Leveraging file-system metadata for direct to cloud object storage optimization

Patent metadata

Field	Value
Publication number	US-12541432-B2
Application number	US-202217952480-A
Country	US
Kind code	B2
Filing date	Sep 26, 2022
Priority date	Sep 26, 2022
Publication date	Feb 3, 2026
Grant date	Feb 3, 2026

Abstract

A method and system for leveraging file-system metadata for direct to cloud object storage optimization. Under cloud object storage architecture, any unstructured data may be managed and stored in the cloud as objects. Objects thus provide an elastic, scalable format through which unstructured data may be maintained for a variety of purposes, including those directed to data backup, archiving, and/or disaster recovery. Cloud object storage, however, tends to be costly—mainly stemming from factors, such as the frequency of read and write operations (also referred to as runtime costs) applied to, as well as the allocation of disk space (also referred to as storage costs) consumed by, any number of objects configured to store file backup copies. In addressing at least the aforementioned, embodiments disclosed herein optimize cloud object storage utilization through the intelligent grouping and object storing of data based on file attribute similarities.

First claim

What is claimed is:

1 . A method for optimizing object-based backups, the method comprising:
obtaining, from a host file-system, file-system metadata descriptive of files; associating, based on the file-system metadata, at least a subset of the files with a shared object group; and performing an object-based backup of the at least subset of the files using at least one shared object, wherein performing the object-based backup, comprises:
identifying files of the at least subset of the files exhibiting shared attributes, wherein the shared attributes include at least one of labeled as read-only, a similar creation timestamp, a similar last modified timestamp, labeled as hidden, or labeled as archived;
partitioning, based on the files of the at least subset of files exhibiting shared attributes, the shared object group into shared object subgroups,
wherein each resulting shared object subgroup comprises a mutually-exclusive portion of the shared object group,
wherein a shared object subgroup of the shared object subgroups comprises a portion of files of the at least subset of the files exhibiting one or more of the shared attributes, and
wherein the partitioning is based on the one or more shared attributes;

assembling the portion of files into at least one shared object subgroup chunk;
generating the at least one shared object comprising the at least one shared object subgroup chunk;
writing the at least one shared object to a backup object storage;
receiving a request to perform an action on a file of the portion of files in the at least one shared object subgroup chunk, wherein the action comprises making a modification to the file of the portion of files; and
applying, based on the request, the action to all files in the at least one shared object subgroup chunk, wherein applying the action comprises making the same modification to all the files in the at least one shared object subgroup chunk.

2 . The method of claim 1 , wherein an object size of each shared object of the at least one shared object does not exceed an optimal object size.

3 . The method of claim 2 , wherein a file size of each file associated with the shared object group is less than a threshold file size, and wherein the threshold file size matches the optimal object size.

4 . The method of claim 1 , wherein the at least subset of the files are partitioned into the shared object subgroups based on file attribute similarities there-between.

5 . The method of claim 1 , the method further comprising:
associating, based on the file-system metadata, at least a second subset of the files with a dedicated object group; and performing a second object-based backup of the at least second subset of the files using at least one dedicated object.

6 . The method of claim 5 , wherein performing the second object-based backup, comprises:
for each file in the at least second subset of the files:
partitioning the file into at least one file chunk;
generating, respectively, the at least one dedicated object comprising the at least one file chunk; and
writing the at least one dedicated object to a backup object storage.

7 . The method of claim 6 , wherein an object size of each dedicated object of the at least one dedicated object does not exceed an optimal object size.

8 . The method of claim 7 , wherein a file size of each file associated with the dedicated object group at least meets a threshold file size, and wherein the threshold file size matches the optimal object size.

9 . A non-transitory computer readable medium (CRM) comprising computer readable program code, which when executed by a computer processor, enables the computer processor to perform a method for optimizing object-based backups, the method comprising:
obtaining, from a host file-system, file-system metadata descriptive of files; associating, based on the file-system metadata, at least a subset of the files with a shared object group; and performing an object-based backup of the at least subset of the files using at least one shared object, wherein performing the object-based backup, comprises:
identifying files of the at least subset of the files exhibiting shared attributes, wherein the shared attributes include at least one of labeled as read-only, a similar creation timestamp, a similar last modified timestamp, labeled as hidden, or labeled as archived;
partitioning, based on the files of the at least subset of files exhibiting shared attributes, the shared object group into shared object subgroups,
wherein each resulting shared object subgroup includes a mutually-exclusive portion of the shared object group,
wherein a shared object subgroup of the shared object subgroups comprises a portion of files of the at least subset of the files exhibiting one or more of the shared attributes, and

wherein the partitioning is based on the one or more shared attributes;
assembling the portion of files into at least one shared object subgroup chunk;
generating the at least one shared object comprising the at least one shared object subgroup chunk;
writing the at least one shared object to a backup object storage; receiving a request to perform an action on a file of the portion of files in the at least one shared object subgroup chunk, wherein the action comprises making a modification to the file of the portion of files; and applying, based on the request, the action to all files in the at least one shared object subgroup chunk, wherein applying the action comprises making the same modification to all the files in the at least one shared object subgroup chunk.

10 . The non-transitory CRM of claim 9 , wherein an object size of each shared object of the at least one shared object does not exceed an optimal object size.

11 . The non-transitory CRM of claim 10 , wherein a file size of each file associated with the shared object group is less than a threshold file size, and wherein the threshold file size matches the optimal object size.

12 . The non-transitory CRM of claim 9 , wherein the at least subset of the files are partitioned into the shared object subgroups based on file attribute similarities there-between.

13 . The non-transitory CRM of claim 9 , the method further comprising:
associating, based on the file-system metadata, at least a second subset of the files with a dedicated object group; and performing a second object-based backup of the at least second subset of the files using at least one dedicated object.

14 . The non-transitory CRM of claim 13 , wherein performing the second object-based backup, comprises:
for each file in the at least second subset of the files:
partitioning the file into at least one file chunk;
generating, respectively, the at least one dedicated object comprising the at least one file chunk; and
writing the at least one dedicated object to a backup object storage.

15 . The non-transitory CRM of claim 14 , wherein an object size of each dedicated object of the at least one dedicated object does not exceed an optimal object size.

16 . The non-transitory CRM of claim 15 , wherein a file size of each file associated with the dedicated object group at least meets a threshold file size, and wherein the threshold file size matches the optimal object size.

17 . A system, the system comprising:
a host device, comprising:
a computer processor;
a host operating system executing on the computer processor and configured to implement a host file-system for managing files; and
a host protection agent also executing on the computer processor and hav

Assignees

• Dell Products Lp

Inventors

• Yadav Sunil
• Chopra Shelesh

Classifications

• G06F2201/84

  Using snapshots, i.e. a logical point-in-time copy of the data · CPC title

• G06F16/164

  File meta data generation · CPC title

• G06F11/1451 Primary

  by selection of backup contents · CPC title