Selecting data nodes in distributed storage system

• US 8,849,756 B2
• Filed: 04/13/2012
• Issued: 09/30/2014
• Est. Priority Date: 04/13/2011
• Status: Active Grant

• Current Assignee: KT Corporation
• Original Assignee: KT Corporation
• Inventors: Kim, Mi-Jeom, Lee, Eo-Hyung, Hwang, Jin-Kyung
• Primary Examiner(s): LE, HUNG D
• Application Number:
  US13/446,052
• Publication Number:
  US 20120284229A1
• Time in Patent Office: 900 Days
• Field of Search: None
• US Class Current: 707/634
• 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 16/275 Synchronous replication
• CPC Class Codes: H04L 67/1095 Replication or mirroring of data, e.g. scheduling or transport for data synchronisation between network nodes
• 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

The disclosure is related to selecting data nodes in a distributed storage system. A server may be provided in a distributed storage system including a plurality of data nodes for providing a storage service. The server may include a receiver configured to receive a replication request and a control unit configured to select data nodes based on at least one of evaluation results and real-time service statuses of the plurality of data nodes. An evaluation result of each data node may be a sum of points of at least one evaluation item assigned to each data node from evaluation of each data node according to at least one evaluating item.

Claims

1. A server in a distributed storage system including a plurality of data nodes for providing a storage service, the server comprising:

a receiver configured to receive a replication request; and

a control processor configured to select data node groups based on evaluation results and real-time service statuses of the data node groups up to a number of replicas to be created based on a node group selection policy for restricting replicas of an object from being stored in data nodes belonging to a same data node group, and to select one data node from each one of the selected data node groups based on evaluation results and real-time service statues of the data nodes,

wherein an evaluation result of each data node is a sum of points of evaluation items assigned to each data node from evaluation of each data node according to the evaluating items, and

wherein the evaluation item includes performance, reliability, availability, and scalability and an evaluation result of each data node is a sum of points of a performance evaluation item point, a reliability evaluation item point, an availability evaluation item point, and a scalability evaluation item point, associated with each data node based on an evaluation of each data node according to one or more of performance, reliability, availability, and scalability.

2. The server of claim 1, comprising:

a monitor processor configured to obtain real-time status of a storage service at each data node by monitoring a status of each data node in real-time.

3. The server of claim 2, wherein the real-time status includes at least one status item and the real-time status is a sum of numeric values of the at least one status item.

4. The server of claim 3, wherein the status item includes an average service response time, an average service failure rate in a unit time, and a remaining available storage space of each data node.

5. The server of claim 2, further comprising:

an update processor configured to adjust the evaluation results of the data nodes stored in a database, based on the obtained real-time statuses of the data nodes, and to update the database with the adjustment results.

6. The server of claim 5, wherein the control processor is configured to apply corresponding weights to the evaluation items of the adjusted evaluation results.

7. The server of claim 1, wherein the control processor is configured to:

determine whether the node group selection policy is applied;

determine a number of replicas created based on the replication request; and

select a number of data node groups based on evaluation results of the data node groups up to the determined number of replicas when the node group selection policy is applied.

8. The server of claim 1, wherein:

a data node group includes data nodes located with a certain area; and

an evaluation result of each data node group is an average of evaluation results of data nodes belonging to each data node group.

9. The server of claim 1, wherein the control processor is configured to select a number of data nodes based on the evaluation results of the data nodes up to the determined number of replicas when the node group selection policy is ignored.

10. The server of claim 1, wherein the control processor is configured to generate replicas and to transmit the generated replicas to the selected data nodes, respectively.

11. The server of claim 1, wherein the control processor is configured to request a proxy server to generate replicas and to transmit the generated replicas to the selected data nodes, respectively.

12. A method for selecting data nodes in a distributed storage system, the method comprising:

receiving a replication request;

selecting data node groups based on evaluation results and real-time service statuses of the data node groups up to a number of replicas to be created based on a node group selection policy for restricting replicas of an object from being stored in data nodes belonging to a same data node group; and

selecting one data node from each one of the selected data node groups based on evaluation results and real-time service statues of the data nodes,

wherein an evaluation result of each data node is a sum of points of evaluation items assigned to each data node from evaluation of each data node according to the evaluating items, and

wherein the evaluation item includes one or more of performance, reliability, availability, and scalability, and an evaluation result of each data node is a sum of points of a performance evaluation item point, a reliability evaluation item point, an availability evaluation item point, and a scalability evaluation item point, associated with each data node based on an evaluation of each data node according to performance, reliability, availability, and scalability.

13. The method of claim 12, further comprising:

obtaining the real-time status of a storage service at each data node by monitoring a status of each data node in real-time;

adjusting the evaluation results of the data nodes, stored in a database, based on the obtained real-time statuses of the data nodes; and

updating the database with the adjustment results.

14. The method of claim 13, wherein:

the real-time statuses include at least one status item and the real-time status is a sum of numeric values of the at least one status item; and

the status item includes an average service response time, an average service failure rate in a unit time, and a remaining available storage space of each data node.

15. The method of claim 12, further comprising:

determining whether the node group selection policy is applied;

determining a number of replicas created based on the replication request;

selecting a number of data node groups up to the determined number of replicas when the node group selection policy is applied; and

selecting a number of data nodes based on the evaluation results and the real-time service status of the data nodes up to the determined number of replicas when the node group selection policy is ignored.

16. The method of claim 12, wherein:

a data node group includes data nodes located with a certain area; and

an evaluation result of each data node group is an average of evaluation results of data nodes belonging to each data node group.

17. A server in a distributed storage system including a plurality of data nodes for providing a storage service, the server comprising:

a receiver configured to receive a replication request; and

a control processor configured to select data nodes based on at least one of evaluation results and real-time service statuses of the plurality of data nodes,

wherein the control processor is configured to;

determine whether a node group selection policy is applied;

determine a number of replicas created based on the replication request;

select data node groups based on evaluation results of the data node groups up to the determined number of replicas when the node group selection policy is applied; and

select a data node from each one of the selected data node group based on evaluation results and real-time service statuses of a plurality of data nodes included in a corresponding data node group,

wherein the node group selection policy restricts replicas of one object from being stored in data nodes belonging to a same data node group.

18. The server of claim 17, wherein:

a data node group includes data nodes located with a certain area; and

an evaluation result of each data node group is an average of evaluation results of data nodes belonging to each data node group.