JP2007164523A - Information processor, distributed synchronous information system, information synchronization method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a distributed synchronous information system capable of efficiently synchronizing data stored by respective apparatuses without requiring a lot of storage area. <P>SOLUTION: A relational database 102 stores an updater updating a line and an event number for deciding whether it is before or after the update in each the updater correspondingly to data of each the line. An update information transmission requirement means 106 of a machine A 100 retrieves a maximum event number in each the updater in reference to the relational database 102, and transmits update information transmission requirement including it to a machine B 110. An update information transmission means 107 of the machine B 110 extracts the update on and after the maximum event number in each the updater included in the update information transmission requirement from the relational database 102, transmits it to the machine A 100, and makes the relational database 102 of the machine A be updated. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an information processing device, a distributed synchronization information system, an information synchronization method, and a program, and more particularly, an information processing device that synchronizes database data among a plurality of information devices having a database, and distributed synchronization information. The present invention relates to a system, an information synchronization method, and a program.

  Conventionally used information systems include LDAP and NIS by Sun Microsystems. These information systems are configured as a server / client system, and when a client terminal is carried while disconnected from the network, the database cannot be accessed from the terminal side. As a method of using the terminal while disconnected from the network, a method of copying the database on the server to the terminal and carrying it can be considered. However, in this case, the database can be referred to even when disconnected from the network. However, when the database is updated at a plurality of terminals, the data in the database is synchronized between the plurality of terminals. I can't.

  When data is updated after the terminal is disconnected from the network, all data in the database is transmitted and received between the terminals when the network is reconnected, and the databases of each terminal are synchronized by comparing both data. be able to. However, in that case, there is a problem that the amount of transmitted / received data becomes enormous. As a conventional technique for solving such a problem, there is a technique described in Patent Document 1, for example. In Patent Document 1, a database on a server is a master and a database on a client is a non-master. The client stores the update after being disconnected from the network in the table as difference information. When the network is reconnected, difference information is transmitted to and received from the server, and the master and the non-master are synchronized.

  As a technique for synchronizing information among a plurality of information devices (information processing apparatuses), there is a technique described in Non-Patent Document 1. In Non-Patent Document 1, each information device is treated equally, and the update information held by each is synchronized when connected to the network. FIG. 21 shows a state of information synchronization in the technique described in Non-Patent Document 1. Each information device holds an update information log 2102 for the data in the database 2101. The log 2102 includes information indicating which part of the database 2101 has been updated to what value, and the ID of the device that performed the update, and the time when the update was performed.

  In Non-Patent Document 1, information synchronization is performed between two information devices, and information synchronization is performed between a plurality of information devices by repeating this. In information synchronization between two information devices, one information device (reception side) scans the log 2102 of its own device, finds the latest update time for each device ID, and uses this as the other information device. Send to (sender). The transmission side device scans the log 2102, and for each device ID, passes the update performed after the update time notified from the reception side device to the reception side device. As a result, information held by the transmitting device and not held by the receiving device can be transmitted to the receiving device. By exchanging the positions of the receiving device and the transmitting device and performing the same operation, the information of both can be synchronized.

JP 2001-142766 A (FIG. 6, FIG. 7, paragraphs 0038 to 0050, etc.) Karin Petersen, Mike J. Spreitzer, Douglas B. Terry, Flexible Update Propagation for Weakly Consistent Replication, Proceedings of the 16th ACM Symposium on Operating Systems Principles (SOSP-16), 1997, pp. 288-301

  In Patent Document 1, there is only one database serving as a master, and all synchronization needs to be performed on a one-to-one basis with the master. Considering the case where non-master clients synchronize updates, the information held by the clients is only the difference from the master, so they cannot exchange unknown information. They cannot synchronize with each other. For example, when considering synchronization of a portable information device such as a household device group or a PDA, Patent Document 1 cannot perform synchronization between portable information devices. In addition, there is a problem that the information device as a master must be always connected to the network and powered on at the time of synchronization.

  In the technique described in Non-Patent Document 1, all update information needs to be stored in the log 2102. Since the log information becomes enormous, there is a problem that a large amount of storage area is required for mounting. It may be possible to determine which part of the log 2102 is unnecessary by some method, but in that case, the algorithm becomes complicated. In Non-Patent Document 1, since update information is managed in a log format, log scanning takes time, and the efficiency at the time of information synchronization is poor. Furthermore, in the prior art including Patent Document 1, information synchronization is limited only to synchronization between two information devices, and when synchronization is performed between a plurality of information devices, There is a problem that it is necessary to repeatedly perform information synchronization, and communication is required many times during information synchronization.

  The present invention solves the above-mentioned problems of the prior art, an information processing apparatus capable of efficiently synchronizing data stored in each device without requiring a large amount of storage area, a distributed synchronization information system, an information synchronization method, and The purpose is to provide a program.

  In order to achieve the above object, a distributed synchronous information system according to the present invention includes a plurality of information processing apparatuses each having a database, and the distributed synchronous information system synchronizes the database among the information processing apparatuses. Each of the information processing devices includes each data of the database, an updater indicating the information processing device that performed the data update, and synchronization information including information for determining before and after the update in the information processing device The update information transmission request means for searching for a storage device that stores the information in association with each other and transmitting an update information transmission request including information indicating the latest update for each updater to another information processing device, and the update information transmission request Update data updated by a newer update than the latest update for each updater included in the update information transmission request from another information processing apparatus that has received It receives the update information including the synchronization information corresponding to the update data, characterized by comprising a first database synchronization means for updating the database and synchronization information in its own device.

The database synchronization method of the present invention includes a plurality of information processing apparatuses each having a database, and the information processing apparatus includes: a database synchronization method in a distributed synchronization information system that synchronizes the database among the information processing apparatuses; When updating the database, update data, an updater indicating the information processing apparatus that performed the data update, and synchronization information including information for determining before and after the update in the information processing apparatus are stored in association with each other. Steps,
The information processing device searches the synchronization information, and transmits an update information transmission request including information indicating the latest update for each updater to another information processing device;
Update data updated by a newer update than the latest update for each updater included in the update information transmission request from another information processing apparatus that has received the update information transmission request, and the update data Receiving update information including synchronization information corresponding to
The information processing apparatus includes a step of updating the database and the synchronization information in the own apparatus based on the received update information.

  An information processing apparatus according to the present invention includes a database and updates data with each data of the database in an information processing apparatus that synchronizes the database with another information processing apparatus connected via a network. Search for a storage device that associates and stores update information indicating the information processing device and synchronization information including information for determining before and after update in the information processing device, and updates the latest information for each updater. Update information transmission request means for transmitting an update information transmission request including information indicating update to another information processing apparatus, and an update included in the update information transmission request from another information processing apparatus that has received the update information transmission request Update information including update data updated by a newer update than the latest update for each user and synchronization information corresponding to the update data is received, and the database in the own device is received. And characterized by comprising a first database synchronization means for updating the synchronization information.

  The program according to the present invention is a program for synchronizing a database of a computer with a database of another computer connected via a network, wherein the computer performs update data and data update when updating the database. And a step of storing in association with synchronization information including information for determining before and after the update in the computer, searching for the synchronization information, and updating the latest update for each of the updaters An update information transmission request including information to be transmitted to another computer, and an update that is newer than the latest update for each updater included in the update information transmission request from another computer that has received the update information transmission request. Update information including update data updated by the method and synchronization information corresponding to the update data Receiving, based on the received update information, characterized in that and a step of updating the database and synchronization information in its own device.

  In the distributed synchronous information system, information processing apparatus, database synchronization method, and program according to the present invention, an updater who has updated the data corresponding to each data in the database, and an update for each information processing apparatus (computer) Synchronization information including information for determining before and after is stored. By referring to this synchronization information, it is possible to know what the latest update is for each updater in the database in the own apparatus. When synchronizing with other information processing devices, refer to the synchronization information to obtain information indicating the latest update for each updater, send this to other information processing devices, from other information processing devices, Data updated by a newer update than the latest update for each updater in the database of the own device is received, and the database is updated. By doing so, it is possible to obtain updates that are not made in the database of the own device without comparing all the data in the database, compared with the case where all the data in the database is transmitted and received and compared. Thus, the amount of data transmitted and received can be reduced. In addition, when an update log is stored, the number of update log lines increases each time an update is performed, and the size of the update log increases as the number of updates increases. Therefore, the size of the synchronization information does not increase even if the number of updates increases, and a smaller recording area is required than when an update log is stored.

  The distributed synchronous information system according to the present invention creates information for determining before and after update in the own device when the database is updated as an updater in the own device, and synchronizes information corresponding to the update data. It is possible to employ a configuration that further includes synchronization information updating means for updating. In the database synchronization method of the present invention, in the step of storing the synchronization information, when the information processing apparatus updates the database as an updater in the own apparatus, information for determining before and after the update in the own apparatus Can be created, and the synchronization information can be updated in correspondence with the update data.

  In the distributed synchronization type information system and database synchronization method of the present invention, information for determining before and after updating in the information processing apparatus is incremented each time the database is updated in the information processing apparatus. A configuration that is an event number for each device can be adopted. In the program of the present invention, it is possible to adopt a configuration in which the information for determining before and after updating in the computer is an event number for each computer that is incremented each time the database is updated in the computer. In this case, by comparing the magnitudes of the event numbers, it is possible to determine before and after updating for each updater who has updated the database. The update time can also be used as information for determining before and after the update. In that case, before and after the update time, the update before and after can be determined for each updater.

  In the distributed synchronous information system according to the present invention, the synchronization information includes an update time, and the first database synchronization means is configured to update the synchronization information stored in the own device, with the update time of the synchronization information included in the update information. If it is newer than the time, a configuration for updating the database and the synchronization information can be adopted. In the database synchronization method of the present invention, the synchronization information includes an update time, and in the step of updating the database and the synchronization information, the information processing apparatus determines that the update time of the synchronization information included in the update information is A configuration in which the database and the synchronization information are updated when the update time of the stored synchronization information is newer can be employed. In the program of the present invention, the synchronization information includes an update time, and in the step of updating the database and the synchronization information, the update time of the synchronization information included in the update information is greater than the update time of the synchronization information stored in the own device. If it is new, a configuration for updating the database and the synchronization information can be adopted. In this case, new data can be prevented from being updated by old data. The update time can be included in the synchronization information in a form added to the information for determining before and after the update for each updater, or the update time is used as information for determining before and after the update for each updater. In some cases, this may be shared.

  In the distributed synchronization type information system and the database synchronization method of the present invention, the synchronization information includes a pair of an information processing apparatus that has been updated so far and information for determining before and after the update in the information processing apparatus. It is possible to employ a configuration that further includes an update history. In the program of the present invention, a configuration can be adopted in which the synchronization information further includes an update history including a set of a computer that has been updated and information for determining before and after the update in the computer. In this case, by referring to the update history, it is possible to know what route is used for the update.

  In the distributed synchronization type information system of the present invention, the first database synchronization unit compares the update history included in the received update information with the update history stored in the own device, and updates included in the received update information. When it is determined that the history includes an update history stored in the device itself, a configuration for updating the database and the corresponding synchronization information can be employed. In the database synchronization method of the present invention, the information processing apparatus further includes a step of comparing an update history included in the received update information with an update history stored in the own apparatus. When it is determined that the update history included in the update information includes the update history stored in the device itself, a configuration for updating the database and the corresponding synchronization information can be employed. In the program of the present invention, the computer further executes a step of comparing the update history included in the received update information with the update history stored in the own device, and in the step of updating the database and the synchronization information, the update is performed. In the step of comparing the histories, when it is determined that the update history included in the received update information includes the update history stored in the device itself, a configuration of updating the database and the corresponding synchronization information can be employed. When the update history included in the synchronization information of the received update information includes the update history of the synchronization information stored in the own device, it is updated on the same route until the update in the middle, and then information other than the own device This indicates that a new update has been made in the processing device. In such a case, there is no problem in updating the database and synchronization information with the received update information, and the database is synchronized with other information processing apparatuses by updating the database with the received update information. can do.

  The distributed synchronous information system of the present invention refers to update propagation information that stores information indicating the latest update for each updater for updates reflected in the database of each information processing apparatus. It is possible to adopt a configuration further comprising history deletion means for deleting the history before update reflected in the database of the information processing apparatus. In the database synchronization method of the present invention, the information processing apparatus refers to the update propagation information that stores information indicating the latest update for each updater with respect to the update reflected in the database of each information processing apparatus. A configuration that further includes a step of deleting the history before update reflected in the database of all information processing apparatuses from the history can be adopted. The program according to the present invention refers to update propagation information that stores information indicating the latest update for each updater, and updates all the computers from the update history. It is possible to adopt a configuration in which the step of deleting the history before the update reflected in the database is further executed. Each information processing apparatus manages, for each updater, the update up to which point in time in the database in each information processing apparatus has been updated for each information processing apparatus based on the update propagation information. By examining this update propagation information, for each updater, it is possible to specify at least what point in time the update has been performed in the databases in all the information processing apparatuses. For example, when an event number is used as information for determining before and after the update, for an updater, information indicating the oldest update among information indicating the latest update of the updater in each information processing apparatus ( If the event number (minimum value) is the event number “8”, it can be specified that the update up to the event number “8” is performed in the databases in all the information processing apparatuses. In this case, it can be determined that the update history before the event number “8” by the updater is unnecessary because it has already been updated in the databases in all the information processing apparatuses. By deleting such an unnecessary update history, it is possible to avoid an increase in the update history as the number of updates increases.

  The distributed synchronous information system of the present invention updates the database in its own device, and updates update means for transmitting update data and synchronization information corresponding to the update data to other information processing devices connected to the network And a second database synchronization unit that receives the update data and the synchronization information transmitted by the update propagation unit of the other information processing apparatus and updates the database and the synchronization information. In the database synchronization method of the present invention, when the information processing apparatus updates the database in its own apparatus, it transmits update data and synchronization information corresponding to the update data to another information processing apparatus connected to the network. And receiving the update data and the synchronization information transmitted in the step of transmitting the update data and the synchronization information corresponding to the update data, and updating the database and the synchronization information. Furthermore, the structure which has is employable. The program according to the present invention includes a step of transmitting update data and synchronization information corresponding to the update data to another computer connected to the network when the database is updated in the computer. A configuration is adopted in which the update data transmitted in the step of transmitting the data and the synchronization information corresponding to the update data and the synchronization information corresponding to the update data are received and the database and the synchronization information are updated. it can. In this case, the database update performed by a certain information processing apparatus is reflected on other information processing apparatuses connected to the network, so that the databases can be synchronized between the information processing apparatuses connected to the network.

  In the distributed synchronous information system of the present invention, when an update information transmission request is received from another information processing apparatus, the update information for each updater included in the received update information transmission request is referred to by referring to the synchronization information stored in the own apparatus. Update information including the update data extracted by the search and the synchronization information corresponding to the update data is sent to the information processing apparatus that has transmitted the update information transmission request. It is possible to employ a configuration further comprising update information transmitting means for transmitting the information. In the database synchronization method of the present invention, when the information processing apparatus receives an update information transmission request from another information processing apparatus, the update information included in the received update information transmission request is referred to by referring to the synchronization information stored in the own apparatus. Searching for an update that is newer than the latest update for each person, and the information processing device includes update information including update data extracted by the search and synchronization information corresponding to the update data, It is possible to employ a configuration further comprising a step of transmitting the update information transmission request to the information processing apparatus that has transmitted the update information transmission request. When the program of the present invention receives an update information transmission request from another computer to the computer, the program updates the latest update for each updater included in the received update information transmission request with reference to the synchronization information stored in the device. The update information including the step of searching for a newer update, the update data of the update extracted by the search, and the synchronization information corresponding to the update data is transmitted to the computer that has transmitted the update information transmission request. It is possible to adopt a configuration that further executes the step of performing. The information processing device on the receiving side of the update information transmission request refers to the synchronization information in its own device, so that the update that is newer than the latest update for each updater included in the received update information transmission request, that is, update It is possible to extract updates that are not performed by the machine that has transmitted the information transmission request but are performed by the machine that has received the update information transmission request. By transmitting the update information including the update data extracted in this way to the information processing apparatus on the side that transmitted the update information transmission request, the database can be synchronized with a small amount of data. Further, since update information is extracted by searching for synchronous information having a smaller amount of information than log information, the search can be performed at a higher speed than when log information is used.

  In the distributed synchronous information system of the present invention, when an information processing apparatus reconnected to the network receives an update information transmission request from one of the other information processing apparatuses connected to the network, the reconnection is performed. The update information transmitting means of the information processing apparatus can employ a configuration in which the update information is multicast transmitted to another information processing apparatus connected to the network. When the information processing apparatus reconnected to the network receives an update information transmission request from one of the other information processing apparatuses connected to the network, the database synchronization method of the present invention transmits the update information to the network It is possible to adopt a configuration further including a step of multicast transmission to another information processing apparatus connected to the. When the program of the present invention receives an update information transmission request from one of the other computers connected to the network after the computer is reconnected to the network, the program of the present invention connects the update information to the network. It is possible to adopt a configuration that further executes a step of multicast transmission to another computer. In this case, when the databases of the other information processing apparatuses connected to the network are synchronized, update information is transmitted between the information processing apparatus connected to the network and each of the other information processing apparatuses connected to the network. The database of each information processing apparatus can be synchronized without transmitting / receiving request and update information. Therefore, the database can be synchronized more efficiently than when the one-to-one communication is repeated to synchronize the database.

  In the distributed synchronous information system, information processing apparatus, database synchronization method, and program according to the present invention, an updater who has updated the data corresponding to each data in the database, and an update for each information processing apparatus (computer) Synchronization information including information for determining before and after is stored. With reference to this synchronization information, information indicating the latest update of each updater in the database in the own apparatus is checked, an update information transmission request including this is transmitted to another information processing apparatus, and other information apparatus The data updated by the new update from the latest update of each updater in the database in the own device and the synchronization information corresponding to the data are received, and the database and the synchronization information are updated. By doing so, it is possible to obtain updates that are not made in the database of the own device without comparing all the data in the database, compared with the case where all the data in the database is transmitted and received and compared. Thus, the amount of data transmitted and received can be reduced. In addition, since synchronization information is stored corresponding to each data in the database, the size of the synchronization information does not increase even if the number of updates increases, and the required recording area is less than when an update log is saved. That's it.

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

First Embodiment FIG. 1 shows the configuration of a distributed synchronous information system according to a first embodiment of the present invention. The distributed synchronization type information system 10 is composed of an arbitrary number of information devices that synchronize information. In the example of the figure, the distributed synchronization information system 10 is composed of three information devices: a machine A100, a machine B110, and a machine C120. The The machine A 100 includes an application 101, a relational database 102, an in-database synchronization information updating unit 103, a reconnection synchronization unit (first database synchronization unit) 104, and a connected synchronization unit (second database synchronization unit) 105. Prepare. Although not shown in the drawing, the machine B110 and the machine C120 each have the same configuration as the machine A100. Each machine is configured as a portable terminal device that performs processing using data in the relational database 102, for example.

  The machine A100, the machine B110, and the machine C120 are each disconnected from a network (not shown) or reconnected to the network at an arbitrary timing. The machine A100, the machine B110, and the machine C120 update the relational database 102 with the application 101. Updating the relational database 102 includes changing an existing line (data), inserting a new line, and deleting an existing line. When each machine updates the relational database 102, the update is propagated to other machines, and the relational database 102 of each machine is synchronized.

  The relational database 102 is updated at an arbitrary timing, both in a state where the machine is connected to the network and in a state where the machine is disconnected from the network. The connected synchronization unit 105 includes an update propagation unit 108, which propagates updates performed in the relational database 102 to other machines connected to the network while the machine is connected to the network. The stored contents of the database 102 are synchronized. The reconnection synchronization unit 104 includes an update information transmission request unit 106 and an update information transmission unit 107. When the machine is reconnected to the network after being disconnected from the network, the reconnected machine and the network connection Updates made on other machines are propagated to synchronize the stored contents of the relational database 102.

  FIG. 2 shows a specific example of the data structure of the relational database 102. The relational database 102 stores, in addition to normal data (“primary key”, “data”), synchronization information added corresponding to each data. In the figure, a portion 201 surrounded by a dotted line corresponds to the synchronization information. The added synchronization information includes an updater who updated the data (row) and information for determining the context of the update. In addition, a deletion flag and an update time are included. When the application 101 updates the relational database 102, the in-database synchronization information update unit 103 (FIG. 1) updates the synchronization information corresponding to the updated data (row).

  As the “information for determining before and after update” of the synchronization information, for example, the time at which the update was performed or a number (event number) assigned in order at each machine every time update processing occurs is used. Can do. As information for determining before and after the update, it is desirable to use an event number. This is because if the update time is used as information for determining before and after the update, the context may be out of order by changing the system time. In the example of FIG. 2, event numbers are used as information for determining before and after the update.

  The updater and the event number will be described. The updater indicates the machine that has updated the data. The event number indicates the update order in each machine. For example, when the relational database 102 is updated for the first time on the machine A100, an event number “1” is assigned to the update. Thereafter, when the relational database 102 is updated in the machine A100, an event number “2” is assigned. This event number is managed for each machine, and is sequentially assigned in each machine regardless of updates in other machines. For example, after the update corresponding to the event number “2” is performed in the machine A100, the relational database 102 is updated in the machine B110, and then the relational database 102 is further updated in the machine A100. Is assigned an event number “3”.

  The deletion flag indicates whether or not the line has been deleted. When deleting data, the application 101 changes the deletion flag to “Y” without actually deleting the line. A row in which the deletion flag is set to “Y” is handled as being deleted from the relational database 102. As described above, the rows that are deleted by using the deletion flag without actually deleting the rows are handled in the synchronization by the synchronization unit 104 at the time of reconnection. If it exists and the row does not exist in the relational database 102 of another machine, it is distinguished whether the row was deleted while disconnected from the network or added on another machine This is necessary.

  Here, consider a case where a machine that is disconnected and a machine that is connected to the network update the same row of each relational database 102 while a machine is disconnected from the network. In this case, when the disconnected machine reconnects to the network and tries to synchronize the relational database 102, the same row is updated separately, which causes inconsistency for that row. As a method for resolving the contradiction, there are a method for allowing the user to select which one is to be given priority, a method for giving priority to a specific machine, a method for giving priority to a new update time, and the like. In the present embodiment, when the synchronization information includes the update time and is updated by two machines with respect to the same row, priority is given to the new update time.

  The synchronization information can be included in the same row as the data in the relational database 102 as shown in FIG. 2, and can also be managed in a separate table separately from the data portion as shown in FIG. it can. In FIG. 3, the relational database 102 is separated into a table 301 including data used by the application 101 and a table 302 including a synchronization information part (201 in FIG. 2). When separating in this way, the rows of the table 301 and the rows of the table 302 may be associated with each row using a unique key, for example, a primary key. When data and synchronization information are separated, there is no need to change the data table configuration that the relational database originally had. Therefore, by adding the synchronization information table to the existing relational database, it is distributed. A type information synchronization system can be constructed.

  FIG. 4 shows an operation procedure of the in-database synchronization information updating unit 103. The application 101 updates the relational database 102 using SQL or the like (step S401). In the update of the relational database 102, when a new row is inserted (added), the machine is disconnected from the network so that it does not overlap with the primary key of the newly inserted row. For example, the machine name may be included in the primary key, or a range of values that can be used as a newly inserted primary key may be set for each machine.

  When the relational database 102 is updated, the in-database synchronization information updating unit 103 determines whether the update content is a row insertion, a row rewrite, or a row deletion (step S402). When the content of the update is a line rewrite, the updater, event number, and update time of the synchronization information (201 in FIG. 2, 302 in FIG. 3) are updated (step S406). For example, when the application 101 rewrites the row of the relational database 102 in the machine A100, the in-database synchronization information update unit 103 in the machine A sets the updater of the synchronization information as “machine A”. For the event number, the maximum event number already assigned in the machine A is stored, and a value obtained by adding 1 to the value is stored. The current time is set as the update time. Here, it is assumed that the time information held by each machine is synchronized by some means.

  If the update content is a row insertion and the synchronization information is separated from the data as shown in FIG. 3, a new row is inserted into the synchronization information table 302 (step S403). At this time, for example, the same key as the primary key of the row of the inserted data table 301 is set as the primary key of the synchronization information, and the correspondence with the inserted data row is taken. Thereafter, in step S406, the updater, the event number, and the update time of the synchronization information (201 in FIG. 2 and 302 in FIG. 3) are updated. When the update content is deletion of a row, if the synchronization information is separated from the data as shown in FIG. 3, the corresponding row is deleted from the data table 301 (step S404). For the synchronization information (201 in FIG. 2 and 302 in FIG. 3), the deletion information is set to “Y” without deleting the synchronization information itself (step S405). Then, it progresses to step S406 and updates synchronous information.

  FIG. 5 shows an operation procedure of the connecting synchronization means 105. When the in-database synchronization information updating unit 103 updates the synchronization information, the update propagation unit 108 of the connected synchronization unit 105 searches for the updated data using the “primary key” or the like, and is a relational database to be propagated. Data on 102 and synchronization information are acquired (step S501). Thereafter, the update propagation unit 108 encodes the data acquired in step S501 into a form suitable for propagation (step S502). For this encoding, for example, a format such as XML can be used.

  The update propagation unit 108 transmits the data encoded in step S502 to the machine currently connected to the network (step S503). In step S503, the machine whose relational database 102 has been updated performs one-to-one communication with another machine connected to the network, and transmits the encoded data. Alternatively, the encoded data is transmitted to all machines connected to the network using multicast. When receiving the encoded data from the other machine, the connecting synchronization means 105 of the machine connected to the network decodes the received data and updates the relational database 102 in its own machine according to the content (step S504). . By such an operation, the relational database 102 of each machine connected to the network is synchronized.

  FIG. 6 shows an operation procedure of the reconnection synchronization means 104. When there is a machine newly connected to the network, the reconnection synchronization means 104 finds a machine that is newly communicable in the network (step S601). In step S601, the reconnection synchronization unit 104 of a machine newly connected (reconnected) to the network discovers each machine that has already been connected to the network as a newly communicable machine. On the other hand, the reconnection synchronization means 104 for a machine that has already been connected to the network finds a machine that is newly connected to the network as a machine that is newly communicable. Discovery of a newly communicable machine can be realized by using a function of UPnP, for example.

  The update information transmission request unit 106 of the reconnection synchronization unit 104 sets the update information to each of the machines discovered in step S601 as “update information reception side machine” and the discovered machine as “update information transmission side machine”. Request to send The update information transmitting unit 107 of the update information transmitting machine that has received this request transmits update information including an update of the relational database 102 that is not performed by the update information receiving machine to the update information receiving machine. The reconnection synchronization unit 104 of the update information transmission side machine updates the relational database 102 based on the received update information (step S602). When there are a plurality of machines found in step S601, the reconnection synchronization means 104 executes step S602 for each discovered machine.

  FIG. 7 shows how the relational database 102 is updated when the network is reconnected. Here, consider a case where the machine A100 and the machine C120 are already connected to the network and the machine B110 is reconnected to the network. In this case, the machine B110 reconnected to the network discovers the machine A100 and the machine C120 that are already connected to the network as newly communicable machines in step S601 in FIG. On the other hand, each of the machine A100 and the machine C120 finds the reconnected machine B110 as a machine that can newly communicate.

  In step S602, the machine B110 sets itself as the “update information receiving machine” and the discovered machines A100 and C120 as the “update information transmitting machine”, and transmits update information to the machine A100 and the machine C120, respectively. Request. Upon receipt of this request, the machine A 100 and the machine C 120 respectively transmit update information that is stored in the relational database 102 (FIG. 1) of the own device and not in the relational database 102 of the machine B110 to the machine B110. The machine B110 updates the relational database 102 based on the information received from the machine A100 and the machine C120. By this update, the update performed in the machine A100 or the machine C120 while the machine B110 is disconnected from the network is reflected in the relational database 102 of the machine B110.

  On the other hand, in step S602, each of the machine A100 and the machine C120 sets itself as an “update information receiving machine” and sets the discovered machine B110 as an “update information transmitting machine” to transmit update information to the machine B110. Request. The machine B 110 transmits update information including updates that are performed in the relational database 102 of the own device and not performed in the machines A 100 and C 120 to the machines A 100 and C 120, respectively. Each of the machine A100 and the machine C120 updates the relational database 102 based on the information received from the machine B110. By this update, the update performed in the machine B110 while the machine B110 is disconnected from the network is reflected in the relational database 102 of the machine A100 and the machine C120.

  8 and 9 show the detailed procedure of the update process at the time of network reconnection in step S602 of FIG. 6, respectively. FIG. 8 shows the operation procedure of the update information receiving machine, and FIG. The operation procedure of the update information transmission side machine is shown. The update information transmission request unit 106 of the reconnection synchronization unit 104 of the update information receiving machine refers to the synchronization information (FIG. 2 or FIG. 3) and searches for a row with the largest event number for each updater. (Step S801), the maximum event number of each updater is acquired. The event number indicates the update order for each updater, and the search in step S801 is equivalent to searching for “what is the closest update for each updater”.

  The search in step S801 can be performed at high speed in a time independent of the table size by using the index function of the relational database 102 or the like. For each updater, a table holding the maximum event number is prepared, and when the database is updated, the table is updated by the synchronization information updating unit 103 in the database, thereby omitting the search for the synchronization information. It is also possible to do. When the update information transmission request unit 106 searches for a row having the largest event number for each updater, the update information transmission request unit 106 transmits an update information transmission request including information on the maximum event number for each machine to the update information transmission side machine (step S802). ).

  FIG. 10 shows a specific example of information exchanged between two machines in the update process at the time of network reconnection. When the machine B110 is an update information transmission side machine, an event number “11” is obtained by searching the maximum event number for the updater “machine A” from the table 1011 stored in the relational database 102 of the machine B110. Similarly, when the maximum event numbers are searched for the updaters “Machine B” and “Machine C”, the event number “8” for the machine B and the event number “5” for the machine C are obtained. The update information transmission request means 106 of the machine B 110 updates the update information transmission request 1020 including “Machine A, maximum event number 11”, “Machine B, maximum event number 8”, “Machine C, maximum event number 5”. The information is transmitted to the machine A100 that is the information transmission side machine, and the maximum event number for each updater is notified to the machine A.

  The update information transmission side machine receives an update information transmission request from the update information reception side machine and receives notification of the maximum event number for each updater (step S901 in FIG. 9). The update information transmitting unit 107 searches the table 1001 of the relational database 102 of the own device, and extracts a row having an event number larger than the maximum event number notified from the update information receiving machine for each updater. (Step S902). Thereafter, the extracted line is encoded, and this is transmitted as update information (difference information) to the update information receiving machine (step S903). If a plurality of lines are extracted in step S902, the extracted lines are transmitted to the update information receiving machine.

  In the example of FIG. 10, when the table 1001 of the relational database 102 of the machine A100 is searched and an event number row exceeding the maximum event number of each updater notified from the machine B110 is extracted, the row of the primary key “3” ( Updater “machine C”, event number “6”) is extracted. This row is a row that is first updated in the machine C120, then propagated to the machine A100, and updated in the relational database 102 of the machine A100. As described above, in step S902, a row updated by a machine different from the update information transmitting machine (machine A) may be extracted. The reconnection synchronization means 104 of the machine A100 transmits the extracted row of the primary key “3” to the machine B as the update information 1021.

  Returning to FIG. 8, the reconnection synchronization unit 104 of the update information receiving machine acquires the update information transmitted from the update information transmitting machine (step S803). The reconnection synchronization means 104 decodes the received update information, and the row corresponding to the row extracted as the row of the event number larger than the maximum event number in the update information receiving side machine in the update information sending side machine is It is determined whether or not it exists in its own relational database 102 (step S804). Whether there is a corresponding row is determined using, for example, a primary key. If it does not exist, the acquired row is inserted into the relational database 102 including the synchronization information (step S808).

  When the corresponding row exists, the “update time” of the acquired row is compared with the “update time” of the corresponding row in the relational database 102, and the “update time” of the acquired row is “ It is determined whether or not it is later than the “update time” (step S805). If the “update time” of the acquired row is earlier in time, the one with the newest update time is prioritized and the acquired row is discarded (step S809).

  If the “update time” of the acquired row is later in time, the reconnection synchronization means 104 determines whether or not the update information indicates deletion of the row (step S806). If the update information indicates deletion, the deletion flag of the corresponding row is set to “Y” (step S807). At this time, as shown in FIG. 3, when the synchronization information is separated from the data, the corresponding row is deleted from the data-side table. If deletion of a row is not indicated, the corresponding row in the relational database 102 is updated with the contents of the acquired row (step S810).

  In the example of FIG. 10, the table 1011 stored in the relational database 102 of the machine B110 includes a row corresponding to the update information 1021 (the row of the primary key “3”) transmitted from the machine A100. When the update times of these rows are compared, the row received from the machine A100 is newer. Therefore, the machine B110 updates the data of the primary key “3” from “data 11” to “data 17” according to the update information 1021 received from the machine A100. In addition, the synchronization information corresponding to this row is overwritten with the synchronization information of the update information 1021. By such an operation, while the machine B110 is disconnected from the network, the contents of the update performed on the machine connected to the network are reflected in the relational database 102 of the machine B110.

  When the reconnection synchronization unit 104 acquires update information including a plurality of rows in step S803, the reconnection synchronization unit 104 performs the processing of steps S804 to S810 for each of the included rows. In that case, it is desirable that the processing order be performed in ascending order of event numbers by comparing event numbers regardless of the updater. This is because if a failure occurs while updating a row with a large event number, it may be impossible to reacquire a row with a small event number.

  In the example of FIG. 10, in addition to the synchronization at the time of network reconnection with the update information receiving machine as machine B110 and the update information transmitting machine as machine A100, the update information receiving machine is set as machine A100 and updated. Synchronization at the time of network reconnection using the information transmission machine as the machine B110 is also performed. In synchronization at the time of network reconnection using the machine A100 as the update information receiving machine, the machine A100 searches the table 1001 of the relational database 102, and determines the maximum event number (“machine A, maximum event number 11”, Update information transmission request 1030 including “Machine B, maximum event number 7”, “Machine C, maximum event number 6”) is transmitted to Machine B110.

  The machine B110 refers to the relational database 102, extracts a row exceeding the maximum event number notified from the machine A100 for each updater, and transmits this to the machine A100. In FIG. 10, the machine B110 extracts the row of the primary key “1” (updater “machine B”, event number “11”) as a row exceeding the maximum event number notified from the machine A. This line is a line updated by the application 101 of the machine B110 while the machine B110 is disconnected from the network. The machine B110 encodes the extracted row of the primary key “1” and transmits it as update information 1031 to the machine A100.

  When the machine A 100 receives the update information 1031, it updates the relational database 102 according to the received update information 1031. More specifically, the data of the primary key “1” is updated from “data 5” to “data 9”. In addition, the synchronization information corresponding to this row is overwritten with the synchronization information of the received update information 1031. As a result, the contents of the update performed by the machine B110 during the network disconnection are reflected in the relational database 102 of the machine A100. As described above, when a newly communicable machine is discovered on the network, it is newly connected (reconnected) to the network by exchanging update information that each other does not know. ) And the machines already connected to the network can synchronize the contents of the relational database 102 in the same procedure.

  In the present embodiment, in correspondence with each row (each data) of the relational database 102, synchronization information including an updater who has updated the data and information for determining the update order for each updater is added. Each machine synchronizes the contents of the relational database 102 by the connecting synchronization means 105 during network connection. When reconnecting to the network, the reconnection synchronization means 104 notifies the machine that is newly communicable to the information indicating the latest update for each updater, requests transmission of update information, The contents of the update after the notified latest update are received, and the relational database 102 is updated. By doing so, the information passed between the machines is only information that the machines do not actually know, and redundant information is not exchanged. Therefore, for example, the relational database 102 can be synchronized more efficiently than when all information in the database is exchanged. Further, unlike the case where the prior art uses a log, by associating synchronization information with a row, it is possible to reduce the capacity of information used at the time of synchronization and to suppress the search cost.

  Here, consider a case where only “update time” is added as the synchronization information, and each machine stores “the time when synchronization was performed with any machine before”. In this case, it may be considered that only the difference information can be acquired by requesting information updated after “time when synchronization with any machine was performed before” as in the present embodiment. However, for example, in the state where the machine C is disconnected from the network, the update occurs at “1:00” in the machine C, and after that, the machine C and the machine B synchronize at “2:00”. Reconnects to the network and machine A and machine C synchronize, machine A is synchronized to “2:00”, so the update made at “1:00” A problem arises that C is not sent to machine A. In the present embodiment, since transmission of update information is requested according to the above procedure, it is ensured that update information is reliably transmitted between machines.

Second Embodiment FIG. 11 shows an operation procedure of synchronization during network connection in the distributed synchronous information system according to the second embodiment of the present invention. The distributed synchronization type information system of this embodiment has the same configuration as the distributed synchronization type information system shown in FIG. 1, and the connecting synchronization means 105 (FIG. 1) replaces the procedure shown in FIG. The procedure shown in FIG. 11 is different from the first embodiment in that the relational database 102 is synchronized with a machine connected to the network. In the present embodiment, the connecting synchronization unit 105 synchronizes the relational database 102 using the reconnection synchronization unit 104.

  When the application 101 updates the relational database 102, the connected synchronization means 105 of the updated machine notifies each machine connected to the network that the data has been updated (step S1101). At that time, the connecting synchronization means 105 includes the name of the machine that has updated the data in the notification that the data has been updated. Receiving this notification, the connecting synchronization means 105 of the other machines uses the update information transmission request means 106 of the reconnection synchronization means 104 to update its own device as an “update information receiving machine” and update the machine that has been updated. The information transmission side machine is requested to transmit update information (step S1102). In response to the request, the machine that has received the update information transmission request transmits the update information to the update information receiving machine in response to the request (step S1103). The update information receiving machine updates the relational database 102 according to the received update information.

  The update information transmission request and the transmission of update information in steps S1102 and S1103 are performed according to the same procedure as the synchronization at the time of network reconnection shown in FIGS. Thus, in the present embodiment, the system configuration can be simplified by performing synchronization during network connection and synchronization during network reconnection in the same procedure.

Third Embodiment FIG. 12 shows an operation procedure of synchronization at the time of network reconnection in the distributed synchronous information system according to the third embodiment of the present invention. The distributed synchronous information system of this embodiment has the same configuration as the distributed synchronous information system shown in FIG. 1, and the reconnection synchronization means 104 (FIG. 1) replaces the procedure shown in FIG. 12 is different from the first embodiment in that the relational database 102 is synchronized with a machine connected to the network in the procedure shown in FIG.

  The reconnection synchronization means 104 of the machine reconnected to the network (newly connected) selects any one of the machines already connected to the network, and designates the machine as “update information transmission side machine”, An update information transmission request is transmitted with itself as the “update information receiving machine”, and the relational database 102 is updated by the procedure shown in FIGS. 8 and 9 (step S1201). Assuming that data is synchronized between machines already connected to the network, the machine reconnected to the network in step S1201 is another machine that is connected to the network and is not its own device. The update can be reflected in the relational database 102 of the own device.

  When the reconnection synchronization means 104 of the machine reconnected to the network updates the relation database 102 of its own device, it selects an arbitrary one from the machines connected to the network, and the maximum event for each updater for that machine. It is instructed to transmit a number (step S1202). Upon receiving this instruction, the update information transmission request means 106 of the machine searches the synchronization information (relational database 102) in the machine, acquires the maximum event number for each updater, and issues an update information transmission request including this. , Send to a machine reconnected to the network.

  The update information transmission unit 107 of the machine reconnected to the network receives the maximum event number for each updater (step S1203), searches the relational database 102, and receives an event number exceeding the maximum event number for each received updater. A row is extracted (step S1204). This step corresponds to step S902 in FIG. Thereafter, the extracted row is encoded, and this is updated as update information and multicast transmitted to all machines connected to the network (step S1205). Each machine connected to the network receives the update information, and updates the relational database 102 by the same procedure as steps S803 to S810 in FIG. By this update, each machine that has already been connected to the network is a machine that has been reconnected to the network, and updates that are not made by the own apparatus can be reflected in the relation database 102 of the own apparatus.

  In the first embodiment, one-to-one communication is performed between a machine reconnected to the network and each machine connected to the network, and the relation database 102 of each machine is updated. Thus, when information is synchronized, useless communication is performed. In this embodiment, a machine reconnected to the network communicates with any of the machines connected to the network, receives update information, and updates the relational database 102. Also, the maximum event number for each updater is received from any of the machines connected to the network, and the update information is multicast-transmitted toward each machine connected to the network. By doing in this way, the frequency | count of communication and a communication fee can be reduced, and useless communication can be suppressed.

Fourth Embodiment A distributed synchronization information system according to a fourth embodiment of the present invention has the same configuration as the distributed synchronization information system 10 shown in FIG. In the present embodiment, “update history” is included in the synchronization information (FIG. 2) instead of “update time”. The update history includes a set of updaters who performed the update and a set of event numbers at that time, and shows a history of when the data was updated on which machine (at which event number). In this embodiment, such an update history is used to resolve inconsistencies when the same row is updated on a plurality of machines when disconnected from the network.

  FIG. 13 shows how the update history is updated. The in-database synchronization information updating unit 103 updates the update history instead of the update time in step S406 of FIG. 4 when updating the database. In the figure, “A3”, “B5”, and the like indicate the name of the machine that performed the update and the event number at that time. The update history of the row 1301 in FIG. 13 is updated when the row of the corresponding primary key “10” is the event number “3” in the machine A, and is updated thereafter when the event number is “5” in the machine B. Further, it is shown that the event number “7” is updated in the machine C. When machine A and machine B each have a row 1301 and machine B disconnects from the network and updates the row of the primary key “10”, the in-database synchronization information updating means 103 of machine B will change to row 1302 in FIG. As shown, “B8” indicating that the event has been updated at the event number “8” in the machine B is added to the update history of the row 1301.

  When the machine B reconnects to the network, the machine A receives the row 1302 as update information from the machine B by the synchronization operation at the time of the network reconnection. When the update history of this update information is compared with the update history of the row 1301 held by the machine A, the update history (A3, B5, C7, B8) of the synchronization information of the acquired update information is the row update of the machine A. The history (A3, B5, C7) is included. This means that row 1301 and row 1302 were not updated separately when disconnected from the network, but originally had the same information and were updated by machine B while it was disconnected from the network. It means that. As described above, when the update history of the acquired update information includes its own update history, the update is performed in a straight line as shown in an order 1303 in FIG. In such a case, there is no contradiction, and since the data in the row 1302 is found to be newer than the data in the row 1301, the machine A may update the row 1301 with the row 1302.

  FIG. 14 shows another example of how the update history is updated. The update history of the row 1401 shown in the figure is “A3, B5, C7, A5, A7”, and the update history of the row 1402 is “A3, B5, C7, B8”. Comparing both update histories, the update up to “C7” is the same as shown in the order 1403 in FIG. 14, and the subsequent updates are different. This means that after the network disconnection, the update is performed separately in the machine connected to the network and the machine disconnected from the network, or in the machine disconnected from the network. In this case, when the machine A and the machine B are reconnected to the network, some contradiction must be resolved. As a method for resolving inconsistencies, there are methods such as giving priority to a specific machine or letting the user select it.

  If any line is prioritized due to conflict resolution, it is treated as a newly updated line in the machine. For example, when the machine C has a row 1402 and the update of “B8” is given priority over the update of “A5, A7” for the rows 1401 and 1402, that is, when the row 1402 is given priority, “ Assume that the relational database 102 has been updated with the data 9 "as it is, the updater being" machine C "and the event number newly assigned to" 11 ". At this time, “C11” is added to the update history, and “priority: A5, A7” is added as a priority instruction indicating that “C11” is prioritized over the update of “A5, A7”. The reason why the priority instruction is added in this way is to prevent the user from requesting the selection again by holding the selection result when, for example, the user selects which update to give priority to. is there.

  Next, consider a case where machine D has row 1401 and this machine D is reconnected to the network. In this case, the machine D receives the row 1404 from the machine C as update information. When the update history of the row 1401 and the update history of the row 1404 are compared, the heads “A3” to “C7” are the same. Further, the part excluding the common part with the update history of the row 1404 from the update history of the row 1401 is “A5, A7”, and the update history of the row 1404 includes “priority A5, A7”. . Thus, when the update history of the row of the update information receiving side machine is the same as the update history of the update information, and the subsequent part is prioritized in the update history of the update information, the acquired update The information update history is treated as including its own update history. By doing in this way, the line including “A5, A7” in the update history can be updated with the line updated by “C11”.

  FIG. 15 shows an operation procedure of the machine reconnected to the network. The procedure shown in the figure is different from the operation procedure of the machine reconnected to the network in the first embodiment in that step S805 of FIG. 8 is replaced with step S1501 and step S1502 is added. In step S801, the update information transmission request unit 106 of the reconnection synchronization unit 104 of the update information receiving machine searches for the maximum event number for each update machine. At that time, the update information transmission request means 106 searches for the maximum event number for each update machine including the update history in addition to the event number of the synchronization information 201 (FIG. 2).

  When the reconnection synchronization unit 104 determines in step S804 that a row corresponding to the update information received from the update information transmitting machine exists in the relational database 102, the update history included in the synchronization information of the update information, The update history of the synchronization information of the corresponding row in the database 102 is compared (step S1501). If, as a result of the comparison, the update history of the acquired update history includes the update history of the corresponding row in its own relational database 102 or can be regarded as including the update history of the corresponding row by priority instruction, the contradiction is Since it has not occurred, the process proceeds to step S806, and the corresponding row in the relational database 102 is updated with the acquired update history row.

  For example, if the update information receiving machine has the row 1301 shown in FIG. 13 and receives the row 1302 as the update information, the update history of the row 1301 is included in the update history of the row 1302 in step S1501. To be judged. In this case, the update information receiving machine that has received the update information proceeds from step S806 to step S807, and updates the relational database 102 in the acquired row 1302. Further, when the update information receiving machine has the row 1401 shown in FIG. 14 and receives the row 1404 as the update information, the update history of the row 1401 is not included in the update history of the row 1404. Since the common part “A3, B5, C7” of the history is “A5, A7” specified by the update history priority instruction in the row 1404, in step S1501, the update history in the row 1401 is It is handled as being included in the update history 1402. Therefore, the row 1401 of the relational database 102 is updated by the update information row 1404.

  When the update history of the acquired update information and the update history of the corresponding row in the relational database 102 are not in an inclusive relationship, or when there is no inclusive relationship of the update history even if the update history priority instruction is referenced, the same row Has been updated separately on multiple machines, causing inconsistencies. In this case, the row is updated so as to resolve the contradiction (step S1502). In step S1502, for example, the user is allowed to select which of the contradicting rows is prioritized, and the relational database 102 is updated with the selected row. Alternatively, the update with the latest update time may be prioritized and updated in that row.

  In the update in step S1502, information for the current update is added to the update history with the updater as the own device and the event number newly assigned the event number for the priority row. At that time, a part of the update history of the unselected row excluding the common part with the update history of the selected row is included in the update history of the updated row as a priority instruction. For example, when priority is given to the row 1402 for the rows 1401 and 1402 in FIG. 14, “A5, A7” excluding the common portion with the update history of the row 1402 in the update history of the row 1401 As shown in FIG. 4, the history update “C11” for the current update is added as a priority instruction. When the update information receiving machine updates the relational database 102 in step S1502, the connected synchronization unit 105 transmits the updated row to each machine connected to the network.

  In the present embodiment, the synchronization information includes an update history indicating in what order (route) the update is performed. By using this update information, whether or not there is a contradiction by checking whether the update history of the update information acquired in the synchronization operation at the time of network reconnection includes its own update history, that is, receiving It can be determined whether the updated information and the corresponding row in the relational database 102 of the own device are updated at the same time after the network disconnection, or updated only on one side. In addition, when conflicts are resolved, the result is included in the update history, so that which update is prioritized when conflicts are resolved can be propagated to other machines connected to the network. Avoid making resolution decisions.

  Here, when the operation of the first or second embodiment and the operation of the present embodiment are combined, as described with reference to FIG. 6, each machine discovers a new communicable machine. Then, each performs synchronization as an “update information receiving machine”. Therefore, when a machine is already connected to the network and a machine is reconnected to the network, update information is received on all machines at the same time, and all machines need to perform conflict resolution. There is sex. On the other hand, when combining the operation of the third embodiment and the operation of the present embodiment, in the third embodiment, update information is multicast-transmitted to machines connected to the network. The priority update information can be propagated to each machine connected to the network, and the conflict resolution operation can be performed only once. Therefore, the operation of the present embodiment is preferably combined with the operation of the third embodiment.

Fifth Embodiment A distributed synchronization information system according to a fifth embodiment of the present invention has a history deletion means in addition to the same configuration as the distributed synchronization information system shown in FIG. In the fourth embodiment, the update history is included in the synchronization information, but this update history increases each time an update is performed. In this embodiment, each machine manages how far the update has been propagated, and deletes unnecessary portions of the update history. Further, in this embodiment, in each machine, it is managed up to which point the update has been propagated, and the row whose synchronization information deletion flag is “Y” is deleted.

  In the present embodiment, each machine stores information on the maximum event number for each updater possessed by each machine as information indicating at what point in time the update has propagated in each machine. FIG. 16 shows a specific example of the maximum event number stored in each machine for each updater possessed by each machine. For example, when the machine A stores the table 1601 shown in the figure, the machine A can update the updater “machine A” up to the event number “20” and the updater “ Machine B ”recognizes (understands) that up to event number“ 11 ”and updater“ machine C ”have updated up to event number“ 15 ”. In the relational database 102 of the machine B, the updater “machine A” is up to the event number “15”, the updater “machine B” is up to the event number “11”, and the updater “machine C” is the event. It is recognized that the number "7" has been updated.

  Creation of a table of maximum event numbers for each updater possessed by each machine stored in each machine will be described. The initial value of the maximum event number in the table 1601 at the start of system operation is 0. When each machine updates its own device relational database 102, the updater “self” of the row “maximum event number of own device” in the table 1601 of the maximum event number for each updater possessed by each machine stored in the own device. The value of “device” is rewritten to the event number newly assigned by the in-database synchronization information updating unit 103. For example, when the relational database 102 is updated in the machine A, the machine A updates the “Machine A Maximum Event Number” row updater “Machine A” in the table 1601 of the maximum event numbers stored in its own device. Rewrite the value of “A” to the newly assigned event number.

  When the data is updated on the machine A, the update is propagated to the machine connected to the network by the connecting synchronization means 105. When an update in another machine is propagated, each machine updates the value of the updater “machine A” in the row “maximum event number held by own device” in the table 1601 of the maximum event number stored in the own device. Rewrite with the event number propagated with the content. At that time, the machine connected to the network is determined, and the value of the updater “machine A” in the row of “the maximum event number of each machine connected to the network” is rewritten to the propagated event number. This is because the update on the machine A is propagated to each machine connected to the network. For example, a machine connected to the network can determine whether or not there is a response when data is transmitted. Alternatively, a machine that can periodically receive the life / death signal using the life / death signal may be determined as a machine connected to the network.

  For example, the machine B receives the update contents in the machine A from the machine A through a synchronization operation during network connection, and updates the relational database 102 with the updater as “machine A”. At this time, the machine B rewrites the value of the “updater“ machine A ”in the row of the maximum event number held by the machine B” in the table 1601 of the maximum event number stored in its own device to the event number propagated from the machine A. . Further, when the machine C is connected to the network, the update in the machine A is also propagated to the machine C. Therefore, the “maximum event number of the machine C” in the table 1601 of the maximum event number stored in the machine B is stored. The value of the updater “machine A” in the row is rewritten to the event number propagated from the machine A. When the machine C is disconnected from the network, the update in the machine A is not propagated to the machine C, so the value of “maximum event number held by the machine C” is not rewritten.

  In the operation at the time of network reconnection, when the update information transmission side machine encodes and transmits the row extracted in step S903 (FIG. 9), the table 1601 of the maximum event number for each updater of each machine stored is stored. Encode and send to the update information receiving machine along with the extracted lines. When the update information receiving machine acquires the update information in step S803 (FIG. 8), a table of the maximum event numbers of each machine received from the update information transmitting machine and the updater of each machine stored in its own device. The table of maximum event numbers to be stored is updated in combination with the table of maximum event numbers for each. Thereafter, the relational database 102 is updated by the synchronization operation at the time of reconnection, and the row of the maximum event number held by the own device in the table of maximum event numbers is updated along with the update.

  FIG. 17 shows how the maximum event number table of each machine is updated in the synchronization operation during reconnection. For example, assume that machine A stores table 1701 and machine B stores table 1702. Assuming that machine B is an update information receiving machine and machine A is an update information transmitting machine, machine B acquires table 1701 from machine A in synchronization at the time of reconnection. The machine B compares the value of each item of the stored table 1702 with the value of each item of the received table 1701, selects the larger value, and determines the maximum event number for each updater possessed by each machine. The table is updated to table 1703. For example, the maximum event number of the updater “machine B” possessed by the machine A in the table 1701 is “11”, and the maximum event number of the updater “machine B” possessed by the machine A in the table 1702 is “7”. In this case, “11” is selected.

  Deletion of a line (synchronization information) in which the deletion flag is set to “Y” will be described. The deletion flag is used to determine whether a row has been newly inserted in a machine disconnected from the network or a row has been deleted in a machine connected to the network during synchronization at the time of reconnection. Therefore, if it can be confirmed that the update for deleting a row has been propagated to all machines, the row itself can be deleted.

  In the table 1601 (FIG. 16) of the maximum event numbers for each updater possessed by each machine, when the minimum value is obtained for each updater, the machine A is “15”, the machine B is “7”, and the machine C is “7”. It becomes. This means that the machine storing the table 1601 recognizes that each machine has been updated up to at least these event numbers for each updater. That is, for all machines, the update with the updater as machine A is up to event number “15”, the update with the updater as machine B is up to event number “7”, and the update with the updater as machine C is up to event number “15”. 7 ”means that it has been propagated. Therefore, if the event number of the line for which the deletion flag is set to “Y” is a value equal to or smaller than the minimum value of the maximum event number of the machine that updated the line, the line can be deleted.

  Each machine has an event number of a row in which the deletion flag is set to “Y” and a maximum event for each updater possessed by each machine at an arbitrary timing, for example, periodically or when synchronizing with other machines. The number is compared with the minimum value of the machine that updated the line, and if the event number of the line is less than or equal to the minimum value of the maximum event number, the line is deleted. As shown in FIG. 3, when the data and the synchronization information are separated, the synchronization information corresponding to the row in which the deletion flag is set to “Y” is deleted. For example, if the updater of the row in which the deletion flag is set to “Y” is the machine C and the event number is “3”, the minimum value of the maximum event number of the updater “machine C” in the table 1601 is “ 7 ”, the row is deleted from the relational database 102.

  Next, deletion of the update history will be described. FIG. 18 shows a line before update history deletion and a line after update history deletion. Regarding the update history, there is no problem even if the update propagated to all machines is deleted from the update history. For example, when the machine A stores the table 1601 shown in FIG. 16, the machine A updates up to “A15”, “B7”, and “C7”, which are the minimum values of the maximum event numbers for each updater. Recognizes that it has propagated to all machines. Therefore, the update history deleting means deletes “A3, B5, C7” from the update history “A3, B5, C7, A20” of this row for the row 1801 shown in FIG. As a result, the update history in the row 1802 after deletion of the update history is only “A20”. The history deletion unit deletes an unnecessary part of the update history at an arbitrary timing, for example, periodically or at the time of synchronization with another machine.

  However, when two arbitrary devices try to synchronize data, the contents deleted from the update history may be different in each machine. This is because the minimum value of the maximum event number for each updater in the table of the maximum event numbers for each updater stored in each machine may be different for each machine. For example, when machine A stores table 1701 shown in FIG. 17 (the same content as table 1601 in FIG. 16) and machine B stores table 1702, in table 1701 stored by machine A, the updater “machine The minimum value of the maximum event number of “A” is “15”, and in the table 1702 stored by the machine B, the minimum value of the maximum event number of the updater “machine A” is “11”. This is because machine A recognizes that updates up to event number “15” have been propagated to all machines for updates for which its own device is an updater, and machine B updates updater “machine A”. Means that updates up to event number “11” are recognized as being propagated to all machines.

  FIG. 19 shows how the update history is deleted in the machine B. The row 1901 shown in the figure has the same primary key as the row 1801 shown in FIG. 18, and shows a state before “A20” is updated. When the machine B has stored the table 1702 shown in FIG. 17 as the maximum event number for each updater possessed by each machine, the minimum value of the maximum event number for the updater “machine C” is “6”. Therefore, when the machine B deletes the update history from the row 1901 shown in FIG. 19, the update history becomes “C7” as shown in the row 1902. In this case, if the machine A and the machine B try to synchronize according to the procedure shown in FIG. 15, the update history in the row 1802 (FIG. 18) and the update history in the row 1902 are not in an inclusive relationship. In spite of this, it is mistakenly recognized as an inconsistent update.

  Therefore, the update information transmission side machine notifies the update information reception side machine of what update has been deleted from the update history. Specifically, when the update information transmission side machine encodes and transmits the extracted row to the update information reception side machine, it stores the “table of maximum event numbers for each updater that each machine has” Encode and send. The update information receiving machine receives the "maximum event number table for each updater on each machine" received from the update information transmission machine, and the "maximum event number table for each updater on each machine" stored in its own device. The larger value of "" is selected, and the "table of maximum event numbers for each updater that each machine waits" is updated. Thereafter, for the updated table, the minimum value of the maximum event number is obtained for each updater.

  In the example of FIG. 17, the table 1701 stored by the machine A and the table 1702 stored by the machine B update the “table of maximum event numbers held by each machine” as shown in the table 1703. Is done. In this case, the minimum value of the maximum event number is “A15”, “B7”, and “C7” from “A11”, “B7”, and “C6” (Table 1702). When the update information receiving machine compares the update histories in step S1501 (FIG. 15), it updates the maximum event number of each updated machine from the update history of the row corresponding to the update information in the relational database 102 of the own device. The update history is deleted using the minimum value of the maximum event number for each updater in the table. Then, the update history of the update information is compared with the update history of the row corresponding to the update information in the relational database 102 to check whether or not a contradiction has occurred.

  For example, when the machine B stores the row 1902 shown in FIG. 19 and receives the row 1802 shown in FIG. 18 as the update information, the updated table 1703 is referred to, and “C7” is obtained from the update history of the row 1902. Is deleted. When the row 1903 from which the update history has been deleted in this way is compared with the update history of the update information, the update history of the row 1903 is “empty”. Therefore, in step S1501, it is not determined that there is a contradiction, and the step Proceeding to S810, the data is updated. If the update information receiving machine has updated the data in step S810, the update information receiving machine corresponds to the update, and the row of the maximum event number of its own device in the "table of maximum event numbers for each updater of each machine" Update. For example, when the machine B updates the row 1903 with the row 1802, the machine B sets the value of the updater “machine A” in the row of the maximum event number held by the machine B in the table 1703 (FIG. 17) to “ 20 ”.

  In this embodiment, each machine memorizes what update has been propagated to each machine for each updater, refers to that information, has already propagated to all machines, and is no longer needed Or, the line (synchronization information) in which the deletion flag is set to “Y” is deleted. By doing so, the storage capacity necessary for storing the update history and the synchronization information can be reduced, and the storage capacity of the storage device can be used effectively.

Sixth Embodiment FIG. 20 shows the configuration of a distributed synchronous information system according to a sixth embodiment of the present invention. In the present embodiment, the distributed synchronous information system 10a is composed of a plurality of computers (three computers 2001 to 2003 in the example shown in the figure) connected via a network. Recording media 2011 to 2013 are connected to each computer, and each recording medium is connected to the application 101 shown in FIG. 1, in-database synchronization information updating means 103, reconnection synchronization means 104, and connected synchronization. A program that functions as the means 105 and a relational database 102 are stored. Each computer reads the program and operates to synchronize the relational database 102 by the operations described in the first to fifth embodiments.

  In FIG. 2, the synchronization information includes “event number” as information for determining before and after update, and “update time” as information used when determining which update should be prioritized. When “update time” is used as information for determining before and after the update, information for determining before and after the update and information for determining which update should be given priority may be shared. it can. When the update time is used as the information for determining before and after the update, the maximum event number described in each of the above embodiments may be read as the latest update time. Further, an event number portion after a certain event number may be read as an update after a certain update time.

  As mentioned above, although this invention was demonstrated based on the suitable embodiment, the information apparatus of this invention, a distributed synchronous type | mold information system, the information synchronization method, and a program are not limited only to the said embodiment example. Those modified and changed from the configuration of the above embodiment are also included in the scope of the present invention.

  The present invention can be applied to applications such as personal computers, information appliances, mobile phones, and PDAs that synchronize information used by them. Further, the present invention can be applied to a use for synchronizing information in a database between a group of servers connected via a network.

1 is a block diagram showing a configuration of a distributed synchronous information system according to a first embodiment of the present invention. The figure which shows the specific example of the data structure of a relational database. The figure which shows another specific example of the data structure of a relational database. The flowchart which shows the operation | movement procedure of the synchronous information update means in a database. The flowchart which shows the operation | movement procedure of the synchronization means during a connection. The flowchart which shows the operation | movement procedure of the synchronization means at the time of reconnection. The block diagram which shows the mode of the update of the relational database at the time of network reconnection. The flowchart which shows the detailed procedure of the update process at the time of the network reconnection of an update information receiving side machine. The flowchart which shows the detailed procedure of the update process at the time of the network reconnection of an update information transmission side machine. The figure which shows the specific example of the information exchanged between two machines in the update process at the time of network reconnection. The flowchart which shows the operation | movement procedure of the synchronization in the network connection in the distributed synchronous information system of 2nd Embodiment of this invention. The flowchart which shows the operation | movement procedure of the synchronization at the time of network reconnection in the distributed synchronous information system of 3rd Embodiment of this invention. The figure which shows the mode of an update history update. The figure which shows another example of the update of update history. The flowchart which shows the operation | movement procedure of the machine reconnected to the network. The figure which shows the specific example of the maximum event number for every updater which each machine memorize | stores. The figure which shows the mode of the update of the table | surface of the largest event number which each machine has in the synchronous operation | movement at the time of reconnection. The figure which shows the line before update history deletion, and the line after update history deletion. The figure which shows the mode of deletion of the update history in the machine B. The block diagram which shows the structure of the distributed synchronous information system of 6th Embodiment of this invention. The figure which shows the mode of the information synchronization in the technique of a nonpatent literature 1.

Explanation of symbols

10: Distributed synchronous information system 100, 110, 120: Information device 101: Application 102: Relational database 103: In-database synchronization information update means 104: Reconnection synchronization means 105: Connected synchronization means 106: Update information transmission request means 107: Update information transmission means 201: Synchronization information 2001, 2002, 2003: Computer 2011, 2012, 2013: Recording medium

Claims (30)

In a distributed synchronous information system that includes a plurality of information processing devices each having a database, and synchronizes the database among the information processing devices,
Each of the information processing devices
A storage device that associates and stores each data of the database, an updater indicating an information processing device that has performed data update, and synchronization information including information for determining before and after the update in the information processing device And an update information transmission request means for transmitting an update information transmission request including information indicating the latest update for each updater to another information processing apparatus,
Update data updated by a newer update than the latest update for each updater included in the update information transmission request and synchronization information corresponding to the update data from another information processing apparatus that has received the update information transmission request. A distributed synchronization type information system comprising: first database synchronization means for receiving update information including and updating the database and synchronization information in the device itself.   When the database is updated as an updater in the own device, the database further includes synchronization information updating means for creating information for determining before and after the update in the own device and updating the synchronization information in correspondence with the update data. The distributed synchronous information system according to claim 1.   The information for judging before and after the update in the information processing apparatus is an event number for each information processing apparatus that is incremented each time the database is updated in the information processing apparatus. The distributed synchronous information system described.   When the synchronization information includes an update time, and the first database synchronization means determines that the update time of the synchronization information included in the update information is newer than the update time of the synchronization information stored in its own device, The distributed synchronous information system according to claim 1, wherein information is updated.   The synchronization information according to any one of claims 1 to 4, wherein the synchronization information further includes an update history including a pair of an information processing apparatus that has been updated so far and information for determining before and after updating in the information processing apparatus. A distributed synchronous information system according to claim 1.   The first database synchronization means compares the update history included in the received update information with the update history stored in the own device, and the update history included in the received update information stores the update history stored in the own device. The distributed synchronous information system according to claim 5, wherein the database and the corresponding synchronization information are updated if it is determined that the database includes the database.   Updates reflected in the database of each information processing apparatus are reflected in the database of all information processing apparatuses from the update history with reference to update propagation information storing information indicating the latest update for each updater. The distributed synchronous information system according to claim 5 or 6, further comprising history deletion means for deleting a history before update. When updating the database in its own device, update propagation means for transmitting update data and synchronization information corresponding to the update data to other information processing devices connected to the network;
8. The apparatus according to claim 1, further comprising: second database synchronization means for receiving update data and synchronization information transmitted by update propagation means of another information processing apparatus and updating the database and synchronization information. Distributed synchronous information system described in 1.   When receiving an update information transmission request from another information processing device, refer to the synchronization information stored in the own device, search for an update that is newer than the latest update for each updater included in the received update information transmission request, Update information transmitting means for transmitting update information including update data extracted by the search and synchronization information corresponding to the update data to the information processing apparatus that has transmitted the update information transmission request; The distributed synchronous information system according to any one of claims 1 to 8, further comprising:   When the information processing apparatus reconnected to the network receives an update information transmission request from one of the other information processing apparatuses connected to the network, the update information transmitting unit of the reconnected information processing apparatus The distributed synchronous information system according to claim 9, wherein the update information is multicast-transmitted to another information processing apparatus connected to the network. In a database synchronization method in a distributed synchronization information system, each comprising a plurality of information processing devices each having a database, and synchronizing the database among the information processing devices,
When the information processing apparatus updates the database, update data, an updater indicating the information processing apparatus that performed the data update, and synchronization information including information for determining before and after the update in the information processing apparatus Storing in association with each other;
The information processing apparatus searches for the synchronization information and transmits an update information transmission request including information indicating the latest update for each updater to another information processing apparatus;
Update data updated by a newer update than the latest update for each updater included in the update information transmission request from another information processing apparatus that has received the update information transmission request, and the update data Receiving update information including synchronization information corresponding to
The information processing apparatus includes a step of updating the database and synchronization information in the own apparatus based on the received update information.   In the step of storing the synchronization information, when the information processing apparatus updates the database as an updater in the own apparatus, the information processing apparatus creates information for determining before and after the update in the own apparatus, and stores the update data in the update data. The method according to claim 11, wherein the synchronization information is updated correspondingly.   The information for judging before and after the update in the information processing apparatus is an event number for each information processing apparatus that is incremented each time the database is updated in the information processing apparatus. The method described.   In the step in which the synchronization information includes an update time, and the database and the synchronization information are updated, the information processing apparatus determines that the update time of the synchronization information included in the update information is greater than the update time of the synchronization information stored in the own apparatus. The method according to claim 11, wherein the database and the synchronization information are updated when new.   15. The synchronization information according to any one of claims 11 to 14, wherein the synchronization information further includes an update history including a pair of an information processing apparatus that has performed an update and information for determining before and after the update in the information processing apparatus. The method described.   The information processing apparatus further includes a step of comparing an update history included in the received update information with an update history stored in the own apparatus, and the information processing apparatus includes an update history included in the received update information. The method according to claim 15, wherein the database and the corresponding synchronization information are updated when it is determined that the update history stored in the device is included.   For the update reflected in the database of each information processing apparatus, the information processing apparatus refers to update propagation information storing information indicating the latest update for each updater, and from the update history, all information processing apparatuses 17. The method according to claim 15 or 16, further comprising the step of deleting the history before update reflected in the database. When the information processing apparatus updates the database in its own apparatus, transmitting update data and synchronization information corresponding to the update data to another information processing apparatus connected to the network;
The information processing apparatus further includes a step of receiving the update data and the synchronization information transmitted in the step of transmitting the update data and the synchronization information corresponding to the update data, and updating the database and the synchronization information. Item 18. The method according to any one of Items 11 to 17. When the information processing apparatus receives an update information transmission request from another information processing apparatus, it refers to the synchronization information stored in its own apparatus, rather than the latest update for each updater included in the received update information transmission request. Searching for new updates;
A step in which the information processing apparatus transmits update information including update data extracted by the search and synchronization information corresponding to the update data to the information processing apparatus that has transmitted the update information transmission request; The method according to claim 11, further comprising:   When the information processing apparatus reconnected to the network receives an update information transmission request from one of the other information processing apparatuses connected to the network, the update information is transmitted to the other information processing apparatus connected to the network. The method of claim 19, further comprising the step of multicasting towards the device. In an information processing apparatus having a database and synchronizing the database with another information processing apparatus connected via a network,
A storage device that associates and stores each data of the database, an updater indicating an information processing device that has performed data update, and synchronization information including information for determining before and after the update in the information processing device And an update information transmission request means for transmitting an update information transmission request including information indicating the latest update for each updater to another information processing apparatus,
Update data updated by a newer update than the latest update for each updater included in the update information transmission request and synchronization information corresponding to the update data from another information processing apparatus that has received the update information transmission request. An information processing apparatus comprising: first database synchronization means for receiving update information including and updating the database and synchronization information in the own apparatus. A program for synchronizing a database of a computer with a database of another computer connected via a network, the computer having:
Storing the update data, the updater indicating the computer that performed the data update, and the synchronization information including information for determining before and after the update in the computer, in association with the database update;
Searching the synchronization information and transmitting an update information transmission request including information indicating the latest update for each updater to another computer;
An update including update data updated by an update newer than the latest update for each updater included in the update information transmission request and synchronization information corresponding to the update data from another computer that has received the update information transmission request Receiving information; and
A program for executing the step of updating the database and the synchronization information in the device based on the received update information.   The program according to claim 22, wherein the information for determining before and after updating in the computer is an event number for each computer that is incremented each time the database is updated in the computer.   In the step of updating the database and the synchronization information, the synchronization information includes an update time, and if the update time of the synchronization information included in the update information is newer than the update time of the synchronization information stored in the device, the database The program according to claim 22 or 23, wherein the synchronization information is updated.   The program according to any one of claims 22 to 24, wherein the synchronization information further includes an update history including a set of a computer that has performed an update and information for determining before and after the update in the computer.   Further causing the computer to perform a step of comparing the update history included in the received update information with the update history stored in the device, and in the step of updating the database and the synchronization information, the step of comparing the update history 26. The program according to claim 25, wherein when the update history included in the received update information is determined to include the update history stored in the device, the database and the corresponding synchronization information are updated.   The update reflected in the database of each computer is referred to the update propagation information storing information indicating the latest update for each updater, and is reflected in the database of all computers from the update history. 27. The program according to claim 25, further comprising executing a step of deleting a history before update. In the computer,
When updating the database in its own device, transmitting update data and synchronization information corresponding to the update data to another computer connected to the network;
Receiving the update data and the synchronization information corresponding to the update data transmitted in the step of transmitting the update data and the synchronization information corresponding to the update data, and further executing the step of updating the database and the synchronization information; The program according to any one of claims 22 to 27. In the computer,
When receiving an update information transmission request from another computer, referring to the synchronization information stored in the own device, searching for an update that is newer than the latest update for each updater included in the received update information transmission request;
The update information including update data extracted by the search and synchronization information corresponding to the update data is further transmitted to the computer that has transmitted the update information transmission request. The program according to any one of 22 to 28. In the computer,
When the computer is reconnected to the network and receives an update information transmission request from one of the other computers connected to the network, the update information is multicast transmitted to the other computers connected to the network. 30. The program according to claim 29, further executing the step of:

Images