JP5494123B2 - Differential backup system, management server and client, differential backup method and program - Google Patents

Description

  The present invention relates to a differential backup system, a management server and a client, a differential backup method, and a program, and more particularly to a differential backup system and the like that can optimize a differential backup block size and complete a differential backup in a short time.

  In today's world, when computers or networks where computers are connected to each other are an important infrastructure as a foundation for companies and society, data loss is a socially important loss. Therefore, data backup is becoming increasingly important. Servers and business terminals that operate business services are usually operated to back up important data or the entire disk in preparation for an unexpected accident.

  At that time, it is required not to stop the operation of the computer, or even if it stops, it is required to restart the operation with the shortest possible stop, so the backup operation must be completed in the shortest possible time There is. Further, since the volume of data handled by each computer is becoming enormous, the volume of backup data to be stored needs to be as small as possible. This necessity is particularly high when the computer is operating some kind of business service.

  To reduce the time required for backup and reduce the amount of data to be saved, after performing a full backup of the entire data to be backed up, files and files that have changed since the full backup was performed In many cases, a differential backup that backs up only the area is performed.

  For example, by performing a full backup once a week and performing a differential backup for this full backup data at the end of each day, it is possible to reproduce the data at the end of each day, rather than taking a full backup every day. The capacity of stored data can be significantly reduced.

  There are the following technical documents related to data backup, particularly differential backup. Among them, Patent Document 1 describes a technique of performing differential backup by extracting differential data on the client side to be backed up and transmitting only the differential data to the server. Patent Document 2 describes a technique of predicting a variation in the data transfer amount of differential backup, thereby detecting and warning a backup work failure.

  Patent Document 3 describes a technique of copying data via a communication path, dividing data before and after duplication into blocks, comparing hash values for each block, and copying only the blocks that are different. . Patent Document 4 describes a technique for estimating a copy time when copying data and selecting an efficient data copy method (differential data copy or all data copy) according to the result. Patent Document 5 describes a technique of selecting an efficient backup method (file backup or volume backup) based on the block size to be copied.

JP 2004-094617 A JP 2004-206611 A JP 2005-100007 A JP 2006-268740 A JP 2008-299441 A

  In differential backup, full backup data and backup target data are divided into blocks having a predetermined block size at the same address, the data is compared between the blocks for each block, and only the different blocks are stored as differential data. For this reason, if the block size is changed, the time required for the differential backup process may differ. As described above, it is desirable to make the time required for backup as short as possible, so it is useful to adjust the block size in this way.

  However, Patent Documents 1 to 5 described above do not describe a technique that focuses on the adjustment of the block size. Naturally, the time required for differential backup and the data capacity cannot be reduced by adjusting and optimizing the block size with these techniques.

  In particular, an environment where the backup management server backs up data to be backed up from multiple client computers connected via a network, and an open environment where multiple types of hardware or software exist in the network. In a network environment, it is difficult to accurately estimate the time required for backup processing and shorten the time required. This problem cannot be solved by the aforementioned Patent Documents 1-5.

  An object of the present invention is to provide a differential backup system, a management server and a client, a differential backup method, and a program capable of accurately adjusting a block size in a differential backup process and reducing the required time. is there.

  In order to achieve the above object, a differential backup system according to the present invention includes a client, which is one or a plurality of computer devices that hold backup target data, and backup management that holds full backup data that is a complete backup of each backup target data. A differential backup system in which servers are connected to each other, wherein a backup management server performs first data transmission / reception with a first fixed storage means for storing full backup data and data communication with a client via a network And first hash data is calculated by dividing the full backup data into blocks at each of the first to third block sizes given in advance and applying a function given in advance to the data of each block. 1 hash data total And the first hash data in each of the first to third block sizes are divided into a test area that is a part of the full backup data and a remaining area other than the test area. A data generation unit for optimization processing that transmits hash data for the storage area to the client, the second data transmitting / receiving unit that performs data communication with the backup management server, and a second data that stores the backup target data And the second hash for calculating the second hash data by dividing the backup target data into blocks and applying a function to the data of each block for each of the first to third block sizes The data calculation unit and the first received the calculated second hash data from the backup management server for each block A hash data comparison unit for comparing with the hash data, a differential backup unit for performing differential backup of the block of the backup target data when the comparison result of the first and second hash data for each block is different, and the first to first For each of the three block sizes, the differential backup unit performs a test differential backup for the test area of the backup target data using the first hash data for the test area, and the first to third The optimal size calculation unit that determines the block size with the shortest required time in the experimental differential backup performed at the block size as the optimal size, and the first hash data for the remaining area in the case of the determined optimal size are backed up Requests to the management server and sent And a differential backup optimizing unit that makes the differential backup unit perform differential backup of the remaining area of the backup target data with the optimum size using the first hash data for the remaining area.

  In order to achieve the above object, a backup management server according to the present invention is connected to a client, which is one or a plurality of computer devices that hold backup target data, and performs full backup data that is a complete backup of each backup target data. A backup management server for holding a fixed storage means for storing full backup data, a data transmission / reception unit for performing data communication with a client via a network, and first to third block sizes given in advance. And a hash data calculator for calculating hash data by dividing the full backup data into blocks and applying a function given in advance to the data of each block, and hash data in each of the first to third block sizes Full back-up It has an optimization processing data creation unit that divides the data into a test area that is a part of the data and a remaining area other than that and transmits hash data about the test area to the client And

  In order to achieve the above object, a client according to the present invention is a client that is a computer device interconnected with a backup management server that holds backup target data and holds full backup data that is a complete backup of the backup target data. The data transmission / reception unit that performs data communication with the backup management server, the fixed storage unit that stores the data to be backed up, and each of these blocks by dividing the data to be backed up into blocks for each of the first to third block sizes A hash data calculation unit that calculates a second hash data by applying a function to the data of the hash data, and a hash data that compares the calculated second hash data with the first hash data received from the backup management server for each block A comparison unit; A differential backup unit that differentially backs up the block of the data to be backed up when the comparison results of the first and second hash data for each lock are different, and a backup for each of the first to third block sizes Using the first hash data for the test area that is part of the target data, the differential backup unit performs a test differential backup for the test area of the backup target data, and the first to third block sizes An optimal size calculation unit that determines the block size with the shortest required time in the experimental differential backup performed in step 1 as the optimal size, and the first hash data for the remaining area other than the test area in the case of the determined optimal size Is requested and sent to the backup management server And having the a first differential backup optimization unit hash data using to perform differential backup of the residual area of the backup target data in the optimal size differential backup portion for the remaining area.

  In order to achieve the above object, a differential backup method according to the present invention includes a client, which is one or a plurality of computer devices that hold backup target data, and backup management that holds full backup data that is a complete backup of each backup target data. In the differential backup system interconnected with the server, the first hash data calculation unit of the backup management server divides the full backup data into blocks at each of the first to third block sizes given in advance, The first hash data calculation unit of the backup management server calculates the first hash data by applying a function given in advance to the data of each block, and the first for each of the first to third block sizes. Backup of hash data The optimization processing data creation unit of the logical server divides the data into the test area that is part of the full backup data and the remaining area, and the hash data for the test area is stored in the backup management server. The optimization processing data creation unit transmits to the client, and for each of the first to third block sizes, the second hash data calculation unit of the client divides the backup target data into blocks of the block size, The second hash data calculation unit of the client calculates the second hash data by applying the function to the data of the second hash data calculated for each block and the first hash data received from the backup management server Are compared by the client hash data comparison unit, and the first and When the comparison result of the second hash data is different, the differential backup unit of the client differentially backs up the block of the backup target data, and the differential backup for the test area performed with the first to third block sizes. The optimal size calculation unit of the client determines the block size with the shortest required time as the optimal size, and the differential backup optimization unit of the client performs backup management of the first hash data for the remaining area in the case of the determined optimal size. The request is sent to the server, and the data transmission / reception unit of the backup management server sends the first hash data for the requested remaining area to the client, and the client uses the first hash data for the remaining area to optimize the differential backup of the client. Is the differential backup unit A differential backup of the remaining area of the backup target data is performed.

  In order to achieve the above object, a differential backup program according to the present invention includes a client, which is one or a plurality of computer devices that hold backup target data, and backup management that holds full backup data that is a complete backup of each backup target data. In a differential backup system connected to a server, a procedure for dividing full backup data into blocks in each of first to third block sizes given in advance to a computer provided in the backup management server, Procedure for calculating hash data by applying a function given in advance to the data, and for each of the first to third block sizes, the hash data is for the test area that is part of the full backup data and Less than Of the remaining area, the procedure for transmitting the hash data for the test area to the client, and the procedure for transmitting the first hash data for the remaining area requested by the client to the client. It is characterized by that.

  In order to achieve the above object, another differential backup program according to the present invention holds a client, which is one or a plurality of computer devices that hold backup target data, and full backup data that is a complete backup of each backup target data. In the differential backup system interconnected with the backup management server, a procedure for dividing the backup target data into blocks of the block size for each of the first to third block sizes in the computer provided in the client, A procedure for calculating the second hash data by applying a function to the block data, a procedure for comparing the second hash data calculated for each block with the first hash data received from the backup management server, and for each block First A procedure for differentially backing up the block of the backup target data when the comparison results of the second hash data are different, and a test area that is a part of the backup target data performed in the first to third block sizes Of determining the block size with the shortest required time in the differential backup of the first as the optimum size, and requesting the backup management server for the first hash data for the remaining area other than the test area in the case of the determined optimum size And a procedure for performing a differential backup of the remaining area of the backup target data using the first hash data for the remaining area that has been sent.

  As described above, the present invention performs differential backup by dividing the test area into blocks with the first to third block sizes, and divides the remaining area into blocks with the block size having the shortest required time. Since it is configured to perform backup, the differential backup system, management server and client, and differential that can reduce the time required for differential backup by adjusting the block size accurately so that the required time can be shortened A backup method and program can be provided.

FIG. 3 is an explanatory diagram showing a more detailed configuration of the backup management server shown in FIG. 2. It is explanatory drawing which shows the structure of the differential backup system which concerns on this embodiment. It is explanatory drawing shown about the concept of the full backup data shown in FIG. 1, and difference data. FIG. 3A shows the full backup data, and FIG. 3B shows the difference data. It is explanatory drawing shown about the data structure of the hash data of the difference data shown in FIG. It is explanatory drawing which shows the content of each data memorize | stored in the working area of the backup management server shown in FIG. It is explanatory drawing shown about the structure of the client shown in FIG. 1, and a boot image. It is explanatory drawing shown about the data memorize | stored in the working area of the client shown in FIG. It is explanatory drawing which shows the outline | summary of the differential backup process which concerns on this embodiment. FIG. 8A shows “Step 1” as the first process, and FIG. 8B shows “Step 2” as the second process. It is a continuation of FIG. FIG. 9C shows “step 3” as the third process, and FIG. 9D shows “step 4” as the fourth process. It is explanatory drawing shown about the hash data produced at each process shown to FIGS. 10A shows hash data for “all areas” of the block size “current (512 kilobytes)”, FIG. 10B shows hash data for “all areas” of the block size “join (1024 kilobytes)”, FIG. (C) is hash data for “all areas” of block size “divided (256 kilobytes)”, FIG. 10 (d) is hash data for “test areas” of block size “current”, and FIG. 10 (e) is a block FIG. 10F shows hash data for the “test area” with the block size “division”, respectively. It is a continuation of FIG. FIG. 11G shows the hash data for the “remaining area” with the block size “current”, FIG. 11H shows the hash data for the “remaining area” with the block size “join”, and FIG. Hash data for “remaining area” in FIG. 3 is a flowchart showing optimization hash data calculation processing performed as a differential backup preparation processing by the optimization processing data creation unit shown in FIG. 1 in the backup management server. 13 is a flowchart showing hash data calculation processing performed by a hash data calculation unit shown as steps S302, 304, and 306 in FIG. It is a flowchart shown about the process of the optimization differential backup which the differential backup optimization part shown in FIG. 6 performed by the client. It is a flowchart shown about the optimal size calculation process which the optimal size calculation part shown as step S401 of FIG. It is a continuation of FIG. 16 is a flowchart illustrating differential backup processing by the differential backup unit and the hash data comparison unit illustrated as steps S452, 454, and 457 in FIG. 15 and S404, 405, and 406 in FIG. FIG. 17 is a flowchart showing processing at the backup management server corresponding to processing at the client shown in FIGS.

(First embodiment)
Hereinafter, the structure of the 1st Embodiment of this invention is demonstrated based on attached FIGS. First, the basic content of the present embodiment will be described, and then more specific content will be described.
The differential backup system 1 according to the present embodiment includes a client 21 that is one or a plurality of computer devices that hold backup target data 221 and a backup management server 10 that holds full backup data 121 that is a complete backup of each backup target data. Is a differential backup system connected to each other. The backup management server 10 includes a first fixed storage unit 103 that stores the full backup data 121, a first data transmission / reception unit 112 that performs data communication with the client 21 via the network 30, and a first given in advance. The first hash data calculation unit 114 calculates the first hash data 160 by dividing the full backup data 121 into blocks for each of the third block sizes and applying a predetermined function to the data of each block. The first hash data in each of the first to third block sizes is divided into a test area that is a part of the full backup data 121 and a test area that is divided into other areas. Data for optimization processing for transmitting hash data 164 to 166 for the area to the client And a generation unit 111. One client 21 includes a second data transmission / reception unit 211 that performs data communication with the backup management server 10, a second fixed storage unit 203 that stores backup target data 221, and each of the first to third block sizes. In this case, the backup object data 221 is divided into blocks, a second hash data calculation unit 215 that calculates a second hash data by applying a function to the data of each block, and the calculated second hash data The data to be backed up when the comparison result of the first and second hash data for each block differs from the hash data comparison unit 216 that compares the first hash data 164 to 166 received from the backup management server for each block. Differential backup to perform differential backup of the block For each of 214 and the first to third block sizes, a trial differential backup is performed in the differential backup unit for the test area of the backup target data using the first hash data 164 to 166 for the test area. And an optimal size calculation unit 213 that determines the block size with the shortest required time in the experimental differential backup performed with the first to third block sizes as the optimal size, and the determined optimal size Request the first hash data for the remaining area to the backup management server, and use the first hash data for the remaining area that has been sent to the differential backup unit to determine the remaining area of the backup target data with the optimum size. Differential backup optimization unit 212 that performs differential backup Having.

  Here, the second block size is larger than the first block size, the third block size is smaller than the first block size, and the optimization processing data creation unit 111 of the backup management server 10 determines the remaining area. The optimum size received when returning the first hash data to the client is stored as the first block size for the same processing next time.

  Further, if the time required for the experimental differential backup performed with the second block size is shorter than the time required for the experimental differential backup performed with the first block size, the optimum size calculation unit 213 of the client 21 at that time. The second block size is determined as the optimum size.

  The backup management server 10 causes the processor included in the client to function as a second data transmission / reception unit, a second hash data calculation unit, a hash data comparison unit, a differential backup unit, an optimal size calculation unit, and a differential backup optimization unit A boot image storage unit 117 that stores a boot image 141 as a program in the first fixed storage unit 103, and a boot image transmission unit 113 that transmits the boot image to a client to operate the client.

By providing this configuration, the differential backup system 1 can accurately adjust the block size, thereby shortening the time required for differential backup.
Hereinafter, this will be described in more detail.

  FIG. 2 is an explanatory diagram showing the configuration of the differential backup system 1 according to the present embodiment. The differential backup system 1 includes a backup management server 10 and a plurality of clients 21, 22,... Connected to each other via a network 30. The backup management server 10 is a computer device equipped with software for backing up data of the clients 21, 22,... Via the network 30.

  The clients 21, 22,... Are computer devices equipped with software for backing up data by the backup management server 10. Hereinafter, any one of these clients will be referred to as a client 21 as a representative. One or more clients 21 may be provided for one backup management server 10. The network 30 is a LAN (Local Area Network) or a WAN (Wide Area Network) for connecting the backup management server 10 and the clients 21, 22,.

  FIG. 1 is an explanatory diagram showing a more detailed configuration of the backup management server 10 shown in FIG. The backup management server 10 is a computer device that includes a processor 101, a temporary storage unit 102, a fixed storage unit 103, and a communication unit 104. The temporary storage means 102 is a storage device such as a RAM (Random Access Memory) that stores data that the processor 101 is working on, and the fixed storage means 103 is a storage device such as a hard disk that stores a large amount of data. is there.

  In the processor 101, an optimization processing data creation unit 111, a data transmission / reception unit 112, a boot image transmission unit 113, a hash data calculation unit 114, a full backup data storage unit 115, a difference data storage unit 116, and a boot image storage unit 117 Executed as a computer program. The optimization processing data creation unit 111 creates hash data that is used by the client 21 for optimization of differential backup in the processing described later. The data transmission / reception unit 112 controls the communication unit 104 to perform data communication with the client 21. The boot image transmission unit 113 transmits the boot image to the client 21 to operate. The hash data calculation unit 114 calculates a hash value for each block obtained by dividing full backup data by an arbitrary size.

  In addition, a full backup data storage area 120, a differential data storage area 130, and a boot image storage area 140 are secured on the fixed storage unit 103, and a full backup data storage unit 115 is provided in the full backup data storage area 120. The difference data storage unit 116 stores the full backup data 121 and the difference data 131 in the difference data storage area 130. The boot image storage unit 117 stores the boot image 141 in the boot image storage area 140. Here, the full backup data storage unit 115, the differential data storage unit 116, and the boot image storage unit 117 can use a DBMS (Database Management System) that stores and manages a large amount of data.

  The communication unit 104 performs data communication with the client 21 via the network 30. The temporary storage unit 102 has a working area 150 that is an area for temporarily storing data being worked on by the optimization processing data creation unit 111. The working area 150 stores hash data 160, which will be described later, and a current size setting value 151, which will be described later.

  FIG. 3 is an explanatory diagram showing the concept of the full backup data 121 and the difference data 131 shown in FIG. 3A shows the full backup data 121, and FIG. 3B shows the difference data 131. The full backup data 121 shown in FIG. 3A corresponds to the backup target data 221, that is, a full copy of the disk area of the client 21, and is a full copy of data from address 0 to the maximum address N of the disk.

  On the other hand, the difference data 131 shown in FIG. 3B is the difference between the full backup data 121 of the previous collection and the current backup target data 221 in the disk area of the client 21. Only a certain part is copied.

  Assuming that the size of the entire full backup data 121 is N, both the full backup data 121 of the previous collection and the current backup target data 221 are divided into blocks with the same block size n, and the data is completely the same in each block. If there is a difference, “no difference” is set, and if there is a difference, “difference exists”, and the head address of the block that is “with difference” is stored in a temporary list 260 described later. Then, only the block whose head address is stored in the temporary list 260 is backed up as “difference data”.

  As shown in FIG. 3B, the difference data 131 is composed of data such as a size 131a corresponding to the block size to be divided, an address 131b corresponding to the head address of the block, and data 131c corresponding to the actual block data. The In FIG. 3B, the “no difference” block is blank, and the “difference” block is indicated by a hatched pattern. If the first block is 0th, the first and ath blocks (a is a natural number, and 2 ≦ a <k−1, k = N / n) become “difference” and become differential data 131. An example of being backed up is shown.

  Here, the optimization processing data creation unit 111 calculates the hash data 160 of each block and stores it in the working area 150. FIG. 4 is an explanatory diagram showing the data structure of the hash data 160 of the difference data 131 shown in FIG. The hash data 160 is data obtained by collecting hash values for each block obtained by dividing the backup target data by the block size n. The hash data 160 is a table storing a block size 160a corresponding to the block size n to be divided, a head address 160b of each block, and a hash value 160c of the block.

  The hash value here means a fixed-length value obtained by summarizing arbitrary data. For example, a value obtained by applying a function uniquely determined as a summary to arbitrary data, such as the GetHashCode function in the C # language, to the data. This function is called a hash function. In the present embodiment, the comparison of the data for each block shown in FIG. 3B does not directly compare the block data itself, but determines whether or not there is a difference by comparing the hash values of the block data. doing.

  In this embodiment, as will be described later, the block size “current”, “combine”, “partition”, and hash value calculation targets are “all areas”, “test areas”, and “remaining areas” in FIG. Since the data shown is created, the hash data 160 is a generic term for them, and each block size and calculation object is called hash data 161-169.

  The optimization processing data creation unit 111 and the hash data calculation unit 114 calculate the hash data 160 as a differential backup preparation process while referring to the backup data storage area 120. The data transmission / reception unit 112 transmits the hash data 160 to the client 21 and receives differential data from the client 21. Details of these processes will be described later.

  FIG. 5 is an explanatory diagram showing the contents of each data stored in the working area 150 of the backup management server 10 shown in FIG. In the working area 150, hash data 160 (161 to 169) created by the optimization processing data creation unit 111 and the hash data calculation unit 114 (shown in FIG. 4) and a current size setting value 151 are stored. Yes.

  FIG. 6 is an explanatory diagram showing the configuration of the client 21 and the boot image 141 shown in FIG. The boot image transmission unit 113 of the backup management server 10 transmits the boot image 141 to the client 21 in order to execute the differential backup of the client 21. By transmitting the boot image 141 to the client 21 and restarting the client 21, various programs operate on the processor 201 of the client 21 and differential backup is executed.

  Similar to the backup management server 10, the client 21 is a computer device that includes a processor 201, a temporary storage unit 202, a fixed storage unit 203, and a communication unit 204. In the processor 201, each of the data transmission / reception unit 211, the differential backup optimization unit 212, the optimal size calculation unit 213, the differential backup unit 214, the hash data calculation unit 215, and the hash data comparison unit 216 operates. The boot image 141 is a computer program for causing the processor 201 to function as these operation units.

  The boot image 141 may be a server program that complies with the PXE (Preboot eXecution Environment) protocol defined in RFC (Request For Comments) 4578, for example, or the boot image is transferred to a so-called client program operating on the client 21 by a push method. You may send in. Since transmitting the boot image 141 to the client 21 and restarting the client 21 with the boot image is a known technique, it will not be described in detail.

  The fixed storage unit 203 has a backup target data storage area 220 in which backup target data 221 is stored. In the temporary storage unit 202, a working area 250, which is an area for temporarily storing data that the differential backup optimization unit 212 is working, is secured. The working area 250 stores hash data, processing time, and optimum size necessary for differential backup processing. Details of this will be described later.

  The data transmission / reception unit 211 receives the hash data 164 to 166 from the backup management server 10 and transmits the differential data created on the client 21 to the backup management server 10. The differential backup optimizing unit 212 calculates a block size that minimizes the differential backup processing time, and performs differential backup processing with that block size. The optimum size calculation unit 213 tries differential backup for the test area of the backup target data 221. The differential backup unit 214 performs actual differential backup processing. The hash data calculation unit 215 calculates hash data for the backup target data 221. The hash data comparison unit 216 compares the calculated hash data with that received from the backup management server 10. Details of each of the above processes will be described later.

  FIG. 7 is an explanatory diagram showing data stored in the working area 250 of the client 21 shown in FIG. In the working area 250, the “test area” hash data 164 to 166 received from the backup management server 10 and the block sizes “current”, “combined”, and “divided” are performed on the “test area”. The backup backup required times 251 to 253, the optimum block size 254 determined by the optimum size calculation unit 213 therefrom, and the temporary list 260 used during the backup operation (shown in FIG. 3B) are stored. The

  In the working area 250, one of the hash data 167 to 169 of the “remaining area” of the block size determined as the “optimal size” in the processing described later among the block sizes “current”, “combine”, and “division”. One is also stored for the “remaining area” backup operation.

  8-9 is explanatory drawing which shows the outline | summary of the differential backup process which concerns on this embodiment. This differential backup process consists of four steps of steps 1 to 4. 8 to 9 are divided into two drawings for the sake of space. Among them, FIG. 8A showing the first step “Step 1” shows the full backup data 121 in the backup management server 10 and the backup target data 221 of the client 21. In the full backup data 121, a specific area and capacity are set as a test area 121a, and other areas are set as remaining areas 121b. Correspondingly, the same area of the backup target data 221 is set as a test area 221a and a remaining area 221b.

  Among them, the entire full backup data 121 is divided by the block size n and the hash value for each block is calculated. The block size n at that time is set to “current size (hereinafter simply referred to as current)” “current × 2”. Three types of “current / 2” are prepared, and in each case, the hash data calculation unit 114 calculates hash values and creates hash data 161 to 163 for “all areas”.

  Although the current can be determined arbitrarily, in this embodiment, 512 kilobytes is set as the current size setting value 151, that is, the current default value. Hereinafter, the respective block sizes are referred to as “current”, “combination”, and “division”, and the hash value calculation targets are referred to as “all areas”, “test areas”, and “remaining areas”.

  In FIG. 8B showing “step 2” as the second process, a specific area and capacity are set as test areas in the full backup data 121 and the backup target data 221, and the remaining areas are the remaining areas. And Then, the optimization processing data creation unit 111 divides the hash data 161 to 163 calculated in Step 1 into hash data 164 to 166 on the test area 121a and hash data 167 to 169 on the remaining area 121b.

  The size of the test area can be arbitrarily determined. In this embodiment, 2048 kilobytes from the beginning are used as test areas, and the remaining areas are used as remaining areas. Then, the hash data 164 to 166 for the test area are transmitted to the client 21.

  In FIG. 9C showing “step 3” as the third process, the optimum size calculation unit 213 uses the block size of “current”, “join”, and “divide” for the test area 221a. Under the control, the hash data comparison unit 216 compares the hash values of the blocks while the differential backup unit 214 executes the differential backup, calculates the time required for actual processing in each case, and calculates the optimum size calculation unit 213. It is determined whether the required time is the shortest.

  In FIG. 9D showing the fourth step “Step 4”, any one of the hash data 167 to 169 for the remaining area 221b of the block size whose required time is the shortest in Step 3 is stored. It is transmitted from the backup management server 10 to the client 21. Under the control of the differential backup optimization unit 212, the hash data comparison unit 216 compares the hash value of each block, the differential backup unit 214 executes the differential backup for the remaining area, and completes the differential backup for the entire backup target data 221. To do.

  FIGS. 10-11 is explanatory drawing shown about the hash data 160 produced at each process shown to FIGS. The hash data 160 is for each block size whose block size 160a is "current", "combine", and "divided", and for each "all area" and for each of "test area" and "remaining area". In the nine ways, the head address 160b of each block and the hash value 160c of that block are stored.

  FIG. 10A shows hash data 161 for “all areas” of the block size “current (512 kilobytes)”, and FIG. 10B shows hash data 162 for “all areas” of the block size “join (1024 kilobytes)”. FIG. 10C shows hash data 163 for “all areas” with a block size “partition (256 kilobytes)”, and FIG. 10D shows hash data 164 for “test areas” with a block size “current”. e) is the hash data 165 for the “test area” for the block size “join”, FIG. 10F is the hash data 166 for the “test area” for the block size “divided”, and FIG. Hash data 167 for “remaining area” of “current”, FIG. Hash data 168 for the "remaining region", FIG. 11 (i) shows the hash data 169 block size of the "split" to "remaining area", respectively.

(Optimization hash data calculation processing)
FIG. 12 is a flowchart showing optimization hash data calculation processing performed as the differential backup preparation processing by the optimization processing data creation unit 111 shown in FIG. 1 in the backup management server 10. First, when the current size setting value 151 is not set, the optimization processing data creation unit 111 sets an initial value in the system (step S301). Here, the initial value in the system is not defined in the present invention because an appropriate value should be set for each system in accordance with the area to be applied.

  When the current is set in the working area 150, the hash data calculation unit 114 is called by calling the hash data calculation unit 114 with “full backup data 121” as the calculation target data and “current” as the block size as inputs, and shown in FIG. The process is executed (step S302), and the output result is stored in the working area 150 as hash data 161 for “all areas” of the block size “current” (step S303).

  Subsequently, the hash data calculation process shown in FIG. 13 is executed by calling the hash data calculation unit 114 with “full backup data 121” as the calculation target data and “current × 2 (combination)” as the block size (step S304). The output result is stored in the working area 150 as hash data 162 for “all areas” of the block size “join” (step S305).

  Similarly, the hash data calculation process shown in FIG. 13 is executed by calling the hash data calculation unit 114 with “full backup data 121” as the calculation target data and “current / 2 (partition)” as the block size as inputs (step S306). The output result is stored in the working area 150 as hash data 163 for “all areas” of the block size “division” (step S307).

  Each of the hash data 160 obtained as described above is divided into the test area and the remaining area as shown in FIGS. 10 to 11 by the optimization processing data creation unit 111 (step S308). The process ends. This preparation process can be performed at an arbitrary timing. That is, it may be performed immediately before executing the differential backup, or may be performed for the next differential backup after executing the differential backup.

  FIG. 13 is a flowchart showing hash data calculation processing performed by the hash data calculation unit 114 shown as steps S302, 304, and 306 in FIG. The hash data calculation unit 114 is started by specifying the calculation target data and the block size from the optimization processing data creation unit 111, and based on this, the calculation target data is first divided into the block size (step S351). A hash value is calculated for each block (step S352), and hash data 160 that combines the block size 160a, the head address 160b, and the hash value 160c is output (step S353).

  When the backup management server 10 is instructed to perform differential backup by a user operation or the like in the state where the preparation processing of FIG. 12 has been performed, the boot image transmission unit 113 of FIG. 2 transmits the boot image 141 to the client 21, Thereafter, the client 21 is restarted. The restarted client 21 operates according to the boot image 141 shown in FIG. 6 and starts the optimized differential backup process.

(Optimized differential backup processing)
FIG. 14 is a flowchart illustrating optimized differential backup processing performed by the differential backup optimization unit 212 illustrated in FIG. The differential backup optimization unit 212 calls the optimal size calculation unit 213 to execute the optimal size calculation process (step S401).

  FIGS. 15 to 16 (shown divided into two for convenience of paper) are flowcharts showing the optimum size calculation processing performed by the optimum size calculator 213 shown as step S401 in FIG. First, the optimum size calculation unit 213 requests the backup management server 10 via the data transmission / reception unit 211 to request and receive hash data 164 to 166 for each block size (current, combined, and divided) of the test area (see FIG. Step S451).

  In response to this, the optimum size calculation unit 213 receives the block size “current” and the target “test area” as input, causes the differential backup unit 214 to execute differential backup processing (step S452), and sets the processing time to “current” “ It is stored in the working area 250 as the required test backup time 251 for “test area” (step S453).

  Subsequently, the optimum size calculation unit 213 receives the “join” as the block size and the “test area” as the target, causes the differential backup unit 214 to execute the differential backup process (step S454), and sets the processing time to “join” and “ It is stored in the working area 250 as the required test backup time 252 for “test area” (step S455).

  Here, the test backup time 251 of “current” “test area” and the test backup time 252 of “join” “test area” are compared (step S456). It is determined that “join” is the optimum block size, and this is stored as the optimum block size 254 in the working area 250 (step S462), and the optimum size calculation process is terminated.

  On the other hand, if the required backup time 251 for “current” and “test area” is shorter than the required backup time 252 for “join” and “test area” in step S456, then the block size is “divided”. The differential backup unit 214 is caused to execute the differential backup process by inputting “test area” as an object (step S457), and the processing time is set in the working area 250 as the required backup time 253 for “division” and “test area”. Store (step S458).

  Then, the test backup time 251 of “current” “test area” is compared with the test backup time 253 of “division” “test area” (step S459). If the latter is shorter, “division” is performed. Is the optimum block size, and is stored as the optimum block size 254 in the working area 250 (step S461). If the former is shorter, it is determined that “current” is the optimum block size, and this is stored as the optimum size in the working area 250 (step S460). Thereafter, the optimum size calculation process is terminated.

  In an environment where the backup target data of multiple client computers connected via a network is backed up collectively as shown in this embodiment, the data transfer speed on the network requires differential backup. It has the greatest impact on time.

  If the data transfer speed on the network is extremely fast and the effect on the processing time due to the data transfer speed can be ignored, differential data calculation (such as calculation and comparison of hash data) that is an overhead in backup processing should be performed as much as possible. Decreasing the number will shorten the time required for differential backup. Therefore, it is desirable to increase the block size as much as possible. It is most ideal to make the entire backup target data one block.

  However, in reality, the data transfer speed on the network often becomes an obstacle to backup work. Therefore, in this embodiment, as shown in FIGS. 15 to 16, the algorithm is configured to increase the block size as much as possible within a range where the influence of the data transfer rate can be ignored.

  If the required backup time 252 for “join” and “test area” is shorter than the required test backup time 251 for “current” and “test area” in step S456 described above, “join” is set to the optimum block size 254. Is based on this idea. In this case, since it can be determined that the data transfer rate is not in the range of an obstacle, the test backup for the block size “division” is not performed.

  Returning to FIG. 14, the differential backup optimization unit 212 determines the calculation result of the optimal block size by the optimal size calculation unit 213, that is, the optimal block size 254 (steps S 402 to S 403), and if this is “current” ( In step S402: YES), the data transmission / reception unit 211 requests the backup management server 10 to receive the hash data 167 for the block size “current” of the remaining area, receives it, calls the differential backup unit 214 as an input, and remains. An area differential backup process is executed (step S406).

  If the optimum block size 254 is “join” (step S402: NO, step S403: YES), the data transmitting / receiving unit 211 requests the backup management server 10 for the hash data 168 for the block size “join” of the remaining area. Receiving this, the differential backup unit 214 is called with this as an input, and the differential backup process of the remaining area is executed (step S405).

  When the optimum block size 254 is not “current” nor “combined” but “divided” (step S402: NO, step S403: NO), the data transmitting / receiving unit 211 notifies the backup management server 10 of the remaining area block size “divided”. The hash data 169 is requested and received, and is received as an input, and the differential backup unit 214 is called to execute the differential backup process for the remaining area (step S404).

  After the differential backup process is completed, the optimum block size 254 is transmitted to the backup management server 10 so as to be used as the current size in the next process (step S407), and the optimized differential backup process is thus completed. Although there is a process of merging the differential data 131 into the full backup data 121 after the optimized differential backup process, the merge process can be realized by a known technique and is not within the scope of the present invention. Will not be explained.

  FIG. 17 is a flowchart showing the differential backup processing by the differential backup unit 214 and the hash data comparison unit 216 shown as steps S452, 454, and 457 in FIG. 15 and S404, 405, and 406 in FIG. The differential backup unit 214 is activated by designating a block size and a backup target area from the differential backup optimization unit 212 or the optimal size calculation unit 213 and receiving the hash data 160 of that area.

  The activated differential backup unit 214 reads i-th area backup target data 221 with i = 1 as an initial value (step S502), calculates a hash value 251c of the area (step S503), and compares hash data The unit 216 compares this with the hash value 160c of the i-th area input by the hash data 160 (step S504). If they are different, the head address 160b of the block having different hash values 160c and 251c is shown in FIG. ) Is output as a temporary list 260 (step S505).

  The processes in steps S502 to S505 are repeated until the value of i is incremented by 1 until i reaches N / n (N is the total block size, n is the block size), that is, all the backup target data 221 It performs with respect to a block (steps S506-507). When the comparison of the hash values 160c and 251c is completed for all the blocks, the block data corresponding to the head address 160b output as the temporary list 260 is transmitted to the backup management server 10 as difference data (step S508). .

  FIG. 18 is a flowchart showing processing in the backup management server 10 corresponding to the processing in the client 21 shown in FIGS. First, in the backup management server 10 for which the hash data 164 to 166 of the test area is requested in the process of step S451 in FIG. (Step S601).

  Then, the data transmission / reception unit 112 receives the differential backup data performed using them from the client 21, and the differential data storage unit 116 stores it as the differential data 131 (step S602).

  Subsequently, the client 21 determines the optimum size, and requests hash data 167 to 169 of the remaining area with the optimum size in the processing of steps S404 to S406 in FIG. 14 (step S603). In the backup management server 10, the data transmitting / receiving unit 112 receives this, and transmits the hash data 167 to 169 corresponding to the request to the client 21 (step S604).

  Then, the differential backup data performed with the remaining area hash data 167 to 169 transmitted in step S601 is received from the client 21 by the data transmission / reception unit 112, and the differential data storage unit 116 stores it as differential data 131 ( Step S605).

  Thereafter, the differential data storage unit 116 discards the test differential data other than the optimum size received in the process of step S603 (step S606), and the optimization process data creation unit 111 sets the received optimum size to the current size setting value. 151 is stored in the working area 150 (step S607).

(Example of specific processing)
The above operation will be described with a more specific operation example. The backup target data 221 has 0 to 16384 kilobytes as the address range of the target area. It is assumed that each block size of “512 kilobytes” is set as the current size setting value 151 and “2048 kilobytes” is set as the test area size.

  The entire backup target data 221 is stored in the fixed storage means 103 of the backup management server 10 as full backup data 121. On the other hand, the operation of taking a differential backup of the backup target data 221 in the fixed storage means 203 of the client 21 will be specifically described below.

  First, as a differential backup preparation process, the optimization processing data creation unit 111 of the backup management server 10 performs the optimization hash data calculation process shown in FIG. The calculation target data is “full backup data 121” shown in FIG. Since “512 kilobytes” is set as the current size setting value 151, the hash data calculation unit 114 performs the hash data calculation processing shown in step S302 and FIG. 13, and the result shown in FIG. The hash data 161 for “all areas” of the block size “current (512 kilobytes)” is stored in the working area 150 (FIG. 12, step S303).

  Subsequently, the hash data calculation unit 114 performs the same processing for the block size “combined (1024 kilobytes)” and the block size “divided (256 kilobytes)”, and the hash data 162 for each block size “all areas” is obtained. And 163 are stored in the working area 150 (FIG. 12, steps S305 and 307).

  After calculating the hash data 160 of FIGS. 10A to 10C, the optimization processing data creation unit 111 performs hashing based on the size of the test area (2048 kilobytes) set in the system in advance. The data is divided (step S308), and as a result, the hash data 164 to 169 shown in FIGS. 10 (d) to 11 (i) are stored in the working area 150.

  When differential backup is started by an operation from the user or the arrival of a set time, the boot image transmission unit 113 transmits the boot image 141 to the client 21 and restarts the client 21. The restarted client 21 starts the boot image 141, and the processor 201 functions as each operation unit illustrated in FIG. Then, the differential backup optimization unit 212 starts the optimized differential backup process shown in FIG.

  When the optimized differential backup process is started, the differential backup optimization unit 212 calls the optimal size calculation unit 213 to execute the optimal size calculation process shown in FIGS. 15 to 16 (step S401). When the optimum size calculation process is started, the optimum size calculator 213 receives the hash data 164 to 166 for each block size (current, combined, divided) of the test area from the backup management server 10 (FIG. 15). Step S451) First, the differential backup unit 214 is called with the block size “current” and the target “test area” as inputs to execute the differential backup process (step S452 in FIG. 15).

  The differential backup processing time was 500 ms. The optimum size calculation unit 213 stores the processing time in the working area 250 as the required test backup time 251 for “current” and “test area” (step S453 in FIG. 15).

  Subsequently, the optimum size calculation unit 213 executes the differential backup processing in the differential backup unit 214 with “join” as the block size and “test area” as the target as input (FIG. 15, step S454). The processing time for this differential backup was 250 ms. The processing time is stored in the working area 250 as a test backup required time 252 for “join” and “test area” (FIG. 15, step S455).

  Here, when the required backup time 251 for “current” and “test area” is compared with the required backup time 252 for “join” and “test area” (step S456 in FIG. 15), the former is 500 ms and the latter. Was 250 ms, and the latter was shorter. Therefore, it is determined that “join” is the optimum block size, and this is stored as the optimum block size 254 in the working area 250 (step S462 in FIG. 15).

  Subsequently, the data transmission / reception unit 211 receives the hash data 168 for the block size “combined” of the remaining area from the backup management server 10 and calls the differential backup unit 214 as an input to perform the differential backup process for the remaining area. It performs (FIG. 14, step S405, FIG. 18, step S603-604). The differential backup data acquired here and transmitted from the client 21 to the backup management server 10 is received by the data transmission / reception unit 112, and the differential data storage unit 116 stores it as the differential data 131 (FIG. 18, step S605). .

  When this differential backup processing is completed, the data transmitting / receiving unit 211 transmits the optimum size “1024 kilobytes” to the backup management server 10 as the current size of the next processing (FIG. 14, step S405). In the backup management server 10, the optimization processing data creation unit 111 stores this in the working area 150 as the current size setting value 151 (FIG. 18, step S607).

(Overall operation of the first embodiment)
Next, the overall operation of the above embodiment will be described. In the differential backup method according to the present embodiment, the client 21 that is one or a plurality of computer devices that hold backup target data and the backup management server 10 that holds full backup data that is a complete backup of each backup target data are mutually connected. In the differential backup system 1 connected to the first backup data, the first hash data calculation unit of the backup management server divides the full backup data into blocks for each of the first to third block sizes given in advance. The first hash data calculation unit of the backup management server calculates the first hash data by applying a function given in advance to the data (steps S302 to S307 in FIG. 12), and the first to third block sizes. In each case, the first The optimization data generator of the backup management server divides the data into the test area that is a part of the full backup data and the other area (step S308 in FIG. 12). The optimization processing data creation unit of the backup management server transmits the hash data for the area to the client (step S601 in FIG. 18), and the second hash data of the client for each of the first to third block sizes. The calculation unit divides the backup target data into blocks of the block size, applies a function to the data of each block, and the second hash data calculation unit of the client calculates the second hash data (FIG. 17, step S503). , Second hash data calculated for each block and backup management The hash data comparison unit of the client compares the first hash data received from the server (step S504 in FIG. 17), and the backup is performed when the comparison result of the first and second hash data is different for each block. The differential backup unit of the client differentially backs up the block of the target data (step S508 in FIG. 17), and the block size with the shortest required time in the differential backup for the test area performed with the first to third block sizes Is determined as the optimum size by the client's optimum size calculation unit (FIG. 15, step S456, FIG. 16, step S459), and the first hash data for the remaining area in the case of the decided optimum size is used as the client differential backup optimum. Request from the backup management server (Figure 18. Step S603), the data transmission / reception unit of the backup management server transmits the first hash data for the requested remaining area to the client (FIG. 18, Step S604), and blocks the remaining area of the backup target data at the optimum size. The differential backup optimization unit of the client causes the differential backup unit to perform differential backup (steps S402 to 406 in FIG. 14).

  Here, the second block size is larger than the first block size, and the third block size is smaller than the first block size. Then, the optimization processing data creation unit of the backup management server stores the optimum size received when the first hash data for the remaining area is returned to the client as the first block size at the next same processing (FIG. 18). Step S607).

Here, each of the above-described operation steps may be programmed to be executable by a computer, and may be executed by the backup management server 10 and the client 21 which are computers that directly execute the respective steps.
With this configuration and operation, the present embodiment has the following effects.

  In this embodiment, the differential backup for the test area is performed with three types of block sizes, and as a result, the differential backup is performed for the remaining area with the block size having the shortest required time. An optimum block size can be obtained in conformity with actual data and processing contents, thereby reducing the time required for differential backup.

  In addition, since the optimum block size obtained by the processing is stored as the “current size” at the next processing, the numerical value becomes more and more optimized as the same differential backup processing is repeated many times. Of course, since this processing is performed in each of the clients 21, 22,..., Differential backup can be performed with an optimum block size for each client.

  The present invention has been described with reference to the specific embodiments shown in the drawings. However, the present invention is not limited to the embodiments shown in the drawings, and any known hitherto provided that the effects of the present invention are achieved. Even if it is a structure, it is employable.

  About each embodiment mentioned above, it is as follows when the summary of the novel technical content is put together. In addition, although part or all of the said embodiment is summarized as follows as a novel technique, this invention is not necessarily limited to this.

(Supplementary note 1) In a differential backup system in which a client, which is a computer device or a plurality of computer devices that hold backup target data, and a backup management server that holds full backup data that is a complete backup of each backup target data are interconnected There,
The backup management server is
First fixed storage means for storing the full backup data;
A first data transmitting / receiving unit that performs data communication with the client via a network;
A first hash data is calculated by dividing the full backup data into blocks at each of first to third block sizes given in advance and applying a given function to the data of each block. A hash data calculator,
The first hash data in each of the first to third block sizes is divided into a test area that is a part of the full backup data and a remaining area other than the test area. An optimization processing data creation unit that transmits hash data about the test area to the client;
The client
A second data transmitting / receiving unit that performs data communication with the backup management server;
Second fixed storage means for storing the backup target data;
A second hash data calculation unit that calculates the second hash data by dividing the backup target data into blocks and applying the function to the data of each block for each of the first to third block sizes When,
A hash data comparison unit that compares the calculated second hash data with the first hash data received from the backup management server for each block;
A differential backup unit that differentially backs up the block of the backup target data when the comparison result of the first and second hash data for each block is different;
For each of the first to third block sizes, a trial differential backup is performed on the differential backup unit for the test area of the backup target data using the first hash data for the test area. An optimal size calculation unit that determines the block size with the shortest required time in the experimental differential backup performed with the first to third block sizes as an optimal size,
Requesting the backup management server for the first hash data for the remaining area in the case of the determined optimum size, and using the first hash data for the remaining area sent thereto, the differential backup A differential backup system, comprising: a differential backup optimization unit that causes a differential backup of the remaining area of the backup target data to be performed at the optimal size.

(Supplementary Note 2) The second block size is larger than the first block size, and the third block size is smaller than the first block size,
The optimization processing data creation unit of the backup management server returns the first block size at the time of the same processing next time to the optimum size received when returning the first hash data for the remaining area to the client. The differential backup system according to appendix 1, characterized by having a function of storing as

(Supplementary Note 3) The time required for the experimental differential backup performed with the second block size by the optimum size calculation unit of the client is greater than the time required for the experimental differential backup performed with the first block size. 3. The differential backup system according to appendix 2, characterized in that if it is shorter, it has a function of determining the second block size as the optimum size at that time.

(Supplementary Note 4) The backup management server includes a processor included in the client as the second data transmission / reception unit, the second hash data calculation unit, the hash data comparison unit, the differential backup unit, the optimum size calculation unit, A boot image storage unit that stores a boot image that is a program that functions as the differential backup optimization unit in the first fixed storage unit;
The differential backup system according to appendix 1, further comprising: a boot image transmission unit that transmits the boot image to the client for operation.

(Supplementary Note 5) A backup management server that is connected to a client, which is one or a plurality of computer devices that hold backup target data, and that holds full backup data that is a complete backup of each backup target data,
Fixed storage means for storing the full backup data;
A data transmitting / receiving unit for performing data communication with the client via a network;
A hash data calculation unit for calculating hash data by dividing the full backup data into blocks at each of first to third block sizes given in advance and applying a function given in advance to the data of each block;
The test area is divided by dividing the hash data in each of the first to third block sizes into a test area that is a part of the full backup data and a remaining area other than the test area. A backup management server, comprising: an optimization processing data creation unit that transmits hash data for the client to the client.

(Additional remark 6) It is a client which is a computer apparatus which hold | maintains backup object data, and is mutually connected with the backup management server which hold | maintains the full backup data which carried out full backup of the said backup object data,
A data transmission / reception unit for performing data communication with the backup management server;
Fixed storage means for storing the backup target data;
A hash data calculation unit that calculates the second hash data by dividing the backup target data into blocks for each case of the first to third block sizes and applying the function to the data of each block;
A hash data comparison unit that compares the calculated second hash data with the first hash data received from the backup management server for each block;
A differential backup unit that differentially backs up the block of the backup target data when the comparison result of the first and second hash data for each block is different;
For each of the first to third block sizes, the difference for the test area of the backup target data using the first hash data for the test area that is part of the backup target data. An optimal size calculation unit for making the backup unit perform a test differential backup, and determining a block size having the shortest required time in the test differential backup performed in the first to third block sizes as an optimal size;
The first hash data for the remaining area other than the test area in the case of the determined optimum size is requested to the backup management server, and the sent first hash data for the remaining area is used. And a differential backup optimizing unit that causes the differential backup unit to perform differential backup of the remaining area of the backup target data with the optimal size.

(Supplementary note 7) A differential backup system in which a client, which is one or a plurality of computer devices that hold backup target data, and a backup management server that holds full backup data that is a complete backup of each backup target data are interconnected There,
The first hash data calculation unit of the backup management server divides the full backup data into blocks at each of the first to third block sizes given in advance,
The first hash data calculation unit of the backup management server calculates the first hash data by applying a function given in advance to the data of each block,
For each of the first to third block sizes, the first hash data is used for the test area in which the optimization processing data creation unit of the backup management server is part of the full backup data. Divide it into other remaining areas,
Hash data about the test area is transmitted to the client by the optimization processing data creation unit of the backup management server,
For each case of the first to third block sizes, the second hash data calculation unit of the client divides the backup target data into blocks of the block size,
The second hash data calculation unit of the client calculates second hash data by applying the function to the data of each block,
The hash data comparison unit of the client compares the second hash data calculated for each block and the first hash data received from the backup management server,
When the comparison result of the first and second hash data is different for each block, the differential backup unit of the client differentially backs up the block of the backup target data,
The optimal size calculation unit of the client determines the block size that required the least amount of time in the differential backup for the test area performed with the first to third block sizes as the optimal size,
The differential backup optimization unit of the client requests the backup management server for the first hash data for the remaining area in the case of the determined optimum size,
The data transmission / reception unit of the backup management server transmits the first hash data for the remaining area requested to the client,
A differential backup method, wherein the differential backup optimization unit of the client uses the first hash data for the remaining area to cause the differential backup unit to perform differential backup of the remaining area of the backup target data.

(Supplementary Note 8) The second block size is larger than the first block size, and the third block size is smaller than the first block size,
The optimization processing data creation unit of the backup management server sets the optimum size received when returning the first hash data for the remaining area to the client as the first block size at the same time next time. The differential backup method according to appendix 7, wherein the differential backup method is stored.

(Supplementary Note 9) A differential backup system in which a client, which is one or a plurality of computer devices that hold backup target data, and a backup management server that holds full backup data that is a complete backup of each backup target data are interconnected There,
In the computer provided in the backup management server,
A procedure for dividing the full backup data into blocks in each of first to third block sizes given in advance;
A procedure for calculating hash data by applying a function given in advance to the data of each block,
A procedure for dividing the hash data for each of the first to third block sizes into a test area that is a part of the full backup data and a remaining area other than the test area;
A procedure for transmitting hash data about the test area to the client;
And a differential backup program for causing the client to execute a procedure of transmitting the first hash data for the remaining area requested by the client to the client.

(Supplementary Note 10) In a differential backup system in which a client, which is one or a plurality of computer devices that hold backup target data, and a backup management server that holds full backup data that is a complete backup of each backup target data are interconnected There,
A computer included in the client;
A procedure for dividing the backup target data into blocks of the block size for each of the first to third block sizes;
A procedure for calculating the second hash data by applying the function to the data of each block,
Comparing the second hash data calculated for each block with the first hash data received from the backup management server;
A procedure for differentially backing up the block of the backup target data when the comparison result of the first and second hash data for each block is different;
A procedure for determining, as an optimum size, a block size that requires the least amount of time in a differential backup for a test area that is a part of the backup target data performed in the first to third block sizes;
Requesting the backup management server for first hash data for the remaining area other than the test area in the case of the determined optimum size;
And a differential backup program for executing a procedure for performing differential backup of the remaining area of the backup target data by using the first hash data of the remaining area sent thereto.

  The present invention can be applied to a backup system. Particularly, it is suitable for a backup system in which backups for a plurality of client computers are collectively performed by a backup management server.

DESCRIPTION OF SYMBOLS 1 Differential backup system 10 Backup management server 21 Client 30 Network 101, 201 Processor 102, 202 Temporary storage means 103, 203 Fixed storage means 104, 204 Communication means 111 Optimization processing data creation part 112 Data transmission / reception part 113 Boot image transmission part 114 Hash data calculation unit 115 Full backup data storage unit 116 Differential data storage unit 117 Boot image storage unit 120 Backup data storage area 121 Full backup data 121a, 221a Test area 121b, 221b Remaining area 130 Differential data storage area 131 Differential data 140 Boot image storage area 141 Boot image 150, 250 Working area 151 Current size setting value 160-1 9 hash data 211 data transceiver 212 differential backup optimization unit 213 optimum size calculation unit 214 differential backup unit 215 hash data calculating unit 216 hash data comparator 220 backup target data storage area 221 backup target data

Claims (10)

A differential backup system in which a client, which is one or a plurality of computer devices that hold backup target data, and a backup management server that holds full backup data in which each backup target data is completely backed up are mutually connected,
The backup management server is
First fixed storage means for storing the full backup data;
A first data transmitting / receiving unit that performs data communication with the client via a network;
A first hash data is calculated by dividing the full backup data into blocks at each of first to third block sizes given in advance and applying a given function to the data of each block. A hash data calculator,
The first hash data in each of the first to third block sizes is divided into a test area that is a part of the full backup data and a remaining area other than the test area. An optimization processing data creation unit that transmits hash data about the test area to the client;
The client
A second data transmitting / receiving unit that performs data communication with the backup management server;
Second fixed storage means for storing the backup target data;
A second hash data calculation unit that calculates the second hash data by dividing the backup target data into blocks and applying the function to the data of each block for each of the first to third block sizes When,
A hash data comparison unit that compares the calculated second hash data with the first hash data received from the backup management server for each block;
A differential backup unit that differentially backs up the block of the backup target data when the comparison result of the first and second hash data for each block is different;
For each of the first to third block sizes, a trial differential backup is performed on the differential backup unit for the test area of the backup target data using the first hash data for the test area. An optimal size calculation unit that determines the block size with the shortest required time in the experimental differential backup performed with the first to third block sizes as an optimal size,
Requesting the backup management server for the first hash data for the remaining area in the case of the determined optimum size, and using the first hash data for the remaining area sent thereto, the differential backup A differential backup system, comprising: a differential backup optimization unit that causes a differential backup of the remaining area of the backup target data to be performed at the optimal size. The second block size is larger than the first block size and the third block size is smaller than the first block size;
The optimization processing data creation unit of the backup management server returns the first block size at the time of the same processing next time to the optimum size received when returning the first hash data for the remaining area to the client. The differential backup system according to claim 1, wherein the differential backup system has a function of storing as:   If the time required for the experimental differential backup performed by the second block size is shorter than the time required for the experimental differential backup performed by the first block size by the optimal size calculation unit of the client, The differential backup system according to claim 2, further comprising a function of determining the second block size as the optimum size at a time point. The backup management server is
The processor included in the client functions as the second data transmission / reception unit, the second hash data calculation unit, the hash data comparison unit, the differential backup unit, the optimal size calculation unit, and the differential backup optimization unit. A boot image storage unit for storing a boot image as a program in the first fixed storage unit;
The differential backup system according to claim 1, further comprising: a boot image transmission unit configured to transmit the boot image to the client for operation. A backup management server that is connected to a client, which is one or more computer devices that hold backup target data, and that holds full backup data that is a complete backup of each backup target data,
Fixed storage means for storing the full backup data;
A data transmitting / receiving unit for performing data communication with the client via a network;
A hash data calculation unit for calculating hash data by dividing the full backup data into blocks at each of first to third block sizes given in advance and applying a function given in advance to the data of each block;
The test area is divided by dividing the hash data in each of the first to third block sizes into a test area that is a part of the full backup data and a remaining area other than the test area. A backup management server, comprising: an optimization processing data creation unit that transmits hash data for the client to the client. A client that is a computer device that holds backup target data and is interconnected with a backup management server that holds full backup data that is a complete backup of the backup target data,
A data transmission / reception unit for performing data communication with the backup management server;
Fixed storage means for storing the backup target data;
A hash data calculation unit that calculates the second hash data by dividing the backup target data into blocks for each case of the first to third block sizes and applying the function to the data of each block;
A hash data comparison unit that compares the calculated second hash data with the first hash data received from the backup management server for each block;
A differential backup unit that differentially backs up the block of the backup target data when the comparison result of the first and second hash data for each block is different;
For each of the first to third block sizes, the difference for the test area of the backup target data using the first hash data for the test area that is part of the backup target data. An optimal size calculation unit for making the backup unit perform a test differential backup, and determining a block size having the shortest required time in the test differential backup performed in the first to third block sizes as an optimal size;
The first hash data for the remaining area other than the test area in the case of the determined optimum size is requested to the backup management server, and the sent first hash data for the remaining area is used. And a differential backup optimizing unit that causes the differential backup unit to perform differential backup of the remaining area of the backup target data with the optimal size. In a differential backup system in which a client, which is one or a plurality of computer devices that hold backup target data, and a backup management server that holds full backup data in which each backup target data is completely backed up are mutually connected,
The first hash data calculation unit of the backup management server divides the full backup data into blocks at each of the first to third block sizes given in advance,
The first hash data calculation unit of the backup management server calculates the first hash data by applying a function given in advance to the data of each block,
For each of the first to third block sizes, the first hash data is used for the test area in which the optimization processing data creation unit of the backup management server is part of the full backup data. Divide it into other remaining areas,
Hash data about the test area is transmitted to the client by the optimization processing data creation unit of the backup management server,
For each case of the first to third block sizes, the second hash data calculation unit of the client divides the backup target data into blocks of the block size,
The second hash data calculation unit of the client calculates second hash data by applying the function to the data of each block,
The hash data comparison unit of the client compares the second hash data calculated for each block and the first hash data received from the backup management server,
When the comparison result of the first and second hash data is different for each block, the differential backup unit of the client differentially backs up the block of the backup target data,
The optimal size calculation unit of the client determines the block size that required the least amount of time in the differential backup for the test area performed with the first to third block sizes as the optimal size,
The differential backup optimization unit of the client requests the backup management server for the first hash data for the remaining area in the case of the determined optimum size,
The data transmission / reception unit of the backup management server transmits the first hash data for the remaining area requested to the client,
A differential backup method, wherein the differential backup optimization unit of the client uses the first hash data for the remaining area to cause the differential backup unit to perform differential backup of the remaining area of the backup target data. The second block size is larger than the first block size and the third block size is smaller than the first block size;
The optimization processing data creation unit of the backup management server sets the optimum size received when returning the first hash data for the remaining area to the client as the first block size at the same time next time. The differential backup method according to claim 7, wherein the differential backup method is stored. In a differential backup system in which a client, which is one or a plurality of computer devices that hold backup target data, and a backup management server that holds full backup data in which each backup target data is completely backed up are mutually connected,
In the computer provided in the backup management server,
A procedure for dividing the full backup data into blocks in each of first to third block sizes given in advance;
A procedure for calculating hash data by applying a function given in advance to the data of each block,
A procedure for dividing the hash data for each of the first to third block sizes into a test area that is a part of the full backup data and a remaining area other than the test area;
A procedure for transmitting hash data about the test area to the client;
And a differential backup program for causing the client to execute a procedure of transmitting the first hash data for the remaining area requested by the client to the client. In a differential backup system in which a client, which is one or a plurality of computer devices that hold backup target data, and a backup management server that holds full backup data in which each backup target data is completely backed up are mutually connected,
A computer included in the client;
A procedure for dividing the backup target data into blocks of the block size for each of the first to third block sizes;
A procedure for calculating the second hash data by applying the function to the data of each block,
Comparing the second hash data calculated for each block with the first hash data received from the backup management server;
A procedure for differentially backing up the block of the backup target data when the comparison result of the first and second hash data for each block is different;
A procedure for determining, as an optimum size, a block size that requires the least amount of time in a differential backup for a test area that is a part of the backup target data performed in the first to third block sizes;
Requesting the backup management server for first hash data for the remaining area other than the test area in the case of the determined optimum size;
And a differential backup program for executing a procedure for performing differential backup of the remaining area of the backup target data by using the first hash data of the remaining area sent thereto.

Images