Method, device and computer program product for backuping data
Embodiments of the present disclosure provide a method, device and computer program product for backing up data. The method comprises obtaining a data attribute of specific data to be backed up from a client to a server, a resource utilization rate at the client, and a network condition between the client and the server. The method further comprises setting, based on the data attribute, the resource utilization rate and the network condition, a plurality of parameters for performing stream backup, wherein the plurality of parameters at least comprises a concurrent number of stream transmission and a concurrent number of data parsing. The method further comprises parsing, according to the set plurality of parameters, the specific data and backing up the specific data from the client to the server.
1. A method of backing up data, comprising:
obtaining a data attribute of specific data to be backed up from a client to a server, a resource utilization rate at the client, and a network condition between the client and the server;
setting, based on the data attribute, the resource utilization rate and the network condition, a plurality of parameters for performing stream backup, the plurality of parameters at least comprising a concurrent number of stream transmission and a concurrent number of data parsing, wherein setting the plurality of parameters comprises:
sending, to a trained decision tree model, the data attribute, the resource utilization rate and the network condition; and
receiving, from the trained decision tree model, the concurrent number of the data parsing, a size of a buffer area for storing parsed data, the concurrent number of the stream transmission, and a chunk size of the stream transmission; and
parsing, according to the plurality of parameters, the specific data and backing up the specific data from the client to the server.
2. The method of claim 1 , further comprising:
collecting historical data related to stream backup as a training data set, each record in the training data set at least comprising the data attribute, the resource utilization rate, the network condition and corresponding optimization parameters; and
training the decision tree model using the collected training data set to obtain the trained decision tree model.
3. The method of claim 2 , wherein training the decision tree model using the collected training data set comprises:
building a very fast decision tree (VFDT) model by using Hoeffding bound as a split condition.
4. The method of claim 3 , further comprising:
deploying the built VFDT model in a backup system;
collecting optimization parameters configured by a user as new training data during user utilization of the backup system; and
training the VFDT model using the new training data in an incremental way.
5. The method of claim 1 , wherein obtaining the data attribute and the resource utilization rate comprises:
obtaining a type of the specific data, a size of the specific data, a proportion of new bytes in the specific data, a utilization rate of a processing resource at the client, a utilization rate of a storage resource at the client, and a utilization rate of input/output (I/O) resources at the client.
6. The method of claim 5 , wherein obtaining the network condition comprises:
obtaining a duration of last backup, and a network bandwidth and a network latency between the client and the server.
7. The method of claim 5 , wherein the type of the specific data comprises any of a database, a virtual machine file and an email file.
8. An electronic device, comprising:
a processing unit; and
a memory coupled to the processing unit and storing instructions, the instructions, when executed by the processing unit, causing the device to perform a method, the method comprising:
obtaining a data attribute of specific data to be backed up from a client to a server, a resource utilization rate at the client, and a network condition between the client and the server;
setting, based on the data attribute, the resource utilization rate and the network condition, a plurality of parameters for performing stream backup, the plurality of parameters at least comprising a concurrent number of stream transmission and a concurrent number of data parsing, wherein setting the plurality of parameters comprises:
sending, to a trained decision tree model, the data attribute, the resource utilization rate and the network condition; and
receiving, from the trained decision tree model, the concurrent number of the data parsing, a size of a buffer area for storing parsed data, the concurrent number of the stream transmission, and a chunk size of the stream transmission; and
parsing, according to the plurality of parameters, the specific data and backing up the specific data from the client to the server.
9. The electronic device of claim 8 , wherein the acts further comprise:
collecting historical data related to stream backup as a training data set, each record in the training data set at least comprising the data attribute, the resource utilization rate, the network condition and corresponding optimization parameters; and
training the decision tree model using the collected training data set to obtain the trained decision tree model.
10. The electronic device of claim 9 , wherein training the decision tree model using the collected training data set comprises:
building a very fast decision tree (VFDT) model by using Hoeffding bound as a split condition.
11. The electronic device of claim 10 , wherein the method further comprises:
deploying the built VFDT model in a backup system;
collecting optimization parameters configured by a user as new training data during user utilization of the backup system; and
training the VFDT model using the new training data in an incremental way.
12. The electronic device of claim 8 , wherein obtaining the data attributes and the resource utilization rate comprises:
obtaining a type of the specific data, a size of the specific data, a proportion of new bytes in the specific data, a utilization rate of a processing resource at the client, a utilization rate of a storage resource at the client, and a utilization rate of input/output (I/O) resources at the client.
13. The electronic device of claim 12 , wherein obtaining the network condition comprises:
obtaining a duration of last backup, and a network bandwidth and a network latency between the client and the server.
14. The electronic device of claim 12 , wherein the type of the specific data comprises any of a database, a virtual machine file and an email file.
15. A computer program product being tangibly stored on a non-transitory computer-readable medium and comprising computer-executable instructions, which, when executed, causes a machine to perform a method, the method comprising:
obtaining a data attribute of specific data to be backed up from a client to a server, a resource utilization rate at the client, and a network condition between the client and the server;
setting, based on the data attribute, the resource utilization rate and the network condition, a plurality of parameters for performing stream backup, the plurality of parameters at least comprising a concurrent number of stream transmission and a concurrent number of data parsing, wherein setting the plurality of parameters comprises:
sending, to a trained decision tree model, the data attribute, the resource utilization rate and the network condition; and
receiving, from the trained decision tree model, the concurrent number of the data parsing, a size of a buffer area for storing parsed data, the concurrent number of the stream transmission, and a chunk size of the stream transmission; and
parsing, according to the plurality of parameters, the specific data and backing up the specific data from the client to the server.
16. The computer program product of claim 15 , wherein the method further comprises:
collecting historical data related to stream backup as a training data set, each record in the training data set at least comprising the data attribute, the resource utilization rate, the network condition and corresponding optimization parameters; and
training the decision tree model using the collected training data set to obtain the trained decision tree model.
17. The computer program product of claim 16 , wherein training the decision tree model using the collected training data set comprises:
building a very fast decision tree (VFDT) model by using Hoeffding bound as a split condition.