STORAGE SYSTEM FOR ELIMINATING DUPLICATED DATA

• US 20110191305A1
• Filed: 09/18/2009
• Published: 08/04/2011
• Est. Priority Date: 09/18/2009
• Status: Active Grant

• Current Assignee: Hitachi, Ltd.
• Original Assignee: Hitachi, Ltd.
• Inventors: Yamamoto, Akira, Iwasaki, Masaaki, Nakamura, Takaki, Takaoka, Nobumitsu, Ishii, Yohsuke
• Granted Patent: US 8,285,690 B2
• Time in Patent Office: Days
• US Class Current: 707/692
• CPC Class Codes: G06F 16/1748 De-duplication implemented within the file system, e.g. based on file segments de-duplication techniques in storage systems for the management of data blocks G06F3/0641
• CPC Class Codes: G06F 3/0608 Saving storage space on storage systems
• CPC Class Codes: G06F 3/0641 De-duplication techniques
• CPC Class Codes: G06F 3/0658 Controller construction arrangements
• CPC Class Codes: G06F 3/067 Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS]

Abstract

A storage system 103 carries out first and second de-duplication processes in response to receiving a write request from a client. First, a determination is made as to whether a write target data item overlaps with any of the stored data items of a part of a stored data item group, which is a user data item group stored in a storage device 209, and if so, the write target data item is prevented from being stored in the storage device. Second, a determination is made as to whether a target stored data item, which is not finished being evaluated as to whether it overlaps with the stored data item in the first de-duplication process, overlaps with another stored data item, and if so, the target stored data item or the same data item overlapping with the target stored data item is deleted from the storage device 209.

Claims

1. A storage system, comprising:

a communication interface device (207) that is coupled to a communication network (102) to which a client (101) is coupled;

a storage device (209) in which user data written/read to/from the client is stored; and

a controller (202) that is coupled to the communication interface and the storage device,

wherein when receiving a write request from the client, the controller carries out a first de-duplication process and thereafter carries out a second de-duplication process at an appropriate time,

in the first de-duplication process, the controller determines whether a write target data item, which is a write target user data item, overlaps with any of stored data items of a part of a stored data item group, which is a user data item group stored in a storage device, and does not store the write target data item in the storage device when a result of the determination is positive, and

in the second de-duplication process, the controller determines whether a target stored data item, which is not finished with an evaluation of whether or not it overlaps with the stored data item in the first de-duplication process, overlaps with another stored data item, and deletes the target stored data item or the same data item overlapping with the target stored data item from the storage device, when a result of the determination is positive.

2. The storage system according to claim 1, further comprising:

a memory (204) that has stored therein first process information indicating comparison candidates that are data items, out of the stored data items, that are likely to be the target of comparison in the first de-duplication process, wherein

stored data information indicating information related to the stored data items is stored in the storage device or the memory,

wherein the controller;

(A) carries out the following processes (A1) to (A4) in the first de-duplication process;

(A1) comparing a write target data item specified by the write request with a stored data item included in the comparison candidates, and carrying out a first duplication determination to determine whether the write target data item is a duplicated data item overlapping with any of the stored data items, a non-duplicated data item that does not overlap with any of the stored data items, or a duplication unknown data item that cannot be determined as the duplicated data item or non-duplicated data item;

(A2) carrying out a process for associating the write target data item with the same data overlapping with the write target data item, without storing the write target data item in the storage device, when the write target data item is determined as the duplicated data item in the first duplication determination;

(A3) storing the write target data item in the storage device, and storing, in the stored data information, the completion of evaluation of whether the write target data item is the duplicated data item or not, when the write target data item is determined as the non-duplicated data item in the first duplication determination; and

(A4) storing the write target data item in the storage device, and storing, in the stored data information, the fact that evaluation of whether the write target data item is the duplicated data item or not is not completed, when the write target data item is determined as the duplication unknown data in the first duplication determination; and

(B) carries out the following processes (B1) to (B3) in the second de-duplication process;

(B1) comparing a target stored data item, which is not finished with an evaluation of whether the data item is the duplicated data item or not, with another stored data item, and carrying out a second duplication determination to determine whether the target stored data item is a duplicated data item overlapping with any of the other stored data items, or a non-duplicated data item that does not overlap with any of the other stored data items;

(B2) deleting the target stored data item from the storage device, and carrying out a process for associating the target stored data item with the same data overlapping with the target stored data item, when the target stored data item is determined as the duplicated data item in the second duplication determination; and

(B3) storing, in the stored data information, the completion of evaluation of whether the target stored data item is the duplicated data item or not, when the target stored data item is determined as the non-duplicated data item in the second duplication determination.

3. The storage system according to claim 2, wherein

rank information that shows a rank applied to each of the stored data items for each generation is stored in the memory,

the rank applied to each of the stored data items is updated on a regular or irregular basis, and the generation of the rank is shifted to a next generation when the rank is updated, and

the controller carries out the following:

(1) in the first de-duplication process;

(1-1) determining that the write target data item is the duplicated data item, when the write target data item conforms with any of the stored data items included in the comparison candidates or when feature data related to the write target data item conforms with feature data related to any of the stored data items included in the comparison candidates;

(1-2) selecting, from the stored data items included in the comparison candidates, stored data items to be compared with the write target data item, in descending order of the ranks applied to the stored data items, and ending the first duplication determination when the write target data item is determined as the duplicated data item; and

(1-3) storing, in the storage device or the memory, comparison completion information indicating up to which rank of stored data item in a certain generation the comparison is completed in the first de-duplication process performed on the write target data item, when the write target data item is determined as the duplication unknown data;

(2) calculating a de-duplication ratio, which is a value indicating to what degree duplication storage is eliminated, and carrying out the second de-duplication process when a current de-duplication ratio is equal to or lower than a target value; and

(3) in the second de-duplication process;

(3-1) determining that the target stored data item is the duplicated data item, when the target stored data item conforms with any of the other stored data items or when feature data related to the target stored data item conforms with feature data related to any of the other stored data items;

(3-2) selecting, from the other stored data items, stored data items to be compared with the target stored data item, in descending order of the ranks applied to the stored data items, and ending the second duplication determination when the target stored data item is determined as the duplicated data item; and

(3-3) referring to the comparison completion information to compare the target stored data item with a stored data item of the other stored data items, which is not finished with the comparison.

4. The storage system according to claim 1, wherein the controller calculates a de-duplication ratio, which is a value indicating to what degree duplication storage is eliminated, and carries out the second de-duplication process when a current de-duplication ratio is equal to or lower than a target value.

5. The storage system according to claim 2, wherein

stored data presence/absence information which indicates, for each hash value within a hash space, whether there exists a stored data item to be the hash value when a hash value is generated, is included in the first process information,

the controller carries out the following processes (a1) to (a3) in the first duplication determination:

(a1) generating a hash value related to the write target data item;

(a2) referring to the stored data presence/absence information to determine whether there exists the stored data item to be the generated hash value; and

(a3) determining that the write target data item is the non-duplicated data item, when the stored data item to be the generated hash value does not exist.

6. The storage system according to claim 5, wherein the hash value is a hash value that is generated from the entire user data or a user data main body or a hash value that is generated from data added to a user data main body.

7. The storage system according to claim 2, wherein

rank information, which indicates a rank applied to each of the stored data items and referenced when determining an order of comparison, is stored in the memory, and

the controller:

selects, from the stored data items included in the comparison candidates, stored data items to be compared with the write target data item, in descending order of the ranks applied to the stored data items, in the first duplication determination, and ends the first duplication determination when the write target data item is determined as the duplicated data item; and

selects, from the other stored data items, stored data items to be compared with the target stored data item, in descending order of the ranks applied to the stored data items, in the second duplication determination, and ends the second duplication determination when the target stored data item is determined as the duplicated data item.

8. The storage system according to claim 3, wherein the controller applies a higher rank to a stored data item that is expected to be likely to overlap, or applies a higher rank to a data item for which the effect of reducing usage of the storage device is great.

9. The storage system according to claim 7, wherein

the rank is managed for each generation in the rank information,

the rank applied to each of the stored data items is updated on a regular or irregular basis, and the generation of the rank is shifted to a next generation when the rank is updated, and

the controller:

stores, in the storage device of the memory, comparison completion information indicating up to which rank of stored data item in a certain generation the comparison is completed in the first de-duplication process performed on the write target data item, when the write target data item is determined as the duplication unknown data in the first duplication determination; and

refers to the comparison completion information in the second de-duplication process to compare the target stored data item with a stored data item of the other stored data items, which is not finished with the comparison.

10. The storage system according to claim 2, wherein the controller:

determines that the write target data item is the duplicated data item, when, in the first duplication determination, the write target data item conforms with any of the stored data items included in the comparison candidates or when feature data related to the write target data item conforms with feature data related to any of the stored data items included in the comparison candidates; and

determines that the target stored data item is the duplicated data item, when, in the second duplication determination, the target stored data item conforms with any of the other stored data items or when feature data related to the target stored data item conforms with feature data related to any of the other stored data items.

11. The storage system according to claim 2, further comprising:

a plurality of storage servers (201),

wherein each of the storage servers has the communication interface device, the storage device, the memory, and the controller,

the memory further has stored therein server management information that indicates a correspondence relationship between a data hash space, which is a hash space related to user data, and the storage server managing the user data included in the data hash space,

the controller of a request receiving server, which is the storage server that receives a write request from the client;

(X1) generates a hash value related to a write target data item specified by the write request;

(X2) refers to the server management information to determine a data server, which is the storage server that manages the write target data item; and

(X3) transmits the write target data item to the data server,

and wherein the controller of the data server carries out the first de-duplication process when receiving the write target data item from the request receiving server, and thereafter carries out the second de-duplication process at an appropriate time.

12. The storage system according to claim 11, wherein, when storing the write target data item in the storage device, the controller of the data server:

(Y1) determines whether the storage device of the data server has a sufficient free space;

(Y2) stores the write target data item in the storage device of the data server when the storage device of the data server has a sufficient free space; and

(Y3) stores the write target data item in the storage device of the storage server other than the data server, when the storage device of the data server does not have a sufficient free space.

13. The storage system according to claim 11, wherein

the server management information includes information that indicates a correspondence relationship between a name hash space, which is a hash space related to a global path of user data, and the storage server managing a path name of the global path included in the name hash space, and

the controller of the request receiving server, which is the storage server that receives a write request from the client:

(Z1) generates a hash value related to a write target global path, which is a global path of the write target data item;

(Z2) refers to the server management information to determine a name server, which is the storage server that manages the write target global path; and

(Z3) transmits the path name of the write target global path to the name server,

and wherein the controller of the name server stores the path name of the write target global path into the storage device of the name server when receiving the path name of the write target global path from the request receiving server.