JP2017033435A - Data decentralization management system and operation method thereof - Google Patents

Abstract

PROBLEM TO BE SOLVED: To determine whether there is a possibility that data are inconsistent with each other when some of a plurality of data are restored.SOLUTION: A plurality of data are backed up at timing of mutual consistency thereof and the backup timing is stored. When only some of the data are restored, it is determined whether the timestamp of data which is not restored indicates timing after the backup timing (T1). When YES at T1, it is determined that the data may be inconsistent with each other supposing that they are restored (T3). When a timestamp of each of data which are not restored indicates timing before the backup timing (T5: NO), it is determined that the data are consistent with each other supposing that they are restored (T7).SELECTED DRAWING: Figure 5

Description

  The present invention relates to a data distribution management system and an operation method thereof.

  Conventionally, for example, a plurality of data is distributed and stored in a distributed database server and backed up to a backup server. In the distributed database server, when data is lost due to a failure of a storage device storing certain data, and then the storage device is restored, the data backed up to the backup server is transmitted to the distributed database server and restored.

JP 2001-282761 A JP 2001-156778 A

  By the way, for example, other data is updated during a failure, so that the restored data becomes relatively old, that is, the data may not match each other.

  In this case, when the restored data is updated and becomes new, the data are consistent with each other.

  Therefore, until then, there is an inconvenience that data cannot be used for service provision or the like.

  The present invention has been made in view of the above-described conventional problems, and provides a technique that makes it possible to determine whether data may not match each other when a part of a plurality of data is restored. With the goal.

  In order to solve the above-described problem, a data distribution management system according to the first aspect of the present invention includes a distributed database server that stores a plurality of pieces of data in association with time stamps indicating update timings of the respective data; A backup server that acquires and stores at least one backup target data as the data from the distributed database server at a timing when the data is consistent with each other, and later transmits the backup target data to the distributed database server for storage Storing a backup timing that is a timing at which the backup target data is stored in the backup server, comparing a timing of a time stamp of each data excluding the backup target data with the backup timing, If one of the time stamps indicates a timing after the backup timing, it is determined that the data may not be consistent with each other when the backup target data is transmitted to the distributed database server and stored. And a control device that determines that the data matches each other when the backup target data is transmitted to the distributed database server and stored if each time stamp indicates a timing before the backup timing. And

  The operation method of the data distribution management system according to the second aspect of the present invention is such that the data distribution management system distributes and stores a plurality of data in association with time stamps indicating the update timing of each data, Backup in which at least one backup target data, which is the data, is acquired from the distributed database server and stored at a timing when a plurality of data are consistent with each other, and then the backup target data is transmitted to the distributed database server for storage A server and a control device that stores a backup timing that is a timing at which the backup target data is stored in the backup server, and the operation method includes: a time stamp of each data excluding the backup target data; The timing is compared with the backup timing, and if at least one of the time stamps indicates a timing after the backup timing, the control device transmits the backup target data to the distributed database server for storage. In some cases, when it is determined that the data may not be consistent with each other, if each time stamp indicates a timing before the backup timing, the backup target data is transmitted to the distributed database server and stored. Is characterized in that it is determined that the data match each other.

  According to the present invention, it is possible to determine whether or not data may not match each other when a part of a plurality of data is restored.

It is a figure which shows the data distribution management system concerning embodiment of this invention. It is a figure which shows the example which disperse | stores and memorize | stores several data relevant to each other. It is a figure which shows the example of data and a time stamp. It is a sequence diagram which shows the example of operation | movement of a data distribution management system. It is a flowchart which shows the determination method of step S61 in FIG. It is a figure which shows the reason which can determine whether data may not mutually match.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is a diagram showing a data distribution management system according to an embodiment of the present invention.

  The data distribution management system 1 is a system in which a distributed database server 11, a backup server 12, and a control device 13 are connected via a communication line, and the distributed database server 11 is connected to the service providing apparatus 2 via a communication line.

  The distributed database server 11 includes a plurality of storage devices A1, A2, A3,..., And the backup server 12 also includes a plurality of storage devices B1, B2, B3,.

  The distributed database server 11 distributes and stores a plurality of data related to each other in the storage devices A1, A2, A3,. The distributed database server 11 stores a plurality of such a plurality of data. When the distributed database server 11 is requested for data from the service providing apparatus 2, the distributed database server 11 transmits the requested data to the service providing apparatus 2. With this data, the service providing apparatus 2 provides a service to the user. The service is, for example, a communication service, but may be another service.

  The backup server 12 receives a plurality of data related to each other stored by the distributed database server 11, and stores them as backups in a distributed manner in the storage devices B1, B2, B3,. When the data is lost due to a failure of the storage device of the distributed database server 11 and then the storage device is restored, the backup server 12 transmits the data stored as a backup to the distributed database server 11 and restores the data.

  FIG. 2 is a diagram illustrating an example of storing a plurality of data related to each other in a distributed manner.

  The storage device A1 stores data D1 among the data D1, D2, D3, and D4 related to each other. The storage device A2 stores data D2, and the storage device A3 stores data D3 and D4.

  The storage device B1 stores data D1 and D2. The storage device B2 stores data D3, and the storage device B3 stores data D4.

  Note that this is an example of a set of a plurality of data related to each other, and the other data need not be the same. The number of storage devices is not limited to three.

  FIG. 3 is a diagram illustrating an example of data and a time stamp.

  In the distributed database server 11, each data is stored in association with a time stamp indicating the data update timing. The figure shows how data D1 is stored in association with a time stamp indicating that it was updated at 10:00 on October 1. Note that the granularity of the time stamp is arbitrary, and may indicate up to seconds.

  FIG. 4 is a sequence diagram illustrating an example of the operation of the data distribution management system.

  The distributed database server 11 stores data D1 to D4 as shown in FIG. 2, and a time stamp is associated with the data. At this time, it is assumed that the data D1 to D4 are consistent with each other (S1).

  For example, the data D1 and the time stamp are updated in the storage device A1 (S3), so that, for example, the data D1 to D4 do not match each other (S5).

  Next, the data D2 and the time stamp are updated in the storage device A2 (S13). For example, the data D1 to D4 remain inconsistent with each other (S5).

  Next, the data D3, D4 and the time stamp are updated in the storage device A3 (S23), and thereby, for example, the data D1 to D4 match each other (S25).

  The backup server 12 performs backup at the timing at which the data D1 to D4 are consistent with each other. That is, the distributed database server 11 transmits the data D1 to D4 to the backup server 12 (S31), and the backup server 12 receives the data D1 to D4 (S33). Whether or not data is transmitted and received in association with a time stamp is arbitrary. In addition, it is not necessary to perform backup every time the data is consistent with each other.

  For example, as shown in FIG. 2, the storage device B1 stores data D1 and D2 (S35), the storage device B2 stores data D3 (S37), and the storage device B3 stores data D4 (S39).

  The backup server 12 notifies the control device 13 of the backup timing that is the timing at which data is stored in the backup server 12 (S41), and the control device 13 stores the backup timing (S43).

  On the other hand, in the distributed database server 11, for example, the storage device A2 fails and the data D2 is lost (S51). Thereafter, the storage device A2 is restored (S53), but the data D2 has not yet been restored to the storage device A2 (does not exist).

  In order to store the data D2 in the storage device A2, the distributed database server 11 restores the control device 13 including time stamps (three time stamps) associated with the data D1, D3, and D4 other than the data D2. A request (restore request for data D2) is transmitted (S55).

  Whether the control device 13 receives the restore request and temporarily transmits the data D2 from the backup server 12 to the distributed database server 11 and stores it (if it is temporarily restored), whether or not the data may be inconsistent. Is determined (S61). Details of this determination will be described later.

  For example, in the case where the data D2 is restored, the control device 13 transmits a restore command for the data D2 to the control device 13 if it is determined that the data match each other (S63).

  The backup server 12 transmits the data D2 to the distributed database server 11 by the restore command (S71), and the distributed database server 11 receives the data D2 (S73).

  The storage device A2 stores the received data D2 (S75). As determined, the data D1 to D4 match each other (S77).

  For example, the distributed database server 11 transmits the requested data to the service providing apparatus 2 to enable service provision.

  Note that the determination in step S61 may be made after completing step S75. If it is determined in step S61 that the data are consistent with each other, for example, the distributed database server 11 transmits the requested data to the service providing apparatus 2 to enable service provision. On the other hand, if it is determined that there is a possibility that the data does not match each other, the distributed database server 11 does not transmit data, thereby preventing the occurrence of a problem due to inconsistency.

  FIG. 5 is a flowchart showing the determination method of step S61 in FIG.

  First, for one time stamp in the restore request, the time stamp timing is compared with the backup timing, and it is determined whether or not the time stamp indicates a timing after the backup timing (T1). If the time stamp indicates a timing after the backup timing (T1: YES), it is determined that there is a possibility that the data do not match each other when the data is restored (T3), and the processing is terminated.

  On the other hand, if the time stamp indicates a timing before the backup timing (T1: NO), it is determined whether or not there is a time stamp that has not been determined in step T1 in the restore request (T5). If there is a time stamp that has not been determined in step T1 in the restore request (T5: YES), the process returns to step T1, and one of them is determined (T1). On the other hand, if there is no time stamp that has not been determined in step T1 in the restore request (T5: NO), it is determined that the data are consistent with each other when the data is restored (T7), and the process ends.

  FIG. 6 is a diagram illustrating the reason why it can be determined whether there is a possibility that data does not match each other.

  FIG. 6A shows an example of a case where data may not match each other.

  It is assumed that the time stamps (t) of the data D3 and D4 updated in step S23 indicate 10:00 on October 1.

  In addition, the backup timing (T) stored in step S43 indicates 0:00 on October 2.

  Thereafter, the data D1 is updated, and the time stamp (t) indicates 8:00 on October 2 (S301). By updating the data D1, for example, the data D1 to D4 do not match each other (S303).

  In this case, in step T1 in FIG. 5, the time stamp in step 301 (t = October 2, 8:00) indicates a timing after the backup timing (T = October 2, 0:00) ( T1: YES), it is determined that there is a possibility that the data D1 to D4 do not match each other when the data D2 is restored (T3). That is, a correct determination is made.

  In this case, for example, the distributed database server 11 does not transmit data to the service providing apparatus 2, thereby preventing occurrence of a problem due to inconsistency.

  In addition, even if the data D1 in S301 is updated, the data D1 to D4 may be consistent with each other. However, in the present embodiment, it is assumed that data may not match each other as a safety measure.

  FIG. 6B shows an example where the data matches each other.

  It is assumed that the time stamp (t) of the data D1 updated in step S3 indicates 8:00 on October 1st.

  It is assumed that the time stamps (t) of the data D3 and D4 updated in step S23 indicate 10:00 on October 1.

  In addition, the backup timing (T) stored in step S43 indicates 0:00 on October 2.

  Thereafter, there is no data update (S401), and the data D1 to D4 are consistent with each other (S403).

  In this case, in step T1 of FIG. 5, the time stamp (t = October 1, 8:00) of step S3 indicates the timing before the backup timing (T = October 2, 0:00) (T1: NO), because the time stamp of step S23 (t = October 1st, 10:00) indicates the timing before the backup timing (T = October 2nd, 0:00) (T1: NO), data When D2 is restored, it is determined that the data D1 to D4 match each other (T7). That is, a correct determination is made.

  Therefore, in this case, for example, the data D2 is actually restored (S63 to S75), and the data D1 to D4 match each other (S77).

  Since the data match each other, the distributed database server 11 can transmit the requested data to the service providing apparatus 2, and the service providing apparatus 2 can provide a service to the user based on this data.

  As described above, according to the data distribution management system of the present embodiment, a plurality of data (D1 to D4) are distributed and stored in association with the time stamp indicating the update timing of each data (S3, S13, S23). The distributed database server 11 and the data to be backed up (D2), which is at least one data, are acquired from the distributed database server 11 and stored at a timing when a plurality of data are consistent with each other (S25) (S33 to S39). The backup target data (D2) is transmitted to and stored in the distributed database server 11 (S71), and the backup timing that is the timing at which the backup target data (D2) is stored in the backup server 12 is stored (S43). Backup target data (D2) The time stamp timing and backup timing of each data to be excluded are compared (T1), and if at least one time stamp indicates a timing after the backup timing (T1: YES), the backup target data (D2) is stored in the distributed database. When the data is transmitted to the server 11 and stored (when restored), it is determined that the data may not match each other (T3). On the other hand, if each time stamp indicates a timing before the backup timing (T5: NO), when the backup target data (D2) is transmitted to and stored in the distributed database server 11 (when restored), it is determined that the data match each other (T7). If you restore a part of Data can be determined whether there is a possibility not to conform to each other.

  In the present embodiment, all of the data D1 to D4 are backed up, that is, all are data to be backed up, but the data to be backed up may not be all. For example, only data (for example, data D2) stored in a storage device with a high possibility of failure may be backed up as data to be backed up and restored.

  The determination in step S61 may be performed by the distributed database server 11 or the backup server 12.

  A computer program for causing a computer to function as the distributed database server 11, the backup server 12, and the control device 13 can be recorded on a computer-readable recording medium such as a semiconductor memory, a magnetic disk, an optical disk, a magneto-optical disk, or a magnetic tape. Further, it can be widely distributed by being transmitted through a communication network such as the Internet.

1 Data distribution management system 11 Distributed database server 12 Backup server 13 Control devices A1-A3, B1-B2 Storage devices D1-D4 Data

Claims (4)

A distributed database server that stores a plurality of data in a distributed manner in association with time stamps indicating the update timing of the data;
At a timing when the plurality of data are consistent with each other, at least one backup target data that is the data is acquired from the distributed database server and stored, and then the backup target data is transmitted to the distributed database server for storage. A backup server,
A backup timing which is a timing at which the backup target data is stored in the backup server is stored, a time stamp timing of each data excluding the backup target data is compared with the backup timing, and at least one time stamp is If the timing after the backup timing is indicated, when the backup target data is transmitted to the distributed database server and stored, it is determined that the data may not be consistent with each other. If the timing before the backup timing is indicated, the control device determines that the data is consistent with each other when the backup target data is transmitted to the distributed database server and stored. A data distribution management system characterized by that. 2. The data distribution management system according to claim 1, wherein if it is determined that the data match each other, the backup server transmits the backup target data to the distributed database server for storage. An operation method of a data distribution management system,
The data distribution management system includes: a distributed database server that stores a plurality of pieces of data in association with time stamps indicating the update timing of each data; and at least one piece of the data at a timing when the plurality of pieces of data match each other Backup data that is acquired from the distributed database server and stored, and later the backup target data is transmitted to the distributed database server for storage, and the timing at which the backup target data is stored in the backup server Comprising a control device for storing backup timing,
The operation method is as follows:
The control device compares the timing of the time stamp of each data excluding the backup target data and the backup timing,
If the control device indicates that the at least one time stamp indicates a timing after the backup timing, the data may not match each other when the backup target data is transmitted to the distributed database server and stored. On the other hand, if each time stamp indicates a timing before the backup timing, it is determined that the data match each other when the backup target data is transmitted to the distributed database server and stored. A method of operating a data distribution management system characterized by the above. 4. The method of operating a data distribution management system according to claim 3, wherein if the backup server determines that the data are consistent with each other, the backup target data is transmitted to the distributed database server for storage.

Images