JP2017174000A - Dispersion storage system, dispersion storage program and dispersion storage method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a server capable of quickly collecting pieces of dispersion data from plural servers when collecting pieces of dispersion data and restoring original data.SOLUTION: A dispersion storage system in which a representative server and plural sub servers can mutually communicate through a communication network, comprises: division means for dividing original data which is a storage object into plural pieces of division data; dispersion storage means for dispersing the pieces of division data into plural sub servers and storing the division data, so that same division data is allocated to two or more sub servers; storage destination recording information generation means for generating storage destination recording information for recording, which division data is stored in which sub server, in the dispersion storage means; parallel collection means for, when there is a restoration request of the original data, using the storage destination recording information for collecting plural pieces of division data forming the original data from the plural sub servers simultaneously in a parallel manner; and restoration means for coupling the collected division data for restoring the original data.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a distributed storage technique for electronic data.

  Conventionally, when storing electronic data in a server, it is common to perform backup in preparation for the server going down. In addition, there is a security concern about storing electronic data in a readable state on one server, so electronic data is divided into multiple pieces and each piece of data does not make sense and is distributed and stored Techniques to do this have been proposed.

  For example, in Patent Document 1, a data dividing unit that divides single data into a plurality of divided data, N (N is a plurality) servers that store the plurality of divided data, and (N−1) A data distributed storage system including a data restoration unit that restores the single data based on the plurality of divided data stored in the server is proposed.

Japanese Patent Application Laid-Open No. 2002-268752

  According to this patent document 1, each of the divided data is always stored in a plurality of servers, so that even if one of the servers is down, the original data is restored on the other remaining servers. However, if data is distributed to a large number of servers, it takes time to aggregate the data from each server when generating the original data by consolidating the data. there were.

  The present invention has been made in view of the above problems, and an object thereof is to provide a server capable of aggregating distributed data from a plurality of servers at a high speed when the distributed data is aggregated and the original data is restored. .

  A distributed storage system according to the present invention includes a representative server and a plurality of secondary servers, and the representative server and the plurality of secondary servers can communicate with each other via a communication network. A dividing unit that divides the original data to be stored into a plurality of divided data, and a distributed storing unit that distributes and stores the divided data among the plurality of secondary servers so that the same divided data is allocated to two or more secondary servers; Storage destination recording information generating means for generating storage destination recording information recording which divided data is stored in which secondary server in the distributed storage means, and when there is a request for restoring the original data, the storage destination recording Combines the parallel collection means that collects the divided data that constitutes the original data using information from multiple secondary servers in parallel and the divided data collected in parallel. Characterized in that it comprises a restoring means for restoring the original data.

  A distributed storage program according to the present invention is a distributed storage program for causing a representative server to realize various functions in a distributed storage system in which a representative server and a plurality of secondary servers can communicate with each other via a communication network. A division function for dividing the original data into a plurality of divided data, a distributed storage function for distributing the divided data to a plurality of secondary servers so that the same divided data is allocated to two or more secondary servers, and the distribution A storage destination recording information generation function for generating storage destination recording information that records which divided data is stored in which secondary server in the storage means, and when there is a request for restoring the original data, the storage destination recording information is used. Parallel collection function that collects the divided data that constitutes the original data from multiple secondary servers in parallel and the divided data collected in parallel Bonded to, characterized in that to achieve the restoration function of restoring the original data.

  A distributed storage method according to the present invention is a distributed storage method for distributing and storing data in a distributed storage system in which a representative server and a plurality of secondary servers can communicate with each other via a communication network. A division process for dividing the target original data into a plurality of pieces of divided data, a distributed storage process for distributing the divided data to a plurality of secondary servers so that the same divided data is allocated to two or more secondary servers, and When there is a storage destination recording information generation process for generating storage destination recording information that records which divided data is stored in which secondary server in the distributed storage means, and when there is a restoration request of the original data, the storage destination recording information is Use the parallel collection process that collects the divided data that constitutes the original data from multiple secondary servers in parallel and the divided data collected in parallel. Characterized in that it comprises a restoring process for restoring data.

  According to the present invention, storage destination recording information that records which secondary server stores the divided data is generated, and when there is a request to restore the original data, the storage destination recording information is used to configure the original data Since the divided data to be collected is simultaneously collected from a plurality of secondary servers in parallel, distributed data can be aggregated from a plurality of servers at high speed.

It is a block diagram showing the structure of the representative server 10 in the distributed storage system 100 of this invention. It is explanatory drawing explaining the example in the case of carrying out distributed preservation | save of original data. It is explanatory drawing explaining the example in the case of decompress | restoring original data. It is a flowchart figure showing the flow in the case of carrying out distributed preservation | save of original data. It is a flowchart figure showing the flow in the case of restoring original data. It is explanatory drawing showing an example of the preservation | save destination recording information.

[One Embodiment]
Hereinafter, an example of the distributed storage system 100 according to the first embodiment will be described with reference to the drawings. The distributed storage system 100 includes a representative server 10 and a plurality of secondary servers 20a to 20n, which are configured to be able to communicate with each other via a communication network.

  FIG. 1 is a block diagram showing the configuration of the representative server 10 in the distributed storage system 100 of the present invention. As shown in FIG. 1, the representative server 10 includes a dividing unit 11 that divides the original data to be stored into a plurality of divided data, and a plurality of pieces of divided data so that the same divided data is allocated to two or more secondary servers. Distributed storage unit 12 that distributes to secondary servers and distributes and stores all the divided data so that it is not stored in one secondary server for security safety, and saves which divided data to which secondary server A storage destination recording information generation unit 13 that generates storage destination recording information that records whether the original data has been restored, and a plurality of secondary servers that divide the divided data that constitutes the original data using the storage destination recording information when there is a request to restore the original data A parallel collection unit 14 that collects data in parallel at the same time, a restoration unit 15 that combines the divided data collected in parallel to restore the original data, and a control that controls processing in each of these units. And parts 16, and a storage unit 17 for storing the necessary information, including the destination recorded information.

  FIG. 6 is an explanatory diagram showing an example of storage destination recording information. The storage destination record information generated by the storage destination record information generation unit 13 includes, for example, the name of the original data, the name of the divided data, the location of the storage destination, the stored content, the data collection history, the collection failure history, and the like. Therefore, it is possible to refer to which secondary server each divided data is stored. In addition, since a data collection history, a collection failure history, and the like are recorded, for example, it is referred to which secondary server 20 has a high utilization rate based on the data collection history, or a secondary server that may have gone down. 20 can be identified from the collection failure history.

  Although not shown, the secondary servers 20a to 20n are configured to be able to communicate with the representative server 10 via a communication network, and at least a control unit having various control functions, and a storage unit that stores divided data. Must be provided. In practice, it is practical to have the same configuration as that of the representative server 10 shown in FIG.

  FIG. 2 is an explanatory diagram for explaining an example in which original data is stored in a distributed manner. As shown in FIG. 2, the representative server 10 divides the original data ABC. The original data ABC is divided by the dividing unit 11 into, for example, divided data A, divided data B, and divided data C. At this time, it is preferable not to simply divide, but to perform an encryption process so that information contents cannot be referred to from each divided data. Various encryption methods can be employed. When a method that requires a common key for encryption and decryption is employed, it is necessary to separately manage the keys.

  After the division, the divided data A to C are distributed to a plurality of secondary servers so that the same divided data is allocated to two or more secondary servers in the distributed storage unit 12, and security safety is ensured. For this reason, the data is distributed and stored in one secondary server so that all the divided data is not stored. In the example of FIG. 2, among the secondary servers 20a to 20c, the divided data A and B are stored in the secondary server 20a, the divided data B and C are stored in the secondary server 20b, and the secondary server 20c is stored. The divided data C and A are stored. At the same time, the storage location recording information generation unit 13 generates storage location recording information that records which divided data is stored in which secondary server, and stores it in the storage unit 17.

  FIG. 3 is an explanatory diagram illustrating an example of restoring original data. As shown in FIG. 3, when a request for restoring the original data ABC is made to the representative server 10, the parallel collection unit 14 of the representative server 10 uses the storage destination recording information to store the divided data that constitutes the original data. Collected in parallel from a plurality of secondary servers 20a to 20c. At this time, a collection command is simultaneously transmitted to two or more secondary servers to which the same divided data is assigned, and the divided data is collected from the secondary server that has made a response or the secondary server that has made the fastest response. . Or you may make it collect the said division | segmentation data from the secondary server with the fastest line speed among the secondary servers which answered.

  In this way, when the collection of all the divided data A to C is completed, the restoration unit 15 combines the divided data A to C to restore the original data ABC. Further, when the encryption process is performed at the time of the division, the decryption unit 15 performs the decryption process at the same time. In this way, the original data ABC can be obtained.

  FIG. 4 is a flowchart showing the flow when the original data is stored in a distributed manner. First, the representative server 10 receives a distributed data storage request (S01). When this request is made, the number n of secondary servers that can be actually designated as storage destinations in the current distributed storage system 100 is confirmed (S02). At this time, if two or more secondary servers 20 exist, distributed storage is possible, but it is preferable that three or more secondary servers 20 exist. If distributed storage is possible, the original data is divided into two or more arbitrary numbers (S03). The number of divisions may be determined based on the data capacity, or may be divided into the number desired by the client.

  After the division, the data is distributed and stored in a plurality of secondary servers. At this time, one divided data is stored in two or more secondary servers 20 so that the original data can be restored even if one of the secondary servers 20 goes down. It is stored and distributed for storage so as not to store all the divided data in one secondary server 20 for security safety (S04). Then, storage destination recording information in which the storage destinations of all the divided data distributed in this manner are recorded is generated and stored in the storage unit 17 (S05).

  FIG. 5 is a flowchart showing a flow when restoring the original data. First, the representative server 10 receives a request for restoring original data (S11). When this request is made, a collection command is transmitted to all the storage destination secondary servers for each divided data based on the storage destination record information (S12). Then, the divided data is collected from the optimum secondary server for each divided data (S13). The optimum secondary server condition may be the secondary server with the fastest answer or the secondary server with the fastest line speed.

  An optimum secondary server is determined for each divided data, and collection of all the divided data is executed in parallel (S13). By collecting in parallel, it is possible to collect all the divided data faster than collecting one by one. When the collection of all the divided data is completed, the original data is restored by combining them (S14).

  The process for selecting the optimum secondary server will be described in more detail. When a collection command is transmitted to all the storage destination secondary servers for each divided data, the processing when there are responses from a plurality of secondary servers becomes a problem. One method is to use the secondary server with the fastest answer. In this case, the collection speed may be slow as a result. Therefore, the collection speed can be further improved by comparing the communication speed with the secondary server that has responded and selecting the secondary server having a high line speed.

  Furthermore, in order to adjust the balance with the collection of other divided data, even if the line speed is high, the secondary server that is collecting other divided data is excluded from the candidates. Faster data collection. By performing such adjustment and selecting the optimum secondary server, it becomes possible to collect all the divided data at high speed.

  As described above, according to the distributed storage system 100 according to the present invention, the storage destination recording information that records which secondary server stores the divided data is generated, and the original data is requested to be restored. Since the division data constituting the original data is simultaneously collected in parallel from the plurality of secondary servers using the storage destination recording information, the distributed data can be aggregated from the plurality of secondary servers at high speed. Also, the same divided data is stored in multiple secondary servers, and at the time of collection, the secondary server with the fastest response or the secondary server with the fast line speed is selected. A secondary server is selected to enable high-speed collection of divided data.

  In the first embodiment, the representative server 10 is not used as the secondary server 20 as the storage destination of the divided data. However, the storage unit 17 of the representative server 10 may be one of the distribution destinations of the divided data.

  In the first embodiment, the master-slave relationship has been described as the representative server 10 and the secondary server 20, but the representative server 10 and the secondary server 20 are configured in the same manner, and both function in the same manner. You may make it do. When there is a request for restoring the original data to the secondary server 20, the representative server 10 that has divided the original data is requested to transmit the storage destination record information, and based on the received storage destination record information. The secondary server 20 may perform the restoration process. Further, at this time, if the storage destination record information is shared between the representative server 10 and the secondary server 20 in advance, it is possible to immediately shift to the restoration process even in the server that has not performed the division process. Become.

  For example, in the medical field, when the electronic medical record is managed by a server, the present invention provides high-speed collection at the time of use and stresses the user while realizing falsification prevention and data backup by encryption and distributed storage. Not useful for managing electronic medical records.

DESCRIPTION OF SYMBOLS 100 Distributed storage system 10 Representative server 11 Division | segmentation part 12 Distributed storage part 13 Storage destination recording information generation part 14 Parallel collection part 15 Restoration part 16 Control part 17 Storage | storage part 20, 20a-20n Subserver

Claims (7)

A distributed storage system comprising a representative server and a plurality of secondary servers, wherein the representative server and the plurality of secondary servers can communicate with each other via a communication network,
The representative server includes dividing means for dividing the original data to be saved into a plurality of divided data,
Distributed storage means for distributing and storing divided data among a plurality of secondary servers so that the same divided data is allocated to two or more secondary servers;
Storage destination record information generating means for generating storage destination record information recording which divided data is stored in which secondary server in the distributed storage means;
When there is a request to restore the original data, parallel collection means for collecting the divided data constituting the original data from the plurality of secondary servers simultaneously in parallel using the storage destination recording information;
A distributed storage system comprising: restoration means for restoring the original data by combining the divided data collected in parallel. The dividing means has a function of performing an encryption process when dividing the original data into a plurality of divided data,
The distributed storage system according to claim 1, wherein the restoring unit has a function of combining and combining the divided data subjected to the encryption processing.   The parallel collection means simultaneously transmits a collection command to two or more secondary servers to which the same divided data is assigned based on the storage destination record information, and the secondary server that has made a reply or the secondary server that has the fastest reply. 3. The distributed storage system according to claim 1, wherein the divided data is collected from a server.   The parallel collection means simultaneously transmits a collection command to two or more secondary servers to which the same divided data is assigned based on the storage destination record information, and has the fastest line speed among the secondary servers that have answered. 3. The distributed storage system according to claim 1, wherein the divided data is collected from the secondary server.   If there is a request for restoring the original data to any of the secondary servers, the secondary server requests the representative server to transfer the storage destination record information, and based on the transferred storage destination record information The distributed storage system according to any one of claims 1 to 4, wherein parallel collection is performed by the secondary server. A distributed storage program for realizing various functions in a representative server in a distributed storage system in which a representative server and a plurality of secondary servers can communicate with each other via a communication network,
A division function for dividing the original data to be saved into a plurality of divided data;
A distributed storage function that distributes and stores divided data among a plurality of secondary servers so that the same divided data is allocated to two or more secondary servers;
A storage location record information generation function for generating storage location record information recording which divided data is stored in which secondary server in the distributed storage means;
When there is a request for restoring original data, a parallel collection function that collects the divided data constituting the original data from the plurality of secondary servers in parallel using the storage destination recording information, and
A distributed storage program characterized by realizing a restoration function that restores original data by combining divided data collected in parallel. A distributed storage method for distributing and storing data in a distributed storage system in which a representative server and a plurality of secondary servers can communicate with each other via a communication network,
The representative server includes a dividing process for dividing the original data to be saved into a plurality of divided data,
Distributed storage processing for distributing and storing divided data among a plurality of secondary servers so that the same divided data is allocated to two or more secondary servers;
A storage location record information generation process for generating storage location record information recording which divided data is stored in which secondary server in the distributed storage means;
When there is a request to restore the original data, a parallel collection process for collecting the divided data constituting the original data from the plurality of secondary servers simultaneously in parallel using the storage destination recording information;
A distributed storage method comprising: a restoration process for restoring original data by combining divided data collected in parallel.

Images