution historical data of the data backup system; and
determining an actual standard deviation of the plurality of concurrent backup jobs, wherein the actual load balance degree is based on the actual standard deviation.
13. A non-transitory computer readable medium comprising machine-executable instructions, which when executed by a computer processor, enables the computer processor to perform a method, the method comprising:
determining a plurality of candidate backup policies for a plurality of clients of a data backup system;
determining an expected load balance degree for the data backup system to perform data backups for the plurality of clients using the plurality of candidate backup policies, wherein determining the expected load balance degree comprises:
simulating a plurality of concurrent backup jobs of the data backup system at a plurality of time points during use of the plurality of candidate backup policies; and
determining an expected standard deviation of the plurality of concurrent backup jobs, wherein the expected load balance degree is based on the expected standard deviation;
determining an actual load balance degree for the data backup system while the data backup system is performing the data backups for the plurality of clients using a plurality of current backup policies; and
selecting, between the plurality of candidate backup policies and the plurality of current backup policies, a plurality of backup policies to be used for the plurality of clients based on a comparison of the expected load balance degree and the actual load balance degree.
14. The non-transitory computer readable medium of claim 13 , wherein determining the plurality of candidate backup policies comprises:
for each of a plurality of available backup policies, determining a degree of suitability of the available backup policy for a client of the plurality of clients, to obtain a set of degrees of suitability of the plurality of available backup policies for the client; and
selecting the plurality of candidate backup policies based at least on the sets of degrees of suitability for the client.
15. The non-transitory computer readable medium of claim 14 , wherein determining the degree of suitability comprises:
for the client:
predicting a backup execution time length for the client based on backup execution historical data of the client; and
determining a matching degree between the backup execution time length and a length of a backup time window of an available backup policy of the plurality of available backup policies.
16. The non-transitory computer readable medium of claim 14 , wherein determining the degree of suitability comprises:
predicting a backup execution time length for the client based on backup execution historical data of the client;
determining a backup execution time period of using an available backup policy of the plurality of available backup policies for the client based on a backup start time indicated by the available backup policy and the backup execution time length for the client;
determining, from the plurality of available backup policies, a number of available backup policies with a backup start time within the backup execution time period; and
modify the degree of suitability to be negatively correlated with the number.
17. The non-transitory computer readable medium of claim 14 , wherein selecting the plurality of candidate backup policies comprises:
obtaining the plurality of candidate backup policies using an augmenting path algorithm of a bipartite graph model.