JP2012089049A - Computer system and server - Google Patents

Abstract

A division number and a division backup destination are dynamically determined according to the amount of backup data and the amount of traffic.
A computer system including a plurality of data centers each including a server and a storage system, wherein the server holds transfer order information in which a priority order for transferring a copy of a file is determined and information on a division size of the file. A control unit; a data division unit that divides data requested to be written; and a transmission unit that transmits the divided data. The control unit determines the number of divisions of data received from the user terminal. , Referring to the transfer order information, determining a data center to which the divided data is transferred, and the data dividing unit divides the file requested to be written into the predetermined number of data, The transmission unit transmits the divided data to the determined data center in parallel.
[Selection] Figure 1

Description

  The present invention relates to a file backup method between storage systems (data centers) for storing data, and more particularly to a technique for dynamically changing a data center as a backup destination.

  In recent years, the amount of data stored in data centers has increased. In order to avoid the risk of losing data stored in the data center, data is backed up regularly. In this backup method, a magnetic tape on which stored data is written is mainly transported off-site, and the magnetic tape is safely stored in preparation for a disaster. Data backup to magnetic tape takes time, and the time required for data backup has been increased due to an increase in the amount of data.

  As a method for solving such an increase in the backup time of data, for example, a method of backing up only pointer information referring to actual data when data overlaps with past data at the time of backup (patent) Reference 1).

  Similarly, as a means for solving the increase in backup time and avoiding the risk of data loss, a method of dividing the data to be backed up into a fixed number and storing the divided data in another device has been proposed (patent) Reference 2).

JP 2009-205201 A JP 2007-42027 A

  As the stored data increases, the execution time of a job for backing up data on a magnetic tape becomes longer. Further, when the backup job is not successful once, the backup job is re-executed from the beginning, and the time until the backup job is completed is further increased. For this reason, it is difficult to increase the backup frequency, and there is a problem that the backup tape is not transported to the off-site on an appropriate schedule and the protection from damage is not sufficient.

  In the technique described in Patent Document 1 described above, when past actual data is lost, data referred to by pointer information is lost, which is not sufficient for avoiding the risk of data loss.

  Further, in the technique described in Patent Document 2 described above, since the number of divisions and the transmission destination of the divided data are fixed, if the backup data is huge, the divided data also becomes large and the problem of increase in backup time is solved. I can't.

  In addition, when backing up to another device online, the backup data transfer time increases as the backup data increases, and even if the data is backed up by dividing it, the number of divisions and the divided backup will depend on the traffic volume, etc. It is desirable to determine the destination.

  It is an object of the present invention to provide a computer system and a storage system that can dynamically determine the number of divisions and the division backup destination according to the amount of backup data and the amount of traffic.

  A typical example of the invention disclosed in the present application is as follows. That is, a computer system including a plurality of data centers including at least a first data center, a second data center, and a third data center, wherein the data center includes a processor, a memory, a storage interface, and a network interface. And a storage system comprising a storage control device, a storage device, and an interface, wherein the first data center server is connected to a user terminal, and the first data center server is connected to the storage system. A control unit for holding transfer order information in which a priority order of data transfer to another data center storing a copy of a file to be written and information on a division size of the file requested to be written; and the user terminal The write requested from A data dividing unit that divides the data, and a transmission unit that transmits the divided data. When the control unit receives a file write request from the user terminal, the file requested to be written The data division unit determines the number of divisions of the received data based on the size and the division size information, determines the data center to which the divided data is transferred with reference to the transfer order information, and the data division unit Divides the file requested to be written into the predetermined number of data, and the transmission unit transmits the divided data in parallel to the determined data center.

  According to the exemplary embodiment of the present invention, it is possible to suppress the extension of the backup time even when the data size is large.

1 is a block diagram illustrating a network system according to a first embodiment of this invention. It is a block diagram which shows the structure of the server of the 1st Embodiment of this invention. It is a figure explaining the structure of the transfer order table of the 1st Embodiment of this invention. It is a figure explaining the structure of the original file management table of the 1st Embodiment of this invention. It is a figure explaining the structure of the division | segmentation file management table of the 1st Embodiment of this invention. FIG. 6 is a sequence diagram of processing for performing split backup according to the first embodiment of this invention. FIG. 6 is a sequence diagram of processing for performing split backup according to the first embodiment of this invention. It is a sequence diagram of the reconstruction process of the divided data according to the first embodiment of this invention. It is a sequence diagram of the setting process of the transfer order table of the 2nd Embodiment of this invention. It is a figure explaining the structure of the traffic data total table of the 2nd Embodiment of this invention.

  First, an outline of an embodiment of the present invention will be described.

  In the embodiment of the present invention, the data transmitted from the user terminal 101 is divided into a predetermined size (size that can be stored in a short time), and is transmitted to another data center via the network (for example, WAN) 104. store. When there is a reference request for the data from the user terminal 101, data stored in the own data center is normally transmitted. However, if the data stored in the own data center is damaged, other data Data divided and stored in the center is collected, the divided data are integrated, and the integrated data is transmitted to the user terminal 101.

  Specifically, the data center includes a transmission unit 205 and a reception unit 201 connected to the user terminal 101, a transmission unit 203 and a reception unit 207 connected to another data center, a control unit 204, a data division unit 202, and data A server 102 having a coupling unit 206 and a storage apparatus 103 having a storage device are provided.

  Data received by the PC-side receiving unit 201 from the user terminal 101 is divided by the data dividing unit in accordance with an instruction from the control unit 204, and the divided data is transmitted in parallel from the data center-side transmitting unit 203.

  On the other hand, when a data reference request is received from the user terminal 101, it is determined whether data is stored in the own data center. When the data stored in the own data center is damaged, the data center storing the divided data of the data is requested to transmit the divided data. The divided data received from the other data centers are combined, and the combined data is transmitted from the PC side transmission unit 205.

  Further, the transfer order table 208 held in the control unit 204 is transferred to each past data center instead of the information set from the maintenance terminal 204 when determining the transfer destination data center for each time period. By creating a transfer order table using multiple regression analysis results based on statistics on the amount of data transferred and information that affects the amount of transferred data, the transfer order table according to traffic for each time zone is automatically created. Generate.

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

<Embodiment 1>
FIG. 1 is a block diagram illustrating a network system according to an embodiment of this invention.

  The network system according to the first embodiment includes a data center W100, a data center X105, a data center Y108, and a data center Z111.

  The data center W100 is connected to the data center X105, the data center Y108, and the data center Z111 via the network 104. The data centers may be connected not by a network but by a dedicated line. Also, the network connecting the data centers and the network connecting the data center W100 and the user terminal may be the same network or different networks. The data center W100 accommodates a user terminal (for example, a personal computer) 101.

  Data center X105, data center Y108, and data center Z111 may be provided at a location (remote location) away from data center W100 or in the same building as data center W100. Note that the present invention is more effective when the distance between data centers is longer because a longer time is required to transfer data.

  The data center W100 includes a server 102 and a storage device 103. The server 102 includes a processor 1021 that executes a program, a memory 1022 that stores a program executed by the processor 1021, a network interface 1023 connected to the network 104, and a storage interface 1024 connected to the storage apparatus 103. It is a computer that provides an interface with. Note that one network interface may be mounted on the server 102 to control both communication between data centers and communication with the user terminal 101.

  The storage device 103 has a storage control device 1031 that controls input / output of data to / from the storage device 1033, a storage interface 1032 connected to the server 102, and a nonvolatile storage device 1033 (for example, a magnetic disk device, flash memory). Then, in accordance with a request from the server 102, data to be stored is stored in the storage device 1033, and data to be referred to is read from the storage device 1033.

  A management terminal 200 is connected to the data center W100. The management terminal 200 is a computer having a processor that executes a program, a memory that stores a program executed by the processor, a network interface connected to the network 104, and a user interface (for example, a keyboard and a display device). The system administrator sets the data division size from the maintenance terminal 200 and inputs the transfer order table 208 to set the data center to which the data is to be sent.

  The data center W100 has been described above. Similarly, the data center X105 includes the server 106 and the storage device 107, the data center Y108 includes the server 109 and the storage device 110, and the data center Z111 includes the server 112 and the storage device. 113. The configuration of each server and each storage device is the same as that of the server 102 and the storage device 103 of the data center W100.

  FIG. 2 is a block diagram showing a configuration of the server according to the embodiment of this invention.

  As described above, the server 102 is a computer having the processor 1021, the memory 1022, the network interface 1023, and the storage interface 1024. The PC side receiving unit 201, the data dividing unit 202, the data center side transmitting unit 203, the control unit 204, A PC-side transmitting unit 205, a data combining unit 206, and a data center-side receiving unit 207 are included.

  The PC-side receiving unit 201 receives data transmitted from the user terminal 101. The data dividing unit 202 divides the data received by the PC-side receiving unit 201 according to the data division size designated by the control unit 204.

  The data center side transmission unit 203 transmits the data divided by the data division unit 202 in parallel to the remote data center in accordance with a predetermined transfer order table 208 (see FIG. 3A). At this time, all the divided data may be transmitted in parallel or a part of the divided data may be transmitted in parallel depending on the capability of the data center side transmission unit 203 (network interface 1023). Then, the data center side transmission unit 203 notifies the control unit 204 of the transmission result of the divided data.

  The control unit 204 has an interface function with the maintenance terminal 200, holds the data division size set from the maintenance terminal 200, and stores the divided data transmission destination, the original file management table 209, and the divided data transmission destination The transfer order table 208 set from the maintenance terminal 200 is stored. The control unit 204 updates the original file management table 209 (see FIG. 3B) based on the transfer record transmitted from the data center side transmission unit 203.

  The data center side receiving unit 207 receives in parallel the divided data transmitted from the remote data center. At this time, depending on the capability of the data center side receiving unit 207 (network interface 1023), all the divided data may be received in parallel, or some of the divided data may be received in parallel. Further, referring to the original file management table 209 held in the control unit 204, reception of data from an unregistered data center may be rejected. Further, when the size of the divided data is registered in the original file management table 209, only the divided data having the correct size may be received.

  The PC-side receiving unit 201, the data center-side transmitting unit 203, the PC-side transmitting unit 205, and the data center-side receiving unit 207 are configured with one or a plurality of network interfaces 1023. Note that one physical network interface 1023 may be logically divided. The data dividing unit 202 and the data combining unit 206 may be implemented by hardware or software executed by the processor 1021. Further, the control unit 204 is implemented by software executed by the processor 1021.

  FIG. 3A is a diagram illustrating the configuration of the transfer order table 208 according to the embodiment of this invention.

  The transfer order table 208 is held by the control unit 204, and a destination data center for each time zone such as a time zone A300 and a time zone B301 is registered. That is, in the transfer order table 208, a data center that is a data transmission destination is set in advance from the maintenance terminal 200.

  FIG. 3B is a diagram illustrating the configuration of the original file management table 209 according to the embodiment of this invention.

  In the original file management table 209, information on files stored in the data center W100 is registered. A divided arrangement destination 315 is included. The original file management table 209 is held in the control unit 204.

  The sender address 310 is the IP address of the user terminal 101 that has requested the file storage. An address used in the network may be used instead of the IP address.

  The file name 311 is the name of the stored file. The size 312 is the size of the stored file. The file storage location 322 is a location (path) where the file is stored. The update date and time 313 is the time when the file is stored. CRC 324 is an error detection code of the stored file.

  In the completion flag 314, “0” or “1” is stored, “0” means start of transmission of divided data, and “1” means completion of transmission of divided data. The division arrangement destination 315 is an identifier of a data center in which data obtained by dividing the file is stored, and is registered in a column corresponding to the number of divisions (transfer order). Note that the size of the divided data may be registered in the divided arrangement destination 315.

  Of the data stored in the original file management table 209, the transmission source address 310, the file name 311, the size 312, the file storage location 322, the update date and time 313, and the divided arrangement destination 315 are transmitted from the data center side transmission unit 203. Based on the transfer record, the control unit 204 updates.

  FIG. 3C is a diagram illustrating the configuration of the divided file management table 210 according to the embodiment of this invention.

  In the divided file management table 210, information on the divided file is registered, and a transmission source data center 316, a divided file name 317, a size 318, a file storage location 323, a divided file number 319, an update date 320, and a completion flag 321 are displayed. Including. The divided file management table 210 is held by a remote data center where the divided files are stored.

  The transmission source data center 316 is an identifier of a data center that has transmitted the divided data (that is, stores a file from which the divided data is divided). The divided file name 317 is a file name of the divided data. The size 318 is the size of the divided data. The file storage location 323 is a location (path) where the divided data is stored. The division file number 319 is the order of the division data in the original file. The update date and time 320 is the time when the divided data is stored. In the completion flag 321, “0” indicating start of divided data transmission or “1” indicating completion of divided data transmission is set.

  Of the data stored in the divided file management table 210, the transmission source data center 316, the divided file name 317, the size 318, the file storage location 323, the divided file number 319, the update date 320, and the completion flag 321 are transmitted on the data center side. Based on the transfer record transmitted from the unit 203, the control unit 204 updates.

  Next, the operation of the network system according to the present embodiment will be described with reference to FIGS. 4A, 4B, and 5. FIG.

  4A and 4B are sequence diagrams of processing for storing data in the local data center and dividing and backing up data to another data center in accordance with a file storage request from the user terminal 101. FIG.

  First, when there is a file saving request 400 from the user terminal 101 to the data center W100, the data center W100 saves the file requested to be saved in the storage device in its own data center (401), and then the saving is requested. The CRC of the selected file is calculated (402). An error detection method and error correction method other than CRC may be used.

  Then, it is determined whether the same file name as the file name of the file requested to be saved is already registered in the original file management table 209 (403). Specifically, it is determined whether or not the file name of the file requested to be saved is already registered in the original file management table 209, and when the file name is already registered in the original file management table 209. Then, it is determined whether or not the size and CRC of the file are the same as the size 318 and CRC 324 registered in the original file management table 209.

  As a result, when it is determined that the file name, size, and CRC are the same, a message for confirming whether or not overwriting is necessary is transmitted to the user terminal 101, and it is determined whether or not to overwrite according to the user's instruction ( 404).

  As a result, when it is determined that the file is to be overwritten, the file is overwritten by registering the data (size 318 and update date / time 320) of the new file in the line of the file name in the original file management table 209. On the other hand, if the file is not overwritten, the file with the same name is managed as a different file according to the update date.

  If the same file name as the file name of the file requested to be saved is not registered in the original file management table 209, the IP address of the user terminal 101 is registered in the source IP address 310 of the original file management table 209. The file name 311, the size 312, the file storage location 322, and the update date / time 313 of the requested file are registered in the original file management table 209 (405). Next, '0' indicating the start of divided data transmission is set in the completion flag 314 in the file name row (406).

  Further, the file size of the file is acquired (407). The file size is stored in the header of the file and / or included in the file save request.

  Then, the number of file divisions is determined according to the division data size preset in the control unit 204 from the maintenance terminal 200 (408). Specifically, the number of divisions can be determined by dividing the file size obtained in step 407 by a preset division data size.

  Next, referring to the transfer order table 208, the data center that is the destination of the divided file at the current time is determined according to the determined number of divisions (409).

  In steps 405 to 407, since the CRC, the number of divisions, and the transmission destination are determined, after registering the CRC 324 and the division arrangement destination 315 in the original file management table 209 (410), the data center W100 sets the file in the specified size. Divide (411). The specified size may be set so that the divided data has the same size. In addition, the maximum size that can be transferred to another data center may be used. In this case, the divided data is generated by dividing the file at the specified size from the beginning of the division source file.

  Further, a sequential division file number according to the order registered in the transfer order table 208 is added after the original file name to determine the file name of the divided data.

  Next, the divided data is transmitted as a divided backup file together with the divided file number to the data center determined as the data transmission destination (412). The divided file number and the divided backup file may be transmitted in different packets, or may be transmitted by including the divided file number in the header of the divided backup file. When the divided file number and the divided backup file are transmitted in different packets, the consistency check between them is further strengthened.

  Finally, '1' indicating completion of divided data transmission is set in the completion flag 314 in the row of the file name in the original file management table 209 (413).

  Next, a process in which another data center that is a transmission destination of divided data stores the received divided data will be described.

  Upon receiving the divided file number and the divided file, the data center X 105 that is the transmission destination of the divided data determines whether the number included in the divided file name matches the divided file number (414).

  As a result of the determination, if the divided file number and the number included in the divided file name do not match, there is a possibility that the divided file to be received is incorrect. Therefore, the data center W100 is requested to retransmit the divided file and the divided file number (415). On the other hand, if the division file number and the number included in the division file name match, the identifier of the data center that is the transmission source of the division data is registered in the transmission source data center 316 of the division file management table 210, and the division file name is registered. 317 is registered in the divided file management table 210 (416).

  Next, '0' indicating the start of saving of the divided data is set in the completion flag 321 of the row of the divided file name (417). Next, the file size of the divided file and the divided file number are acquired from the received data (418, 419), and the divided data is saved (420).

  Then, the size 318, the file storage location 323, the divided file number 319, and the update date / time 320 are registered in the divided file management table 210 (421). Finally, '1' indicating completion of saving of the divided data is set in the completion flag 321 of the line of the file name (422).

  The process of the data center X105 that has received the divided data has been described above, but the other data center (for example, the data center Y108) that has received the divided data also executes the same process.

  By the procedure described above, the storage to the own data center and the divided backup to another data center by the file storage request from the user terminal 101 are completed.

  Next, processing for a file reference request will be described.

  FIG. 5 shows confirmation of data stored in its own data center, transmission request of divided backup data to another data center, and divided data in response to a file reference request from the user terminal 101 in the embodiment of the present invention. It is a sequence diagram of the process of reconstruction.

  First, when there is a file reference request (500) from the user terminal 101 to the data center W100, the server 102 of the data center W100 refers to the original file management table 209 in its own device (501), and the reference is requested. It is determined whether the file is stored in the data center W100 based on whether a file name that matches the file name exists on the table (502). If the file exists, the file is read from the storage apparatus 103, and the size and CRC of the read file are confirmed.

  If both the size and CRC are the same as the values stored in the original file management table 209, it is determined that the file is normal, and the file read from the storage apparatus 103 is transmitted to the user terminal 101 (503). On the other hand, if at least one of the size and CRC does not match, it is determined that the file is damaged, and a divided backup file is obtained from another data center by the following procedure.

  First, when the server 102 determines that the file is damaged, the server 102 refers to the divided arrangement destination 315 of the original file management table 209, and determines the data center that is the divided arrangement destination of the file and each identified data center. The division file number being backed up is acquired (504). The server 102 generates a divided file name from the divided file number and the file name, transmits the file name and the divided file number of the divided file to the data center which is the divided backup of the file, and requests transmission of the divided data ( 505). In this embodiment, since the divided file number is included in the divided file name, if the divided file name is transmitted, the requested file can be specified without transmitting the divided file number.

  The server 106 of the data center X 105 that has received the divided data transmission request refers to the divided file management table 210, and the request source data center matches the sending source data center 316, or the requested divided file name is the divided file name 317. Or whether the divided file number matches the divided file number 319 (506). Thereafter, when it is confirmed that all data matches the data registered in the divided file management table 210, the divided file is read from the storage apparatus 107 and transmitted to the data center W100 (507). On the other hand, when it is determined that any of the data does not match, a response indicating that the requested divided file cannot be transmitted is transmitted to the data center W100.

  The server 102 of the data center W100 determines whether the file name of the divided data transmitted from the data center X 105 matches the requested file name. Then, it is determined whether the number at the end of the file name of the divided data matches the divided file number, or whether the portion other than the number at the end of the file name of the divided data matches the original file name. Then, after the above-described determination for all received divided files, it is determined whether all the divided files are prepared. Specifically, the original file management table 209 is referred to, and it is determined whether or not all the divided files from the divided arrangement destination data center registered in the original file name line have been received. If these determinations confirm that the received divided file is correct, the divided files are combined (508).

  Then, the size of the combined file is compared with the size 312 of the original file management table 209, and it is determined whether or not the size of the combined file is correct. Further, the CRC of the combined file is calculated, and it is determined whether or not the calculated CRC matches the CRC 324 of the original file management table 209 (509).

  If the size and CRC are correct, the original data file has been correctly reconstructed, and the combined file is saved in the storage apparatus 103 instead of the lost file (510). As a result, the damaged file can be complemented. On the other hand, if at least one of the size and CRC is not correct, the system administrator is notified that the system has an abnormality because the original data file has not been correctly reconstructed (511).

  Thereafter, the reconstructed file is transmitted to the user terminal 101 (509), and the process is terminated.

  According to the first embodiment of the present invention, large data is also divided into preset data sizes, the number of data centers matching the data amount of the original file is selected, and the selected data center is parallel to the selected data center. Since the divided data is transmitted to perform backup, the extension of the backup time can be suppressed even when the data size is large. Further, since the data center for transmitting the divided data is changed depending on the time zone, an appropriate transmission destination data center can be selected according to the load on the network and the data center. Furthermore, since the data is divided and backed up, the confidentiality of the data can be improved.

<Embodiment 2>
Next, a second embodiment of the present invention will be described.

  In the first embodiment described above, the transfer order table 208 is set from the maintenance terminal 200. However, the task of registering an appropriate data center according to the time zone is complicated, and it is difficult to select and register an appropriate data center for the change of the traffic volume of the data center according to time and date.

  Therefore, in the second embodiment, based on the multiple regression analysis result using the statistics of traffic data of the data center and the related information (information affecting traffic), the transfer destination data center of the divided data is transferred to the transfer order table. Set to 208. However, in the second embodiment, the transfer order table 208 is generated using the traffic data between the transmission source data center and the transmission destination data center of the divided data. Therefore, when it is necessary to consider the traffic data between the destination data center and another data center, a multiple regression analysis including these data may be performed. It may be set.

  Note that the second embodiment is different only in the processing in which the server 102 sets the transfer order table 208, and the configuration and processing of other devices are the same as those in the first embodiment described above. Description of is omitted.

  As in the first embodiment, the server 102 according to the second embodiment includes a PC-side receiving unit 201, a data dividing unit 202, a data center-side transmitting unit 203, a control unit 204, and a PC-side transmitting unit 205. The data combination unit 206 and the data center side reception unit 207, and the control unit 204 executes the setting process of the transfer order table 208. This setting process is implemented by the processor executing a program stored in the memory.

  Further, a traffic amount measuring unit is provided in the network interface of the server 106 of the data center X105, the server 109 of the data center Y108, and the server 112 of the data center Z111, and measures the amount of data received by each data center.

  FIG. 6A is a sequence diagram of setting processing of the transfer order table 208 according to the second embodiment.

  First, as shown in FIG. 6A, the server 102 of the data center W100 periodically acquires the traffic volume measured by another data center, and records the acquired traffic volume in the traffic data tabulation table (611).

  Specifically, the statistical values of the past traffic volume for each data center that is a candidate for the destination of the divided data are aggregated in a predetermined format and recorded in the traffic data aggregation table 601 for each data center.

  As shown in FIG. 6B, the traffic data tabulation table 601 includes a date 602, a day of the week 603, a time zone 604, a traffic volume 605, a weather 606, a holiday 607, an event 608, and a flag 609. The day of the week 603, the time zone 604, the weather 606, and the holiday 607 are empirically factors that affect the traffic volume. In addition, a flag 609 is provided for registering, as necessary, events that cause fluctuations in traffic volume that cannot be covered by these items, and it is necessary to transmit divided data at a specific date and time for this data center. Used in any case. The items in the traffic data summary table 601 are not limited to the items described above, and items that may affect the traffic volume may be used.

  The traffic amount 605 may be an average traffic amount in the past time zone or a traffic amount at a certain point in the time zone, and a statistically determined traffic amount can be used.

  Note that the traffic data totaling table 601 updated by this process can be corrected from the maintenance terminal 200 at any time as needed (612).

  Next, when it becomes necessary to automatically generate the transfer order table 208, the system administrator instructs the control unit 204 to automatically generate the transfer order table (613).

Upon receiving the transfer order table automatic generation instruction 613, the control unit 204 performs a multiple regression analysis with reference to the traffic data summary table 601 for the data center Y generated in step 611, and performs a multiple regression analysis for each data center on a certain day. A regression analysis formula (the following formula) is generated (614). Specifically, referring to the traffic data summary table 601 for each data center, data of the past specific date (January 11) is extracted, the traffic volume is set as a target variable, and other variables are set as explanatory variables. Perform multiple regression analysis. There are various methods for obtaining the coefficient of the explanatory variable in the multiple regression analysis method, such as a variable reduction method and a variable increase method, and any method can be adopted.

Traffic volume (expected value) = coefficient A × traffic volume + coefficient B × weather + coefficient C × holiday + coefficient D × event + coefficient E × flag

  Next, using the multiple regression analysis formula for each data center generated in step 614, an expected value of the traffic volume on the target day is calculated (615). Specifically, information such as the day of the week on which the predicted value is to be calculated, the time zone, the weather, holidays, and the presence / absence of an event is substituted into the multiple regression analysis formula to calculate the predicted value of traffic volume (615).

  Thereafter, the predicted traffic volume values for the respective data centers calculated in step 615 are compared, and the data center identifiers are registered in the transfer order table 208 in ascending order of the predicted traffic volume values (616).

  Finally, a transfer order table creation completion notification is issued from the control unit 204 to the maintenance terminal 200 (617), and the transfer order table setting process is completed.

  As described above, in the second embodiment of the present invention, the order of transfer is based on the results of multiple regression analysis using the statistical traffic data amount for each past data center, information that affects the traffic data amount, and the like. By setting the table 208, it is possible to flexibly correspond to the actual traffic amount in the order of the data center to which the divided data is transferred.

100, 105, 108, 111 Data center 101 User terminal 102, 106, 109, 112 Server 103, 107, 110, 113 Storage device 104 Network 200 Maintenance terminal 201 PC side receiving unit 202 Data dividing unit 203 Data center side transmitting unit 204 Control unit 205 PC side transmission unit 206 Data combination unit 207 Data center side reception unit 208 Transfer order table 209 Original file management table 210 Split file management table

Claims (12)

A computer system comprising a plurality of data centers including at least a first data center, a second data center, and a third data center,
The data center includes a server including a processor, a memory, a storage interface, and a network interface, and a storage system including a storage control device, a storage device, and an interface.
The server of the first data center is connected to a user terminal;
The server of the first data center is
A control unit for holding transfer order information in which a priority order of data transfer to another data center storing a copy of a file to be written to the storage system and information on a division size of the file requested to be written; ,
A data dividing unit for dividing data requested to be written by the user terminal;
A transmission unit for transmitting the divided data;
The controller is
When a file write request is received from the user terminal, the division number of the received data is determined based on the size of the file requested to be written and the division size information.
Referring to the transfer order information, determine a data center to transfer the divided data,
The data dividing unit divides the file requested to be written into the predetermined number of data,
The transmission system transmits the divided data to the determined data center in parallel. The controller is
Holding file management information in which information of the file requested to be written is registered;
2. The computer system according to claim 1, wherein after the divided data is transmitted, a data center of a transfer destination of the divided data is registered in the file management information.   The file management information includes an address of a user terminal that has requested data writing, a name of a transferred file, and an identifier of a data center to which the divided data is transferred. Computer system. The server of the first data center is
When receiving a file reference request from the user terminal, it is determined whether the file requested for reference can be provided from the storage device of the first data center;
If it is determined that the data cannot be provided from the storage device of the first data center, the data center storing the data obtained by dividing the file requested for reference is specified;
Request the divided data from the identified data center;
Reconstructing the file for which the reference was requested by combining the received split data;
The computer system according to claim 1, wherein the reconstructed file is transmitted to the user terminal.   The computer system according to claim 1, wherein the transfer order information includes a priority order of transfer to different data centers according to a time zone.   The said control part produces the said transfer order information using the multiple regression analysis result based on the data amount transferred to each past data center, and the information which influences the said data amount. The computer system described. A server provided in a data center connected to at least two other data centers,
With processor, memory, storage interface and network interface,
Connected to the storage system via the storage interface, connected to the user terminal,
A control unit for holding transfer order information in which a priority order of data transfer to another data center storing a copy of a file to be written to the storage system and information on a division size of the file requested to be written; ,
A data dividing unit for dividing data requested to be written by the user terminal;
A transmission unit for transmitting the divided data;
The controller is
When a file write request is received from the user terminal, the division number of the received data is determined based on the size of the file requested to be written and the division size information.
Referring to the transfer order information, determine a data center to transfer the divided data,
The data dividing unit divides the file requested to be written into the predetermined number of data,
The server, wherein the transmission unit transmits the divided data to the determined data center in parallel. The controller is
Holding file management information in which information of the file requested to be written is registered;
8. The server according to claim 7, wherein after the divided data is transmitted, a data center to which the divided data is transferred is registered in the file management information.   9. The file management information includes an address of a user terminal that has requested data writing, a name of a transferred file, and an identifier of a data center to which the divided data is transferred. Server. The server
When receiving a file reference request from the user terminal, it is determined whether the file requested for reference can be provided from the storage device of the first data center;
When it is determined that data cannot be provided from the storage device of the data center that has received the file reference request from the user terminal, the data center storing the data obtained by dividing the file requested for reference is identified,
Request the divided data from the identified data center;
Reconstructing the file for which the reference was requested by combining the received split data;
The server according to claim 7, wherein the reconstructed file is transmitted to the user terminal.   8. The server according to claim 7, wherein in the transfer order information, a priority order of transfer to different data centers is determined according to a time zone.   The said control part produces the said transfer order information using the multiple regression analysis result based on the data amount transferred to each past data center, and the information which affects the said data amount, The transfer order information is characterized by the above-mentioned. The listed server.

Images