JP2005071238A - File sharing system and terminal device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a file sharing system that solves disadvantages of a server type file sharing method and enables continuous use of the latest file even if a terminal device shuts down by an asynchronous event such as a hardware failure. <P>SOLUTION: In the file sharing system, a plurality of terminal devices connected to an information communication network (10) are logically connected to a semantic information network (20), and the terminal devices (31 and 32) share files. Each terminal device (31, 32) has storing means (312, 322) for storing files to be shared, and means (3014) for storing management information including duplicate transfer path information on the files. The files to be shared are distributed over the semantic information network (20), and file duplicate paths are topologically configured with the file duplicate transfer path information stored in each terminal device. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a system for sharing a file with a plurality of terminal devices via a transmission medium such as a LAN or the Internet network, and the latest file can be continuously shared even when the plurality of terminal devices are simultaneously terminated due to a failure or the like. The fault-tolerant file sharing system, the file sharing synchronization control method, the non-end device in which these are implemented, and the terminal program.

In a conventional file sharing system, there is a method for centrally managing files to be shared by a central server (hereinafter referred to as a server-type file sharing method). However, the following problems have been pointed out for centrally managing shared files in a server system.
(1) As the number of client terminals increases, the processing load also concentrates on the server system. (2) The file itself may be damaged due to a server system failure.
(3) There is a problem that it is expensive to maintain and operate the server system.
(4) Although a form in which a plurality of file sharing servers are distributed and arranged on the network is also conceivable, it is necessary to transfer lock requests and update data to all servers, and not only takes time to update data. It is assumed that the traffic consumption will increase.

  On the other hand, Takanari Hoshiai invited lecture “New Trend of Internet“ Non-Broker Model ”and its Realization Technology“ SIONet: Semantic Information Network ”SSE2000-235 & IN2000-191, PP65 Semantic information networks that can be identified and that enable end users to identify optimal information are available. The following file sharing systems have been proposed by the present applicant as file sharing systems related to the semantic information network. One proposal is to move shared files between terminals every time they are accessed in response to the problem that a large amount of storage space is required in a way that all terminals hold shared files in the network and perform synchronous control. This makes it possible to reduce the storage area amount (see, for example, “Patent Document 1”). As another proposal, by providing a device for backing up a shared file in a network, it is possible to continue file sharing even after termination of the terminal device (see, for example, “Patent Document 2”). When the terminal device is terminated due to an asynchronous event such as a hardware failure, there is a problem that the old generation file in the backup device is restored.

Japanese Patent Application No. 2003-42909 Japanese Patent Application No. 2003-42743

  As described above, in the conventional server method, in proportion to the increase in the number of files and the number of terminals to be shared, a large-scale server must be installed, which is expensive and difficult to deal with in real time. There is a problem. In addition, the method using the semantic information network without using the server eliminates the problems caused by using the server. However, if the terminal device is terminated due to an asynchronous event such as a hardware failure, the old generation file in the backup device is restored. This makes it impossible to continue file sharing.

Accordingly, an object of the present invention is to realize a file sharing system that eliminates the drawbacks of the server-type file sharing method.
Furthermore, another object of the present invention is to realize a file sharing system that can continuously use the latest file even when the terminal device is terminated due to an asynchronous event such as a hardware failure.

In order to solve the above problems, a file sharing system according to the present invention includes a semantic information network in which a plurality of terminal devices connected to an information transmission network are logically connected to a semantic information network, and each terminal device shares a file. A file sharing system used,
Each terminal device has storage means for storing a file to be shared, and means for storing management information including copy transfer path information of the file,
The file to be shared is distributed and arranged on the semantic information network, and the file duplication path is constructed topologically using the file duplication transfer path information stored in each terminal device.

  In the present invention, a plurality of terminal devices sharing a file are logically connected to a semantic information network. Since each terminal device has storage means for storing the file to be shared, the file to be shared is distributed and arranged on the semantic information network. As a result, even if some terminal devices leave the network due to a failure or the like, each terminal device can share the latest file.

  In a preferred embodiment of the file sharing system according to the present invention, when a terminal device connected to a network updates data for a shared file, the lock request is transferred to the terminal device defined based on the file replication transfer path information. Then, the data to be updated is transferred after locking the use of the file in the terminal device, and unlock release is transferred after the transfer of the update data is completed. In the file sharing system according to the present invention, when updating a file by editing data or the like, it is only necessary to transfer the update data only to the terminal device defined based on the file replication transfer path information, which causes a problem of concentrated traffic. It will be resolved. In addition, the lock request is not transferred to all terminal devices, but only the lock request is transferred to some terminal devices, which eliminates problems caused by the use of shared files being locked. Is done.

  In the preferred embodiment of the file sharing system according to the present invention, each terminal device is connected to the semantic information network via a shared link, and even if some terminal devices are disconnected from the semantic information network due to a failure, Each of the terminal devices is connected to a semantic information network by a shared link self-organization function. Since the shared link that constructs the semantic information network has a self-organizing function, even if some terminal devices are disconnected from the network due to failure or the like, the remaining terminal devices are connected to the semantic information network. The update data of the file is sequentially transferred between the terminal devices based on the replication transfer path constructed topologically, and each terminal device can share the latest file.

In another preferred embodiment of the file sharing system according to the present invention, when there is a file update request from a terminal device, the terminal device should be updated based on the copy transfer path information held in each terminal device. Data is sequentially transferred to other terminal devices. Each terminal device retains the copy transfer path information of the file and sequentially transfers the update information based on the transfer path information. Therefore, when there is a need to update the file data, the update data is updated for all peers at once. Are not transferred, but are sequentially transferred via Peers constituting the semantic information network, so that the amount of communication for update access can be reduced.
Further, in the file sharing method using the semantic information network of the present invention, when the peer leaves, the file information is exchanged and collated between the peers, so that the peer removal can occur in an overwave state during the lock / data update request / reference request processing. Normal operation can continue even if it occurs).

The effects of the present invention are as follows.
(1) A file management system that eliminates the cost of equipment and maintenance of the central server system can be constructed.
(2) By distributing the number of replicas in the network, the probability of returning to the old file even if some peers leave can be reduced.
(3) By performing processing of lock / data update request / data reference request as shown in the first and second embodiments at each peer, normal data reference / update is possible without operating the entire copy. It becomes.

  FIG. 1 is a diagram showing an example of a file sharing system of the present invention. For example, a plurality of terminal devices 31 and 32 are physically connected to the transmission network 10 such as IP. Although only two terminal devices are shown in the drawing, it is assumed that a large number of terminal devices are connected. These terminal devices are logically connected to a semantic information network (Semantic Information-Oriented Network, hereinafter referred to as SIONet) 20. Although the terminal device A31 such as a PC of the first user 41 of the file sharing system is described as the first terminal device 31 and the terminal device of the second file sharing system user 42 is described as the terminal device B32 separately for explanation. It does not mean that the configurations of these terminal devices are different. When it does not matter which terminal device it is, it is described as a terminal device 30.

  The terminal device 31 includes a file sharing application program 311 and a storage device 312 that stores a shared file. The terminal device 32 also includes a file sharing application program 321 and a storage device 322 that stores the shared file. In these terminal apparatuses 31 and 32, the file sharing application program is installed in the event place (in its own terminal apparatus by installing programs such as SI-SW, SI-R, etc. for constructing a SIONet and a program that functions in cooperation with SIONet. Is connected to SIONet by creating an EP) or installing a program that works in conjunction with SIONet and accessing an event place (EP) on another network to establish a session or shared link Assuming that.

The present invention is characterized in that the file contents are continuously guaranteed even when a plurality of peers (terminal devices) are simultaneously terminated while maintaining the consistency of the file contents discretely distributed on the semantic information network. is there. These features will be described with reference to FIGS.
FIG. 2 is a functional block diagram viewed from software according to the embodiment of the terminal device. The terminal device 30 can be applied as both the terminal device 31 of the first file sharing system user 41 and the terminal device 32 of the second file sharing system user 42. The terminal device 30 includes an operation system (OS) 301, a network connection control unit 302 that controls connection between the transmission network 10 and the semantic information network 20, and a file sharing application program. And a file management control unit 303 for controlling updating and the like.

  The network connection unit 302 of the terminal device 30 installs a program that functions in cooperation with SIONet and a program for constructing SIONet on the OS 301 such as UNIX (registered trademark) or Windows (registered trademark). By creating an event place (EP) on the terminal or installing a program that works in cooperation with SIONet and accessing the event place (EP) on other network nodes to establish a session or shared link, Connected to SIONet. The network connection unit 302 is physically connected to the transmission network 10.

  The file sharing application program 3011 functions as a service entity (SE) 3012 that receives information input / output by the file sharing system user and inputs / outputs signals such as screen display, and receives file operation requests from the service entity 3012. The API processing unit 3013 that checks the parameter content, the file operation request received by the API processing unit, the file management function unit 3014 that provides the Open / Read / Write function for each file, and the Read / Write to the file, The exclusive control function unit 3015 executes an exclusive control function (for example, a lock request or the like) for ensuring the consistency of files distributed and deployed in the semantic information network. The function of each part makes it possible to distribute and deploy files within the semantic information network in consideration of access efficiency and reliability, and to perform file synchronization control to ensure consistency of the file contents themselves.

  FIG. 3 shows an example of the update of file file contents distributed in the semantic information network and their arrangement. In the file sharing system according to the present invention, each terminal device that holds a file holds not only the file but also the copy transfer path information of the file as file management information. Then, the file and update replication paths are managed topologically. Further, it is assumed that each terminal device has the replication path topology information in the file management function unit. First, an example of file replication will be described with reference to FIG. Peer1 is a root terminal device having an original copy. Peer1, which is an original holding terminal device, transmits a copy of the file to Peer4 and Peer2 based on the file copy transfer path information. Then, Peer 2 transmits a copy of the file to Peer 7 and Peer 8 based on the stored file copy transfer path information. Further, Peer2 also sends a copy of the file to Peer6. In this way, the file is stored in all peers connected to the semantic information network and making file request access. Next, file update will be described. Peer 6 now sends update information to Peer 2 in order to distribute the latest file on the network. Peer 2 transmits the copied update information to Peer 1, Peer 1 transmits further update information to Peer 4, and further transmits to Peer 7 and Peer 8. Thus, by transmitting update information according to the constructed topology, it is possible to reduce the amount of processing and the amount of communication required for access. By deploying a copy of the file to the peer that requested the file access at the time of file access from the service entity, such discrete distributed deployment becomes possible. In addition, there is a method of limiting the maximum number of replicas for the purpose of reducing the processing related to the synchronous control of the shared file (the processing is reduced but the probability that the state returns to the old file when Peer leaves).

  FIG. 4 is an example of management information of a shared file generated at the time of file duplication. The file name / path name of the shared file, the maximum number of duplicate files, the number of duplicate files existing in the semantic information network, and the origin of duplication The ID of the peer that held the file, the peer ID list of the distribution destination that distributed the copy, the edited peer ID, the data in the peer where the corresponding file exists, including the data managing the peer ID to which the lock was distributed, etc. Saved in. Here, the duplication source PeerID and the duplication destination PeerID prescribe the duplication route topologically.

  FIG. 5 shows a configuration example of the file management information illustrated in FIG. 4 based on the file distribution and deployment example of FIG.

FIG. 6 is a schematic diagram for explaining a method of performing access to a shared file distributed as shown in FIGS. 3 and 5 (for example, updating to a file) as a first embodiment and a second embodiment. In this example, an example is shown in which the topology of the replication route of each terminal device is constructed hierarchically.
The first embodiment exemplifies a case where an access to the shared file (for example, an update access request for the shared file) is generated from any one of the shaded peers in the figure. The access request source Peer transmits access based on the copy path information included in the file management information. A lock order is transmitted from the file management information to the copy source PeerID of the file and the PeerID selected at random from the copy destination to lock access to the file. Then, the information to be updated is transmitted. If the peer that received the lock order and updated the file is the lock order received from the copy source, the lock order is transferred from the copy destination to, for example, a randomly selected PeerID, and if the lock order is received from the copy destination, Distribute the lock order to the copy source. At this time, as a condition for returning a response signal to the lock, consistency of the lock processing is ensured by triggering reception of the lock response signal from the peer of the distribution destination. After these lock processes are completed, the shared file is updated. Update data is distributed to the locked range, and the data is updated after updating the last update time. Then, unlock release is transferred.

  When referring to data, the original holding Peer (Peer whose duplication source PeerID is NULL) is set to a filter that fires at all times, and all the Peers other than the original holding Peer are set to the file management information to be held. It is assumed that a filter that fires when a time newer than the last update time is set is set. The peer requesting the data reference transmits an event in which the last update time of the shared file held by the peer is set, and when the response source peer for the event is the original holding peer, the last update time of the original holding peer is the own peer. If the file update time is newer than the update time of the file held in step 1, access is made after transferring the file contents from the original peer. If the response is from a person other than the original holding peer, the file is unconditionally transferred from the peer and referenced.

  The second embodiment exemplifies a case where an access to the shared file occurs from any one of the shaded peers in the figure. The access request source Peer transmits a lock order to the Peer ID randomly selected from the file management information, for example, the copy source Peer ID of the file and the majority of copy destination peers. If a peer that has received a lock order is receiving a lock order from the replication source, a majority of peers are transmitted from the replication destination to, for example, a randomly selected PeerID. If a lock order is received from the replication destination, A lock order is distributed to, for example, a peer ID selected at random from a majority of peers including the received peers. At this time, the return condition of the response signal to the lock, the data update process, and the reference processing method are the same as in the first embodiment.

  FIG. 7 is a schematic diagram of a processing example when Peer leaves (for example, termination due to an asynchronous event due to a failure or the like) during the transfer of the lock / data update request. In this example, it is assumed that a lock / data update request is transferred from Peer1 to Peer2 and Peer3, and Peer2 is disconnected. When such a disconnection occurs, Peer # 1 in the figure continues to wait for a response from Peer # 2 and cannot continue processing, but it uses the SIONet shared link as the connection method between Peers, and its self-organizing function Is used, a shared link between Peer # 1 and Peer # 3 is newly reproduced, and the lock delivery status data held in the file management information is collated between Peer # 1 and Peer # 3, and the delivery status is confirmed. If there is an inconsistency, normal processing can be continued by resending the lock data update request. That is, in the file sharing system according to the present invention, since each Peer has replication path information, even when peers that are topologically adjacent are separated, the shared link is reproduced by the self-organization of the semantic information network, and continuously connected. The reconstructed topology is reconstructed and the latest file can be used continuously.

It is a diagram showing an example of a file sharing system according to the present invention. It is a diagram which shows an example of a terminal device. It is a diagram showing an example of sequential transfer of files. It is a diagram which shows an example of file management information. It is a diagram which shows an example of the file management information after file generation. It is a diagram which shows the example of access to a shared file. It is a diagram which shows the process example at the time of Peer detachment | leave.

Explanation of symbols

10 Transmission network 20 Semantic information network (SIONEt)
30, 31, 32 Terminal devices 41, 42 File sharing system users 301, 311, 321 File sharing application program 301 OS
302 Network Connection Control Unit 303 File Management Control Units 312 and 322 Storage Device

Claims (7)

A file sharing system in which a plurality of terminal devices connected to an information transmission network are logically connected to a semantic information network, and each terminal device shares a file,
Each terminal device has storage means for storing a file to be shared, and means for storing management information including copy transfer path information of the file,
A file sharing system in which files to be shared are distributed on a semantic information network, and a file replication path is topologically constructed by file replication transfer path information stored in each terminal device .   2. The file sharing system according to claim 1, wherein a terminal device connected to the network transfers a lock request to a terminal device defined based on the file replication transfer path information when updating data for the shared file. A file sharing system, wherein the data to be updated is transferred after locking the use of the file in the terminal device, and unlock release is transferred after the transfer of the update data is completed.   3. The file sharing system according to claim 2, wherein the terminal device that has received the transfer of the update data transfers the received update data to the terminal device defined by the file replication transfer path information, and updates the data in the network. A file sharing system characterized by transferring files in a chain.   4. The file sharing system according to claim 1, wherein each of the terminal devices is connected to a semantic information network via a shared link, and some of the terminal devices are disconnected from the semantic information network due to a failure. In this case, the remaining terminal device is connected to the semantic information network by the self-organization function of the shared link.     5. The file sharing system according to claim 1, wherein the terminal devices are connected in a hierarchical structure on a semantic information network. 6. An information transmission network is a terminal device used in a file sharing system that is physically connected and logically connected to a semantic information network and shares the same file with a plurality of terminal devices,
Storage means for storing files to be shared, means for storing management information including file copy transfer path information,
When updating the file, update the stored file using the transferred file update data and transfer the file update data to a predetermined terminal device based on the file replication transfer path information. Characteristic terminal device.   7. The terminal device according to claim 6, wherein when the received file update data is transferred to another terminal device, the terminal device transfers a lock request to the other terminal device and is stored in the other terminal device. The file update data is transferred after the file being locked is locked.

Images