DISTRIBUTED STORAGE OF DATA IN A CLOUD STORAGE SYSTEM

• US 20120259901A1
• Filed: 03/30/2012
• Published: 10/11/2012
• Est. Priority Date: 03/31/2011
• Status: Active Grant

• Current Assignee: KT Corporation
• Original Assignee: KT Corporation
• Inventors: Kim, Mi-Jeom, Kim, Hyo-Min, Park, Chang-Sik, Hwang, Jin-Kyung, LEE, Eo-Hyung
• Granted Patent: US 9,052,962 B2
• Time in Patent Office: Days
• US Class Current: 707/827
• CPC Class Codes: G06F 16/27 Replication, distribution or synchronisation of data between databases or within a distributed database system; Distributed database system architectures therefor
• CPC Class Codes: G06F 9/5072 Grid computing

Abstract

Provided are apparatus and method for distributedly storing an object in a plurality of data centers included in a cloud storage system. In the method, an object may be divided into a plurality of sub-objects each having a fragmentation size. Then, the sub-objects may be distributedly stored in at least one data center based on weights of the plurality of data centers.

Claims

1. A method for distributedly storing data in a plurality of data centers in a cloud storage system, the method comprising:

dividing an object into a plurality of sub-objects each having a fragmentation size;

distributedly storing the sub-objects in at least one data center based on weights of the plurality of data centers.

2. The method of claim 1, comprising determining each fragmentation size based dividing an average object size by a sum of the weights of the plurality of data centers.

3. The method of claim 1, comprising determining the weights based on storage resources of each of the plurality of data centers.

4. The method of claim 3, wherein the storage resources include storage capacities of the plurality of data centers.

5. The method of claim 1, wherein, for the distributedly storing, corresponding weight of a data center is based on a ratio of a storage capacity of the data center to storage capacities of other data centers.

6. The method of claim 1, wherein the distributedly storing the sub-object comprises determining, a number of sub-objects to be stored in each data center based on a weight of each data center.

7. The method of claim 1, wherein for the distributedly storing the sub-objects, more sub-objects are stored in a data center having a higher weight than a data center having a lower weight.

8. The method of claim 1, comprising storing and managing the weights of the data center in a form of metadata.

9. The method of claim 1, comprising receiving the object from a client with a request of storing the object in the cloud storage system.

10. The method of claim 1, further comprising:

calculating a first hash value of the object; and

assigning the first hash value to the sub-objects as a correlation value.

11. The method of claim 10, comprising storing and managing the hash value of the object and the correlation value of the sub-objects as metadata.

12. The method of claim 10, comprising determining, with the hash value of the object and the correlation value of the sub-objects, whether the object is related to the sub-objects and whether the sub-objects are related to each other.

13. The method of claim 1, further comprising:

calculating a second hash value of each sub-object based on a filename of each sub-objects; and

determining at least one data center to store sub-objects based on the second hash value of a corresponding sub-object.

14. The method of claim 13, comprising:

processing the second hash value of each sub-object into a reference number and indicating, based on the second hash value, at least one data center in metadata stored in a metadata database.

15. A method for retrieving an object in a cloud storage system including a plurality of data centers, the method comprising:

receiving, from a client, a request of retrieving an object from the cloud storage system;

determining data centers in the cloud storage system, which store sub-objects divided from the object;

requesting the determined data centers to transmit the sub-objects;

receiving the sub-objects from the determined data centers;

restoring the object by combining the received sub-objects; and

transmitting the restored object to the client.

16. The method of claim 15, the determining data centers comprises:

determining whether a sub-object is divided from the object based on a correlation value of the sub-object and a hash value of the object.

17. The method of claim 16, comprising determining sub-objects having a correlation value identical to a hash value of the object as the sub-objects divided from the object.

18. An apparatus for distributedly storing an object in a plurality of data centers included in a cloud storage system, the apparatus comprising:

a divider configured to divide an object into a plurality of sub-objects each having a fragmentation size; and

a distributor configured to distributedly store the plurality of sub-objects in at least one data center based on weights of the data centers,

wherein;

the fragmentation size is determined based on a result of dividing an average object size by a sum of the weights of the data centers; and

a weight of a data center is a ratio of a storage capacity of the data center to storage capacities of other data centers.

19. The apparatus of claim 18, further comprising:

a correlation unit configured to calculate a first hash value of the object and assign the first hash value to the sub-objects as a correlation value; and

a storage location decider configured to calculate a second hash value of each sub-object based on a filename of each sub-object and to determine at least one data center to store the sub-objects based on the second hash value of each sub-object.

20. The apparatus of claim 19, wherein:

the storage location decider is configured to i) obtain reference numbers that indicate the at least one data center in metadata stored in a metadata database by processing the second hash value of each sub-object and ii) determine data centers to store the sub-objects based on the reference numbers;

the distributor is configured to determine a number of sub-objects to be stored in the determined data centers based on weights of the determined data centers; and

the distributor is configured to store more sub-objects in a data center having a higher weight than a data center having a lower weight.