JP2008191898A - Data backup system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To ensure confidentiality of backup data. <P>SOLUTION: A data backup system comprises a server and a plurality of distributed computers 2-1 to 2-4 each having a disk device and networked to the server. Backup data is divided into a plurality of division data, and the division data are sent from the server to the distributed computers 2-1 to 2-4 via the network, where the division data are stored in the disk devices of the distributed computers 2-1 to 2-4. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a data backup system for backing up various data.

  Conventionally, when backing up data stored in an information processing system, etc., a user terminal storing data to be backed up is connected to a backup device having a magnetic tape device for backup, and the backup data is sent from the user terminal to the backup device. Was stored in the magnetic tape unit. However, to purchase an expensive backup magnetic tape device by the user himself / herself and to construct a backup system has forced the user to perform complicated work and has a large cost burden.

  Therefore, a backup system has been proposed in which a user terminal and a backup center are connected via the Internet, backup data is transmitted from the user terminal to the backup center, and the backup data is stored in a magnetic tape device installed in the backup center ( For example, see Patent Document 1). According to this backup system, it is not necessary to install a magnetic tape device for backup on the user side, and the burden on the user can be reduced.

By the way, storing backup data in the backup center can easily restore the original data even if the data is lost on the user side, but there is a high possibility that the data leaks. Conventionally, in order to prevent leakage of backup data, measures such as limiting access to the backup data or encrypting the backup data have been taken.
JP 2002-108678 A

  However, even if access to backup data is restricted or encrypted, it is difficult to completely prevent leakage of backup data due to theft of a magnetic tape or unauthorized access.

  The present invention has been made in view of the above circumstances, and an object thereof is to provide a data backup system capable of ensuring the confidentiality of backup data.

  To achieve the above object, a data backup system according to the present invention is a data backup system comprising a server and a plurality of distributed computers each having a disk device and connected to the server via a network. The backup data is divided into a plurality of divided data, the divided data is transmitted from the server to each distributed computer via a network, and the divided data is stored in a disk device of each distributed computer.

  According to this data backup system, since the backup data is divided into a plurality of divided data and distributed and stored in the disk devices of a plurality of distributed computers, even if the divided data leaks alone from the distributed computer as the backup destination, Backup data cannot be restored, and confidentiality can be maintained while the backup data is distributed and stored.

  According to the present invention, in the data backup system, each of the distributed computers is installed in a geographically distant place.

  According to this data backup system, multiple distributed computers are installed in geographically separated locations, so even if some of the distributed computers become unusable due to a disaster, etc. Data can be restored from the divided data or parity data stored in the remaining distributed computers that are installed and immune from disasters, and the reliability of the system can be improved.

  Further, the present invention provides the data backup system, wherein the parity data generated based on the divided data and used to restore a single data and the number of divisions of the divided data are N, and M (N> M) or more The backup data can be restored by using the divided data and the parity data stored in the distributed computer.

  According to this data backup system, since data cannot be restored without collecting divided data from M (N> M) or more distributed computers, even if divided data or parity data leaks alone from some distributed computers, Backup data cannot be restored and confidentiality can be maintained.

  According to the present invention, in the data backup system, parity data is made redundant and stored in separate distributed computers so that the backup data can be restored even if some of the plurality of distributed computers are unusable. It is characterized by that.

  According to this data backup system, since parity data is made redundant and stored in the disk devices of the plurality of distributed computers, the redundant parity data is used even if some percent of the backup destination distributed computers cannot be used. The original backup data can be restored and the reliability of the system can be improved.

  According to the data backup system of the present invention, the confidentiality of backup data can be ensured.

Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.
FIG. 1 is a network configuration diagram to which the data backup system according to the present embodiment is applied. The data backup system according to the present embodiment includes a server 1 and a plurality of distributed computers 2-1 to 2-n serving as backup destinations of original data. The plurality of distributed computers 2-1 to 2-n are preferably set at locations physically separated from the server 1 for the purpose of avoiding risks such as large-scale disasters or fires. It is desirable that 2-n is physically separated from each other.

  The server 1 is arranged on the LAN 3. The server 1 holds the original data and manages backup destination information of a plurality of divided data obtained by dividing the original data. A communication infrastructure is constructed so that the server 1 can be connected to the Internet / intranet 5 from the router 4 via the LAN 3.

  The distributed computers 2-1 to 2-n include disk devices 6-1 to 6-n that store the divided data requested for backup from the server 1. Each distributed computer 2-1 to 2-n is connected to the Internet / intranet 5 via routers 7-1 to 7-n provided therein, and transmits / receives backup data (divided data) to / from the server 1. I do.

  FIG. 2 is a functional block diagram of the server 1. A data backup instruction is issued from the control unit 11. The original data is stored in the disk device 12. The division / redundancy processing unit 13 divides the original data stored in the disk device 12 into a plurality of pieces of divided data and creates parity data. The number of divisions of the divided data is N, and the backup data can be restored by using the divided data and the parity data stored in the distributed computers of M (N> M) or more. In order to increase the reliability of the system, parity data is multiplexed so that restoration can be performed even if some of the distributed computers 2-1 to 2-n used as backup destinations are unusable. It is desirable to do. For example, if distributed storage using RAID 5 is used, the entire data can be restored using parity data even if one piece of divided data is lost. If distributed storage using RAID 6 is used, two pieces of divided data are lost. Can also restore the whole using parity data.

  The storage destination assigning unit 14 determines a plurality of distributed computers as backup destinations of the plurality of divided data and parity data generated by the division / redundancy processing unit 13, and distributes the plurality of divided data and parity data and their backup destinations. The computer information is passed to the network IF 16. The network IF 16 that stores the distributed computer information of the backup destinations of the plurality of divided data and parity data in the distributed storage information database 15 stores the plurality of divided data and parity data given from the division / redundancy processing unit 13. Backup storage processing is performed based on the destination distributed computer information. That is, a plurality of divided data and parity data are respectively transmitted to the backup destination distributed computers 2-1 to 2-n via the network (LAN2, Internet / Intranet 5), and the disks of the distributed computers 2-1 to 2-n. It memorize | stores in apparatus 6-1-6-n.

  Further, when restoring the original data from the backup data, the control unit 11 transmits a read request to each backup destination distributed computer 2-1 to 2-n from the network IF 16. When reading from the backup destination distributed computers 2-1 to 2-n, the backup destination distributed computer information is acquired from the distributed storage information database 15. The network IF 16 inputs the divided data and parity data read from the backup destination distributed computers 2-1 to 2-n and received via the network (LAN2, Internet / Intranet 5) to the restoration processing unit 17.

  The restoration processing unit 17 is a part that restores original data from a plurality of divided data. If some of the distributed computers used as the backup destination are unusable, a restoration process using parity data is executed.

  Next, the operation of the data backup system of the present embodiment configured as described above will be described.

  FIG. 3 is a conceptual diagram of distributed backup and reading by the data backup system of this embodiment. In the server 1, when a backup request is generated for original data, the control unit 11 designates the original data to be backed up to the division / redundancy processing unit 13 and issues a backup instruction.

  The division / redundancy processing unit 13 extracts the original data to be backed up instructed from the control unit 11 from the disk device 12 and divides it into a plurality of divided data and creates parity data. Here, the parity data is multiplexed so that restoration is possible even if a part of the backup destination distributed computer is unusable.

  A backup destination distributed computer for a plurality of divided data and parity data generated by the division / redundancy processing unit 13 is determined by the storage destination allocation unit 14. The storage destination assigning unit 14 may determine a plurality of distributed computers fixed in advance as backup destination distributed computers, or may dynamically assign a plurality of registered computers according to a predetermined condition. The backup destination distributed computer information of the divided data and parity data determined by the storage destination assigning unit 14 is stored in the distributed storage information database 15. In the example shown in FIG. 3, the divided data A is the distributed computer 2-1, the divided data B is the distributed computer 2-2, the divided data C is the distributed computer 2-3, and the parity data is the distributed computer 2-4. As determined.

  The control unit 11 negotiates with each backup destination distributed computer 2-1 to 2-4 based on the backup destination distributed computer information of the plurality of divided data and parity data determined by the storage destination allocating unit 14 to obtain the divided data and the like. Request to save backup. When an acknowledgment response is received from the backup destination distributed computers 2-1 to 2-4, the allocated divided data or parity data is transmitted to the backup destination distributed computers 2-1 to 2-4. The above processing is repeated with all the backup destination distributed computers 2-1 to 2-4.

  As described above, the original data requested for backup is converted into a plurality of divided data and parity data, and the disk devices 6-1 to 6-6 of the plurality of distributed computers 2-1 to 2-4 physically separated. n is distributed and stored. Since the divided data stored in the disk devices 6-1 to 6-n of the backup destination distributed computers 2-1 to 2-4 are fragments, the original data cannot be restored only by the divided data. In this example, the original data cannot be restored unless the divided data is collected from the M or more distributed computers 2-1 to 2-4.

  On the other hand, when the backup data is required due to the original data being lost in the server 1, the control unit 11 generates a read request for the backup data. The control unit 11 acquires the backup destination distributed computer information of the backup data for which a read request has been issued from the distributed storage information database 15. Then, the control unit 11 makes a request to negotiate with the backup destination distributed computers 3-1 to 3-4 via the network IF 16 and read out the divided data and the like. The divided data or parity data read from the disk devices 6-1 to 6-n in response to the read request in the backup destination distributed computers 2-1 to 2-4 is transmitted to the server 1 via the network.

  In the server 1, when the requested divided data and parity data are received by the network IF 16, the restoration processing unit 17 restores the original original data.

  Here, when the backup destination distributed computer storing the divided data or parity data is in an unusable state, the divided data and parity data cannot be read from the backup destination distributed computer.

  In this embodiment, since parity data is created from original data and parity data is multiplexed, the original data can be restored in the restoration processing unit 17 even if a part of the backup destination distributed computer cannot be used. For example, when the distributed computer 2-3 that stores the divided data C in FIG. 3 cannot be used, the complete original data (ABC) is restored by using the parity data read from the distributed computer 2-4. be able to. Note that RAID 5 or 6 can be used for a technique for restoring original data from a plurality of pieces of divided data in which part of the divided data is partially lost by using parity data, but the present invention can be applied to the algorithm of RAID 5 or 6. It is not limited.

  Thus, according to the present embodiment, the backup data is divided into a plurality of divided data and distributed and stored in the disk devices 6-1 to 6-n of the plurality of distributed computers 2-1 to 2-n. Even if the divided data leaks alone from the distributed computers 6-1 to 6-n as the backup destination, the backup data cannot be restored, and the confidentiality can be maintained while the backup data is distributed and stored.

  In addition, according to the present embodiment, since a plurality of distributed computers 2-1 to 2-n are installed in geographically separated locations, some of the distributed computers become unusable due to a disaster or the like. In addition, data can be restored from the divided data or parity data stored in the remaining distributed computers that are installed in geographically distant locations and are immune from disasters, and the reliability of the system can be improved.

  Further, according to the present embodiment, since data cannot be restored unless the divided data is collected from M (N> M) or more distributed computers 2-1 to 2-n, the divided data or parity data from some of the distributed computers. Even if it leaks alone, the backup data cannot be restored, and confidentiality can be maintained.

  According to the present embodiment, the parity data is made redundant and stored in the disk devices 6-1 to 6-n of the plurality of distributed computers 2-1 to 2-n. Even if the data cannot be used, the original backup data can be restored using the redundant parity data, and the reliability of the system can be improved.

  The present invention is applicable to a data backup system that backs up data to a plurality of computers.

1 is a network configuration diagram to which a data backup system according to an embodiment of the present invention is applied. Functional block diagram of the server in the above embodiment Conceptual diagram showing data backup operation and restoration operation in the above embodiment

Explanation of symbols

1 server 2-1 to 2-n distributed computer 3 LAN
4, 7-1 to 7-n Router 5 Internet / Intranet 6-1 to 6-n Disk device 11 Control unit 12 Disk device 13 Partition / redundancy processing unit 14 Storage destination allocation unit 15 Distributed storage information database 16 Network IF
17 Restoration processing unit

Claims (4)

A data backup system comprising a server and a plurality of distributed computers each comprising a disk device and connected to the server via a network,
Data backup characterized in that backup data is divided into a plurality of divided data, the divided data is transmitted from the server to each distributed computer via a network, and the divided data is stored in a disk device of each distributed computer. system.   2. The data backup system according to claim 1, wherein each of the distributed computers is installed in a geographically distant place.   Divided data and parity generated on the basis of the divided data and stored in M (N> M) or more distributed computers, where N is the parity data used for restoration of single data and the number of divisions of the divided data is N 3. The data backup system according to claim 1, wherein the backup data can be restored by using the data. The parity data is made redundant and stored in separate distributed computers so that backup data can be restored even if some of the plurality of distributed computers are unusable. 4. The data distributed storage system according to any one of 3.

Images