Systems and methods for prioritizing file downloads

• US 9,824,094 B1
• Filed: 04/24/2014
• Issued: 11/21/2017
• Est. Priority Date: 04/24/2014
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Meyer, Cayden, Kilpatrick, Ian
• Primary Examiner(s): Singh, Amresh
• Application Number:
  US14/260,519
• Time in Patent Office: 1,307 Days
• Field of Search: None
• CPC Class Codes: G06F 16/182 Distributed file systems
• CPC Class Codes: H04L 67/06 specially adapted for file transfer, e.g. file transfer protocol [FTP]
• CPC Class Codes: H04L 67/1097 for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]

Abstract

Systems and methods are disclosed herein for downloading data from a cloud system. A plurality of files on the cloud system is identified for downloading to a client system, where the plurality of files is associated with metadata. A respective score for each file in the plurality of files is evaluated by applying a ranking scheme to the metadata, where the ranking scheme is based on at least two features of the metadata. A download process is initiated for at least some files in the plurality of files based on each file'"'"'s respective score.

Claims

1. A method for determining an order of files for downloading from a cloud system, the method comprising:

identifying a first plurality of files on the cloud system for downloading to a client system, wherein the first plurality of files is associated with first metadata;

evaluating a first respective score for each file in the first plurality of files by applying a first ranking scheme to the first metadata to generate a first ranking of the first plurality of files, wherein the first ranking scheme is based on weights of at least two features of the first metadata;

initiating a first download process for at least some files in the first plurality of files based on the first ranking of the first plurality of files;

receiving, during the first download process and from a user, a user selection of a first file from the first plurality of files, wherein the user selection prioritizes the first file to be downloaded immediately regardless of a ranking of the first file in the first ranking;

instantiating a ranker training procedure to refine the first ranking scheme based on training data to generate a second ranking scheme, wherein the training data comprises a downloading priority of the first file;

evaluating a second respective score for each file in a second plurality of files by applying the second ranking scheme to second metadata associated with the second plurality of files to generate a second ranking of the second plurality of files, wherein the ranker training procedure generates a predicted access frequency of the first file and compares the predicted access frequency to an actual access frequency of the first file, and wherein the second ranking scheme is generated based at least in part on the comparison; and

initiating a second download process for each file in the second plurality of files based on the second ranking of the second plurality of files.

2. The method of claim 1, further comprising selecting the at least some files in the first plurality of files based on the first respective score of each file in the first plurality of files and an amount of storage availability on the client system.

3. The method of claim 1, further comprising downloading the first plurality of files to the client system in an order defined by the first respective score of each file in the first plurality of files.

4. The method of claim 1, wherein the first respective score for each respective file in the first plurality of files is indicative of a likelihood that the user will access the respective file on the client system.

5. The method of claim 1, wherein the first ranking scheme is trained using a machine learning technique on existing cloud data in the cloud system.

6. The method of claim 5, further comprising separating the existing cloud data into at least two groups of cloud data, and applying the machine learning technique on the at least two groups of cloud data to train at least two ranking schemes.

7. The method of claim 1, wherein the first metadata comprises at least two of:

file size, file type, file history, time of last write, time of last access, time of initial creation, folder depth, or user specific data.

8. The method of claim 1, wherein the first metadata comprises a set of binary features that describe each file in the first plurality of files.

9. The method of claim 1, wherein the cloud system comprises a processor that applies the first ranking scheme to the first plurality of files.

10. The method of claim 1, wherein the cloud system comprises a processor that periodically updates the first ranking scheme based on updated cloud data.

11. A system for determining an order of files for downloading from a cloud system, the system comprising:

at least one memory that stores instructions; and

at least one processor, coupled to the at least one memory, configured to execute the instructions to;

identify a first plurality of files on the cloud system for downloading to a client system, wherein the first plurality of files is associated with first metadata;

evaluate a first respective score for each file in the first plurality of files by applying a first ranking scheme to the first metadata to generate a first ranking of the first plurality of files, wherein the first ranking scheme is based on weights of at least two features of the first metadata;

initiate a first download process for at least some files in the first plurality of files based on the first ranking of the first plurality of files;

receive, during the first download process and from a user, a user selection of a first file from the first plurality of files, wherein the user selection prioritizes the first file to be downloaded immediately regardless of a ranking of the first file in the first ranking;

instantiate a ranker training procedure to refine the first ranking scheme based on training data to generate a second ranking scheme, wherein the training data comprises a downloading priority of the first file;

evaluate a second respective score for each file in a second plurality of files by applying the second ranking scheme to second metadata associated with the second plurality of files to generate a second ranking of the second plurality of files, wherein the ranker training procedure generates a predicted access frequency of the first file and compares the predicted access frequency to an actual access frequency of the first file, and wherein the second ranking scheme is generated based at least in part on the comparison; and

initiate a second download process for each file in the second plurality of files based on the second ranking of the second plurality of files.

12. The system of claim 11, wherein the at least one processor is further configured to select the at least some files in the first plurality of files based on the first respective score of each file in the first plurality of files and an amount of storage availability on the client system.

13. The system of claim 11, wherein the at least one processor is further configured to download the first plurality of files to the client system in an order defined by the first respective score of each file in the first plurality of files.

14. The system of claim 11, wherein the first respective score for each respective file in the first plurality of files is indicative of a likelihood that the user will access the respective file on the client system.

15. The system of claim 11, wherein the first ranking scheme is trained using a machine learning technique on existing cloud data in the cloud system.

16. The system of claim 15, wherein the at least one processor is further configured to separate the existing cloud data into at least two groups of cloud data, and apply the machine learning technique on the at least two groups of cloud data to train at least two ranking schemes.

17. The system of claim 11, wherein the first metadata comprises at least two of:

file size, file type, file history, time of last write, time of last access, time of initial creation, folder depth, or user specific data.

18. The system of claim 11, wherein the first metadata comprises a set of binary features that describe each file in the first plurality of files.

19. The system of claim 11, wherein the at least one processor applies the first ranking scheme to the first plurality of files.

20. The system of claim 11, wherein the cloud system comprises the at least one processor that periodically updates the first ranking scheme based on updated cloud data.