Fine-grained shared multi-tenant de-duplication system

• US 10,380,098 B1
• Filed: 09/30/2015
• Issued: 08/13/2019
• Est. Priority Date: 09/30/2015
• Status: Active Grant

• Current Assignee: Emc IP Holding Company LLC (Dell Technologies Inc.)
• Original Assignee: Emc IP Holding Company LLC (Dell Technologies Inc.)
• Inventors: Marelas, Peter
• Primary Examiner(s): Pham, Michael
• Application Number:
  US14/870,407
• Time in Patent Office: 1,413 Days
• Field of Search: 707692
• CPC Class Codes: G06F 11/1453 using de-duplication of the data
• CPC Class Codes: G06F 11/1456 Hardware arrangements for backup
• CPC Class Codes: G06F 16/2255 Hash tables
• CPC Class Codes: G06F 16/2365 Ensuring data consistency and integrity
• CPC Class Codes: G06F 16/285 Clustering or classification

Abstract

In one example, a method for managing data includes detecting a data stream that is specific to a first user and that includes one or more user objects each having one or more data chunks. Next, the data associated with the first user is salted by associating a user-specific salt with the data chunks of the one or more user objects to form user-specific combinations of the form ((user-specific salt)+(user data chunk)). Finally, an amount of storage capacity consumed by the one or more user objects is determined.

Claims

1. A method, comprising:

detecting a data stream that is generated in part by one or more applications and is specific to a first user and that includes a first user object having one or more data chunks;

salting data associated with the first user, wherein salting data comprises;

associating a user-specific salt with the data chunks of the first user object to form user-specific combinations of the form ((user-specific salt)+(first user data chunk)); and

hashing the ((user-specific salt)+(first user data chunk)) combinations to form first user-specific chunk hashes;

storing the user-specific chunk hashes in a container;

reducing an amount of user-specific data to be stored by performing user-level deduplication in a shared multi-tenant deduplication system to eliminate duplicate user-specific data, wherein performing user-level deduplication comprises de-duplicating only data associated with the first user by performing the following;

comparing a first user-specific chunk hash with a second user-specific chunk hash that is associated with the same user as the first user-specific chunk hash;

when the second user-specific chunk hash is the same as the first user-specific chunk hash, discarding the second user-specific chunk hash;

based on results of the deduplication process, determining a storage capacity consumed by objects of the first user; and

storing the user-specific data remaining after deduplication has been performed.

2. The method as recited in claim 1, wherein the container is tiered to a cloud store after the container is full, or after the user-specific data stream has ended.

3. The method as recited in claim 1, wherein associating a user-specific salt with the first user-specific chunk comprises either prepending the first user-specific chunk with the user-specific salt, or appending the user-specific salt to the first user-specific chunk.

4. The method as recited in claim 1, wherein the user-specific salt persists so long as the first user exists.

5. The method as recited in claim 1, wherein when the detected data stream is associated with a new user, the user-specific salt is created after detection of the data stream.

6. The method as recited in claim 1, further comprising:

detecting another data stream that is specific to a second user that is different from the first user;

salting data of the second data stream; and

deduplicating data of the second data stream.

7. The method as recited in claim 6, wherein salts of the respective first and second users ensure that identical blocks stored by each user will not share the same persistent storage address space.

8. The method as recited in claim 1, further comprising recording the user-specific salt against a namespace of the first user object, and against the first user-specific chunk hashes.

9. The method as recited in claim 1, further comprising:

removing the user-specific salt from the data chunks of the first user object; and

re-computing a hash of the data chunks without the user-specific salt.

10. A non-transitory storage medium having stored therein computer-executable instructions which, when executed by one or more hardware processors, perform the following operations:

detecting a data stream that is generated by one or more applications and is specific to a first user and that includes a first user object having one or more data chunks;

salting data associated with the first user, wherein salting data comprises;

associating a user-specific salt with the data chunks of the first user object to form user-specific combinations of the form ((user-specific salt)+(first user data chunk)); and

hashing the ((user-specific salt)+(first user data chunk)) combinations to form first user-specific chunk hashes;

storing the user-specific chunk hashes in a container;

reducing an amount of user-specific data to be stored by performing user-level deduplication in a shared multi-tenant deduplication system to eliminate duplicate user-specific data, wherein performing user-level deduplication comprises de-duplicating only data associated with the first user by performing the following;

comparing a first user-specific chunk hash with a second user-specific chunk hash that is associated with the same user as the first user-specific chunk hash;

when the second user-specific chunk hash is the same as the first user-specific chunk hash, discarding the second user-specific chunk hash;

based on results of the deduplication process, determining a storage capacity consumed by objects of the first user; and

storing the user-specific data remaining after deduplication has been performed.

11. A non-transitory storage medium having stored therein computer-executable instructions which, when executed by one or more hardware processors, perform the following operations:

detecting a data stream that is generated in part by one or more applications and is specific to a first user and that includes one or more user objects each having one or more data chunks;

salting data associated with the first user, wherein salting data comprises associating a user-specific salt with the data chunks of the one or more user objects to form user-specific combinations of the form ((user-specific salt)+(user data chunk));

using the salted data as a basis for accurately determining, on a user basis, an amount of storage capacity consumed by the one or more user objects; and

using information about the consumed storage capacity to facilitate performance of another process relating to the consumption of data storage.

12. The non-transitory storage medium as recited in claim 11, wherein determining an amount of storage capacity consumed by the one or more user objects comprises keeping a running total of data chunk sizes associated with the user-specific salt.

13. The non-transitory storage medium as recited in claim 12, wherein keeping a running total of data chunk sizes associated with the user-specific salt comprises:

incrementing a salt-dependent counter during intake of data chunks associated with the user-specific salt; and

decrementing the salt-dependent counter during garbage collection of data chunks associated with the user-specific salt.

14. The non-transitory storage medium as recited in claim 11, wherein determining an amount of storage capacity consumed by the one or more user objects comprises:

traversing an object namespace;

identifying objects associated with the user-specific salt; and

totaling live data chunk sizes for all of the objects.

15. The non-transitory storage medium as recited in claim 11, wherein determining an amount of storage capacity consumed by the one or more user objects comprises:

traversing the data chunks; and

totaling live data chunk sizes associated with the user-specific salt.

16. The non-transitory storage medium as recited in claim 11, wherein the user objects comprise deduplicated data.

17. The non-transitory storage medium as recited in claim 11, wherein the operations further comprise reporting the amount of storage capacity consumed by the one or more user objects.