JP2008304980A - Data file distribution system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To surely achieve data file distribution within a time limit in a system for achieving data file distribution in a P2P system. <P>SOLUTION: This data file distribution S includes a server device 1 as a data file distribution source and a plurality of clients (clients 21A to 21C for cedars, clients 22A to 22C). Those server device 1 and clients are network-connected through an Internet line, and data file distribution is performed in a P2P system. The server device 1 manages the distribution circumstances of the data file to each client, and when the client whose distribution is predicted to be delayed is specified, an undistributed data file is directly distributed to the client. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a data file distribution system for distributing various data files such as images, video, and audio contents through a network.

  As a form of a network system for connecting a plurality of computer apparatuses to a network, a client server (C / S) system and a peer-to-peer (Peer to Peer = P2P) system are known. The C / S system is a system in which the master-slave relationship between the server and the client is fixed. For example, the server provides a service for distributing data files, and the client uses this service.

On the other hand, in a P2P system, each peer (for example, a personal computer) connected to a network operates like a server by providing a predetermined service to other peers at one time. This is a system that operates like a client by using a service provided by another peer, and each peer is an equivalent system. According to such a system, there is an advantage that information resources such as data files and computing power can be shared between personal computers. Various technologies have been proposed for such a P2P system. For example, Patent Document 1 discloses a system in which a data file is segmented and distributed from a server and the segmented file can be transferred between peers. Proposed.
JP 2004-213508 A

  However, when attempting to distribute a data file to each peer using a single peer (file server) as a source in the P2P system, a delay in file distribution may occur to some peers. This is because in the distribution using the P2P system, it is more efficient to operate the system in accordance with the load of the peer, the distribution environment, etc., and it is difficult to perform distribution control. Accordingly, there is a problem that the P2P system has low reliability when there is a request for completing the distribution of the data file to each peer by a certain time limit.

  The present invention has been made in view of the above problems, and it is an object of the present invention to ensure that data file delivery can be reliably performed on time in a system capable of data file delivery in the P2P method. To do.

  A data file distribution system according to one aspect of the present invention that achieves the above object includes a server device that is a data file distribution source and a plurality of clients, and the server device and the plurality of clients are connected to each other via a network, and peer-to-peer A data file delivery system capable of delivering a data file by a method, wherein the server device acquires information on a delivery status of a data file for each client, and sends information to each client based on the delivery status. Delivery management means for predicting when delivery is completed is provided (claim 1).

  According to this configuration, the distribution status of the data file to each client can be managed by the distribution management means, and the completion time of distribution to each client can be predicted. Therefore, it becomes possible to manage the progress of data file distribution, to grasp the client whose distribution is delayed, and to take measures for the client.

  In the above configuration, it is desirable that the distribution management unit treats an arbitrary client of the plurality of clients as a seeder that performs a server function, and causes the other client to distribute a data file from the seeder. ).

  According to this configuration, data files can be distributed by hierarchizing clients. Therefore, it is possible to construct a system that can solve the bottleneck problem of the server device and perform efficient data file distribution.

  In any one of the configurations described above, the server device has the data file, and a first server capable of delivering the data file to a client determined by a predetermined condition among the plurality of clients; A second server that holds the data file and can directly deliver the data file to a specific client in file units, and the delivery management means delays delivery of the data file from a predetermined time When the predicted client is specified, the data file can be directly distributed from the second server to the specified client.

  According to this configuration, when a client whose data file delivery is predicted to be delayed from a predetermined time is specified, the data file is directly delivered from the second server to the delayed client in file units. For this reason, even if a delivery delay occurs in some clients in data file delivery by the P2P method between clients, the delay can be compensated by direct delivery from the second server. Therefore, the data file can be reliably delivered within the time limit.

  In any one of the configurations described above, the server device has the data file, and a first server capable of delivering the data file to a client determined by a predetermined condition among the plurality of clients; A third server that holds the data file and can directly distribute one data file to a specific client in units of subdivided small files, and the distribution management means includes: When a client whose distribution is predicted to be delayed from a predetermined time is specified, the data file can be directly distributed from the third server to the specified client. ).

  According to this configuration, when a client whose data file delivery is predicted to be delayed from a predetermined time is specified, the data file is directly delivered from the third server to the delayed client in units of small files. . For this reason, even if a delivery delay occurs in some clients in data file delivery by the P2P method between clients, the delay can be compensated by direct delivery from the third server. Therefore, the data file can be reliably delivered within the time limit.

  In this case, when the distribution to the specified client is completed for some small files among all the small files constituting one data file, the specified from the third server. It is desirable that the remaining small files that have not been distributed are distributed to the client.

  According to this configuration, only the remaining data files that have not been transmitted are distributed to the specified client without discarding the already transmitted partial data files, thereby avoiding redundant distribution waste. Can do.

  In the above-described configuration, the server device has the data file and can distribute the data file to a client determined by a predetermined condition among the plurality of clients, and the data file A second server that can directly deliver the data file to a specific client in units of files, and the data file that is divided into one data file for the specific client. And a third server capable of direct distribution in units of small files, wherein the distribution management means specifies data for the client when the client is predicted to be delayed in data file distribution from a predetermined time. Confirming the delivery completion level of the file, the delivery completion level being a first value below a predetermined threshold The distribution management means causes the data file to be directly distributed from the second server to the specified client in file units, and the distribution completion level is a second value exceeding a predetermined threshold value. The distribution management means can be configured to directly distribute the data file from the third server to the specified client in units of small files (Claim 6).

  In general, the file delivery speed is faster when file-by-file is distributed than when a single file is distributed in small file units. According to the above configuration, when a client whose data file delivery is predicted to be delayed from a predetermined time is specified, the second server or the third server is selected according to the degree of completion of delivery to the client. In addition, the data file is delivered directly to the delivery delay client. Therefore, file distribution efficiency can be further improved.

  Here, in the case where the data file to be distributed includes a first data file and a second data file to be distributed following the first file, the degree of distribution completion of the first data file is the first data file. The distribution management means causes the first server and the second data file to be directly distributed from the second server to the specified client in units of files, and the distribution of the first data file is completed. When the degree is the second value, the distribution management means distributes the first data file directly from the third server to the identified client in units of small files, and then Preferably, the second data file is directly distributed from the second server in file units.

  According to this configuration, when the first and second data files are continuously distributed, when the first data file is in the middle of distribution to the distribution delay client, the second server is selected according to the degree of distribution completion. Alternatively, distribution from the third server is selected. On the other hand, in the subsequent distribution of the second data file, the distribution from the second server having a high file distribution speed is selected. Therefore, file distribution efficiency is further enhanced.

  In any one of the above configurations, the distribution management unit has a function of disconnecting the server device from the network and returning to the network after the disconnection, and the distribution management unit is configured to form a data distribution environment between clients Accordingly, it is desirable to control the detachment of the server device from the network or the return to the network (claim 8).

  According to this configuration, it is possible to reduce the load on the server device by detaching the server device from the network. Therefore, it is possible to make the server device bear another load, and increase the usability of the server device. be able to.

  In this case, the delivery management means treats an arbitrary client of the plurality of clients as a seeder that performs a server function, and causes the seeder to deliver a data file to the other client. The server apparatus is detached from the network when the number of the server is greater than the predetermined number, and then the server apparatus is returned to the network when the number of the seeders is smaller than the predetermined number. (Claim 9).

  According to this configuration, the server device is disconnected from the network at a stage where a sufficient number of seeders to allow other clients to perform data file distribution is secured, while the active seeder is affected by an excessive load or power off. In such a case, the server device can be returned to the network to perform a part of the data file distribution operation so that it can be operated in a very practical manner.

  The server device includes a single server that holds the data file, and the single server distributes the data file to a client that is determined by a predetermined condition among the plurality of clients. A first functional unit that performs direct distribution of the data file to a specific client in units of files and / or a single data file that is subdivided into a small file unit A third function unit that directly distributes at a time, and the distribution management unit disconnects the first function unit from the network at the time of leaving, while the second function unit and / or the third function unit It is desirable to maintain a state connected to the network (claim 10).

  According to this configuration, the first function that distributes the data file to a specific client, the second function unit that rescues and distributes to a specific client whose distribution is delayed, and / or the first server. Three functions are realized. At this time, since the first function is disconnected from the network in a timely manner, it is possible to reduce the load on the server when performing rescue delivery.

  In any one of the above-described configurations, it is desirable to further include a history information database that stores distribution result information of data files for each client.

  According to this configuration, since the distribution result information of the data file for each client is recorded, it is possible to accurately predict when the distribution to each client is completed by referring to such information.

  According to the data file distribution system of the present invention, it is possible to manage the progress of data file distribution, and to grasp the client whose distribution is delayed. Therefore, it is possible to take measures such as directly distributing the data file from the server for rescue, for example, to the delivery delay client. As a result, in a system in which data file distribution by the P2P method is enabled, data file distribution can be reliably performed on time.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a diagram schematically showing a hardware configuration of a data file distribution system S according to the embodiment of the present invention. The data file distribution system S includes a server device 1 that is a data file distribution source and a plurality of clients (seeder clients 21A to 21C, clients 22A to 22C). The server apparatus 1 and a plurality of clients are connected to each other via an Internet line, and can perform data file distribution by the P2P method.

  Here, data files distributed over the network are, for example, color moving image video files compressed by the MPEG (moving picture experts group) method, audio files compressed by the MP3 (MPEG audio layer 3) method, JPEG (joint photographic experts group). ) Method for various content files such as color still image files compressed. The server device 1 is a server device held by a content provider whose business is to distribute such content files to users, for example. The client is a terminal device such as a personal computer held by the user.

  The server device 1 includes a first server (distribution source server 11), a second server (FTP rescue server 12), and a third server (P2P rescue server 13). The client includes seeder clients 21A to 21C and clients 22A to 22C which are used so as to substantially function as a server. The seeder clients 21A to 21C are not fixed, and are appropriately selected from the client group by the distribution management unit 112 of the server device 1 described later. This selection can be performed by, for example, predetermining the number of seeder clients and treating them as seeders in order from the client whose data file has been distributed. As described above, each client is an equivalent terminal device that functions as a peer in P2P communication. However, in order to improve distribution efficiency, each client is flexibly layered.

  The distribution source server 11 is a server that holds the data file and distributes the data file to clients (as a result, the seeder clients 21A to 21C). An FTP (file transfer protocol) rescue server 12 is a server that shares a data file of the distribution source server 11 and directly distributes the data file in units of files to a specific client in which file distribution is delayed. . Similarly, the P2P rescue server 13 shares the data file held by the distribution source server 11 and subdivides one data file into subdivided small files (subpieces) for a specific client whose file distribution is delayed. It is a server that delivers directly in units.

  The general operation of the data file distribution system S configured as described above is as follows. First, under arbitrary conditions, the data file is distributed from the distribution source server 11 to the client, and the clients whose distribution is completed are treated as the seeder clients 21A to 21C. Data files are distributed from the seeder clients 21A to 21C to the clients 22A to 22C by the P2P method. Further, the data files are distributed by the P2P method between the clients 22A to 22C. Thereby, the bottleneck problem of the distribution source server 11 can be solved, and the data file can be distributed efficiently.

  However, when a data file is distributed between the clients 22A to 22C by the P2P method, the file distribution may not proceed as planned to some clients. In this case, when the P2P communication between the clients 22A to 22C is entrusted, when it is necessary to complete the distribution of data files to all clients (clients that have applied for downloading of data files) by a certain time limit, the time limit It may happen that the delivery of the data file cannot be completed by

  Therefore, in the data file distribution system S of the present embodiment, distribution management means (FIG. 2) obtains information on the distribution status of the data file for each client and predicts the completion of distribution to each client based on the distribution status. A distribution management unit 112). The data file is directly rescued and delivered from the FTP rescue server 12 and / or the P2P rescue server 13 to a client whose delivery delay is predicted. The distribution management means and the repair distribution function will be described in detail below.

  FIG. 2 is a block diagram illustrating a functional configuration of the data file distribution system S, particularly a functional configuration of the distribution source server 11. The distribution source server 11 includes a content database (DB) 111 and a distribution management unit 112.

  The content database 111 is a database in which data files such as the above-described video file, audio file, and image file are stored. The data file stored in the content database 111 is appropriately distributed according to the download application from the client.

  The distribution management unit 112 is installed so that the data file of the content database 111 is distributed to each client without delay. The distribution management unit 112 includes a CPU (Central Processing Unit), a ROM (Read Only Memory) that stores a control program, a RAM (Random Access Memory) that is used as a work area of the CPU, and the CPU executes the control program. By doing so, the distribution control unit 31, the management information communication unit 32, the distribution management table 33, the distribution history storage unit 34, the distribution prediction unit 35, the rescue server control unit 36, and the network connection control unit 37 function.

  When distributing the data file, the distribution control unit 31 refers to center master information set in a distribution management table 33 described later, and determines the number of seeder clients according to the center master information and the number of clients to be distributed. To do. Then, a process of distributing the data file stored in the content database 111 via the Internet IN is performed to the client that has made the distribution request.

  Here, a data file transfer method will be described. FIG. 3 is a diagram schematically showing a sub-piece transfer method generally employed for transferring video files and audio files. The bit torrent method that is widely used when transferring video files and audio files by the P2P method also employs such a transfer method. As shown in FIG. 3, in the subpiece transfer method, one data file is subdivided into a number of subpiece files, and file transfer is performed in units of subpieces.

  4 shows files from the distribution source server 11 to the client (seeder clients 21A and 21B or clients 22A to 22C shown in FIG. 2), between the seeder clients 21A and 21B and the clients 22A to 22C, and in the clients 22A to 22C. It is a figure which shows the system of transfer typically. As shown in FIG. 4, in the distribution source server 11 and the like, the above-described sub-piece transfer method is adopted in order to distribute files by the P2P method. That is, when file 1, file 2, and file 3 to be distributed exist, file transfer is performed in units of subpieces in order from file 1.

  FIG. 5 is a diagram schematically showing a transfer method when a data file is remedied and delivered from the FTP rescue server 12 to the delivery delay client. The FTP rescue server 12 performs file transfer to the delivery delay client in units of file 1, file 2, or file 3. FIG. 5 shows an example in which the file 2 is repaired and distributed in units of files. This transfer method has an advantage of high data transfer efficiency because transfer is performed in file units. On the other hand, since transfer can only be performed in units of files, for example, even if a part of the file 2 has already been transferred by the P2P method, the partially transferred file 2 is discarded, and the entire file 2 is newly rewritten. Need to transfer again.

  FIG. 6 is a diagram schematically showing a transfer method when the data file is remedied and delivered from the P2P rescue server 13 to the delivery delay client. The P2P rescue server 13 performs file transfer to the delivery delay client in units of sub-pieces of one file. FIG. 6 shows an example in which a part of the file 1 is repaired and delivered in units of subpieces. In this transfer method, transfer is performed in units of sub-pieces. Therefore, for example, when a part of transfer has already been performed for the file 1 by the P2P method, the file 1 that has already been transferred is not discarded. 1 has an advantage that only untransferred sub-pieces can be distributed by the P2P method.

  Returning to FIG. 2, the management information communication unit 32 performs data communication with each client, and acquires various client information including information related to the distribution status of the data file from each client. For example, when a client is connected to the server apparatus 1, the management information communication unit 32 acquires information such as the address of the client and a file name desired to be distributed, and creates a management table for each client in the distribution management table 33. At the same time, processing is performed for acquiring data relating to the data file distribution progress status and sequentially writing it into the management table.

  The distribution management table 33 is a functional unit that manages the center master information set for the distribution of the data file and the client information acquired by the data communication between the management information communication unit 32 and each client. FIG. 7 is a table format showing an example of a table managed in the distribution management table 33.

  FIG. 7A is an example of a table of center master information. In this example, seeder check timing, delivery start date / time, delivery check date / time, delivery deadline date / time, minimum seeder ratio, and rescue mode switching% are listed as management items.

  The seeder check timing is an interval setting for checking whether or not the seeder client is operating normally. This seeder check is performed to check whether or not a required number of seeder clients exist on the network. In order to perform file transfer quickly and efficiently, it is desirable that a certain number of seeders or more are always operating with respect to the total number of clients that perform distribution. However, it is possible that a client having a seeder qualification (a client for which file transfer has been completed) is turned off, or an overload does not substantially fulfill the role of a seeder. Therefore, a seeder check is performed in order to constantly grasp the number of seeders. The management information communication unit 32 communicates with the seeder client at the seeder check timing set here, and performs a seeder check.

  The distribution start date and time is a setting of the date and time when distribution of one data file (group) to be distributed is actually started. The distribution control unit 31 starts file transfer from the set date and time. The delivery deadline date and time is a date and time setting at which delivery of one data file (group) must be completed. The delivery check date and time is a date and time setting for confirming the delivery status of the data file to each client. The management information communication unit 32 acquires data file distribution progress data from each client at the distribution check date and time.

  The minimum seeder ratio is a setting of the number of seeders necessary for performing efficient data file distribution with respect to the total number of clients performing distribution. Here, the percentage is illustrated, but for example, the number of seeders may be set to 1 if the total number of clients is 50 or less, and the number of seeders may be set to 2 if 50 to 100. In this embodiment, when the minimum seeder ratio is exceeded, the distribution source server 11 is disconnected from the network, and when the minimum seeder ratio is exceeded, the distribution source server 11 is returned to the network and resumes data file distribution.

  Rescue mode switching% is a setting for a threshold value for determining which of the FTP rescue server 12 and the P2P rescue server 13 is to be used when performing rescue delivery to a delivery delay client. For example, when the rescue mode switching% is set to 50%, if the distribution completion degree is 50% or more for one data file, the rescue delivery is performed from the P2P rescue server 13, and the distribution completion degree is less than 50%. If so, distribution from the FTP rescue server 12 is performed.

  FIG. 7B is an example of a client information management table. In this example, the distribution completion rate is managed for each distribution file in addition to the nickname (unique management number assigned to each client), the distribution (DL) mode, and the status. Also, a delivery start date and time and a delivery end date and time are recorded. These data are utilized when recording the distribution history.

  The distribution history storage unit 34 (history information database) stores data file distribution history information for each client. FIG. 8 is a table format showing an example of a table storing distribution history information. Here, an example is shown in which the reception capacity, the average reception speed, the reception speed order, the transmission capacity, the average transmission speed, and the transmission speed order are recorded for each client for the previous and previous delivery results.

  The distribution predicting unit 35 performs a process of predicting the time when the distribution to each client is completed based on the data regarding the distribution status acquired from each client. At this time, the distribution prediction unit 35 refers to the distribution history information stored in the distribution history storage unit 34. By referring to such distribution history information, it is possible to accurately predict the completion of distribution to each client.

  Specifically, the distribution predicting unit 35 obtains an average transfer capacity for each client from the degree of distribution completion from the distribution start date and time to the distribution check date and time (see FIG. 7A), and the distribution completion time is calculated from this average transfer capacity. Predict. In actual processing, a capacity (distributable capacity) that can be distributed by normal P2P communication is obtained based on the average transfer capacity and a period from the delivery check date to the delivery deadline date. Then, this distributable capacity is compared with the data capacity (remaining transfer capacity) that has not been distributed yet at the time of the distribution check date, and it is determined whether or not the distribution can be completed by the distribution deadline date.

  The rescue server control unit 36, when the distribution prediction unit 35 identifies a client (distribution delay client) that is predicted to be unable to complete the distribution of the data file by the distribution deadline date and time, the FTP rescue server 12 and / or the P2P rescue The server 13 is controlled, and the undelivered data file is directly distributed (relief distribution) to the distribution delay client.

  The rescue server control unit 36 performs FTP rescue mode in which rescue delivery is performed using the FTP rescue server 12, P2P rescue mode in which rescue delivery is performed using the P2P rescue server 13, and the FTP rescue server 12 and P2P depending on the situation. One of the composite rescue modes in which rescue delivery is performed using both of the rescue servers 13 is selected. When applying the composite rescue mode, the rescue server control unit 36 refers to the rescue mode switching% in the distribution management table 33 and switches between them.

  The network connection control unit 37 controls detachment of the distribution source server 11 from the network and return to the network after the detachment according to the state of formation of the data distribution environment between the clients. Specifically, the network connection control unit 37 refers to the minimum seeder ratio in the distribution management table 33, and if the minimum seeder ratio is exceeded, the distribution source server 11 is detached from the network and is below the minimum seeder ratio. The distribution source server 11 is returned to the network.

  By performing such control, the distribution source server 11 is disconnected from the network when a sufficient number of seeders is ensured, so that there is no problem in the distribution of the data file, while the load on the distribution source server 11 is reduced. This makes it easy to use the distribution source server 11 for other purposes. On the other hand, when the number of working seeders decreases, the distribution source server 11 is returned to the network, so that the distribution of data files can be prevented from being delayed.

  In addition, a single server device is used as the server device 1, a part that functions as the distribution source server 11 (first function unit), a part that functions as the FTP rescue server 12 (second function unit), and P2P When operating by dividing into a part (third function unit) that functions as the rescue server 13, the first function unit is disconnected from the network, while the second function unit and / or the third function unit is connected to the network. By maintaining the state, the load on the single server device can be reduced, and the function as a rescue server can be executed at high speed.

  Next, the operation of the data file distribution system S described above will be described according to a flowchart. FIG. 9 is a flowchart showing an overall schematic operation of data file distribution processing by the data file distribution system S (distribution management unit 112) according to the present embodiment. When distributing the data file in response to a download application from the client, the distribution management unit 112 first performs various settings necessary for the distribution (step S1). Thereafter, control is performed to distribute the data file to the client (step S2).

  FIG. 10 is a flowchart showing details of the distribution setting process in step S1 shown in FIG. First, registration of a data file distributed in response to a request is accepted in the content database 111 (step S11). If the data file has already been stored in the content database 111, distribution selection is accepted.

  Next, the setting of each item is received as center master information in the distribution management table 33 (step S12). The setting items here are, for example, a seeder check timing, a delivery start date and time, a delivery check date and time, a delivery deadline date and time, a minimum seeder ratio, and a rescue mode switching% as described above (see FIG. 7A).

  Subsequently, the distribution control unit 31 receives an input of the total number of clients that should distribute the data file (step S13). With reference to the total number of clients and the center master information, the distribution control unit 31 determines the minimum number of seeder clients necessary for the current distribution (step S14).

  FIG. 11 is a flowchart showing an outline of the distribution control process in step S2 shown in FIG. When the above-described distribution setting is completed and the set distribution start date and time arrives, the distribution control unit 31 transfers the data file stored in the content database 111 to the network (Internet IN ) Is started via step S21) (step S21).

  After that, when the delivery check date and time arrives, the delivery delay client identification based on the client information (see FIG. 7B) sequentially acquired by the management information communication unit 32 and the relief processing based on the relief delivery are executed. (Step S22). In parallel with this, the network connection control unit 37 executes server detachment return processing for controlling the detachment and return of the distribution source server 11 from the network (step S23).

  FIG. 12 is a flowchart showing details of the server leave return process in step S23 shown in FIG. The network connection control unit 37 refers to the minimum seeder ratio in the distribution management table 33 and determines whether or not the number of seeders has reached a predetermined number (step S31). The seeder count depends on the number of clients whose distribution completion rate is 100% in the client information acquired by the management information communication unit 32.

  If the number of seeders has not reached the predetermined number (NO in step S31), the network connection control unit 37 maintains the state where the distribution source server 11 is connected to the network. On the other hand, when the number of seeders reaches a predetermined number (YES in step S31), the network connection control unit 37 leaves the distribution source server 11 from the network and stops the distribution of the data file from the distribution source server 11 (step S31). S32).

  Thereafter, the network connection control unit 37 waits for the arrival of the seeder check timing set as one of the center master information (step S33), and when the seeder check timing arrives (YES in step S33), the number of seeders is predetermined. It is determined whether or not the number is maintained (step S34).

  When the predetermined number of seeders exists on the network (YES in step S34), the process returns to step S33 and is repeated. On the other hand, when the predetermined number of seeders does not exist on the network (NO in step S34), the network connection control unit 37 returns the distribution source server 11 to the network, and from the distribution source server 11 The distribution of the data file is resumed (step S35). Then, it returns to step S31 and a process is repeated.

Next, details of the relief process in step S22 shown in FIG. 11 will be described. In the present embodiment, relief delivery in the following three modes is possible.
[1] FTP rescue mode: use FTP rescue server 12 [2] P2P rescue mode: use P2P rescue server 13 [3] composite rescue mode: use both FTP rescue server 12 and P2P rescue server 13 Details of each of the modes will be described sequentially.

  FIG. 13 is a flowchart showing details of the rescue processing in the FTP rescue mode. The delivery prediction unit 35 waits for the delivery check date and time set as one of the center master information (step S41). When the delivery check date / time arrives (YES in step S41), the delivery prediction unit 35 refers to the client information in the delivery management table 33 to check the delivery progress for each client, and from the delivery check date to the delivery deadline date / time. It is confirmed whether or not the distributable capacity in the period exceeds the remaining transfer capacity (step S42).

  If the distribution capacity for a specific client is less than the remaining transfer capacity (NO in step S42), it is highly likely that the data file cannot be distributed to the client by the distribution deadline date, so relief distribution is required. . In this case, the DL mode for the delivery delay client is changed from the normal “P2P mode” to the “FTP rescue mode” (step S43), and the FTP rescue server 12 is controlled by the rescue server control unit 36. The undelivered data file is directly distributed from the FTP rescue server 12 (step S44).

  FIG. 14 is a diagram schematically showing a mode of rescue delivery in the FTP rescue mode. Assume that file 1, file 2, or file 3 exists as a data file to be distributed, and that only transfer of file 1 has been completed for a specific client at the distribution check date. When it is determined that the transfer of the files 2 and 3 is not in time for the delivery deadline based on the past transmission results in the normal P2P delivery route, the files 2 and 3 are transferred from the FTP rescue server 12. Relief delivery is performed in units of files to a delivery delay client.

  If it is determined in step S42 that the capacity that can be distributed is smaller than the remaining transfer capacity, the terminal of the system administrator or the client is notified by e-mail or the like that the distribution is delayed. May be.

  If the capacity that can be distributed for a specific client is greater than the remaining transfer capacity (YES in step S42), no special distribution is performed for that client. Thereafter, whether or not the distribution of the data file to the client is 100% complete is confirmed based on the client information (step S45). If the distribution is 100% complete (YES in step S45), an electronic mail notifying the completion of distribution is transmitted from the distribution management unit 112 to the terminal and client of the system administrator (step S46). Thereafter, the server disconnection return processing function by the network connection control unit 37 is stopped (step S47), and the distribution processing is completed.

  FIG. 15 is a flowchart showing details of the relief process in the P2P rescue mode. The distribution predicting unit 35 waits for the distribution check date and time (step S51), similarly to step S41 and step S42 described above with reference to the flowchart of FIG. 13, and whether or not the relationship of distributable capacity> remaining transfer capacity is satisfied at that time. (Step S52).

  If the distribution capacity for a specific client is less than the remaining transfer capacity (NO in step S52), since relief distribution is required, the DL mode for the distribution delay client is changed from the normal “P2P mode” to “P2P rescue mode”. (Step S53). Then, the rescue server control unit 36 controls the P2P rescue server 13, and the undelivered data file is directly delivered from the P2P rescue server 13 to the delivery delay client (step S54).

  FIG. 16 is a diagram schematically illustrating a relief delivery mode in the P2P rescue mode. Assume that file 1, file 2, or file 3 exists as a data file to be distributed, and that only a halfway transfer of file 1 has been completed for a specific client at the distribution check date and time. If it is determined that the delivery deadline will not be met, the delivery delay client will be delivered to the well-separated sub-piece through normal P2P communication, and the remaining sub-piece of file 1 will be sent to the P2P rescue server. 13 is distributed directly in a form that monopolizes the band. Thereafter, the file 2 and the file 3 are sequentially delivered from the P2P rescue server 13 in subpiece units.

  If the capacity that can be distributed for a specific client is greater than the remaining transfer capacity (YES in step S52), no particular relief distribution is executed for that client. Thereafter, it is confirmed whether or not the distribution of the data file is 100% complete (step S55). If the distribution is 100% complete (YES in step S55), the distribution is completed to the terminal and client of the system administrator. Is sent from the distribution management unit 112 (step S56). Thereafter, the server disconnection return processing function by the network connection control unit 37 is stopped (step S57), and the distribution processing is completed.

  FIG. 17 is a flowchart showing details of the rescue processing in the composite rescue mode. The delivery predicting unit 35 waits for the delivery check date and time (step S61), similarly to step S41 and step S42 described above with reference to the flowchart of FIG. 13, and whether or not the relationship of deliverable capacity> remaining transfer capacity is satisfied at that time. (Step S62).

  If the distribution capacity is less than the remaining transfer capacity for a specific client (NO in step S62), rescue distribution is necessary. In order to select the rescue mode, the rescue server control unit 36 Based on the client information, it is confirmed whether or not the distribution completion degree of the file being transferred exceeds a predetermined threshold (rescue mode switching%) (step S63). As described above, this threshold is, for example, 50%.

  If the distribution completion level is below a predetermined threshold (NO in step S63), it is more time-consuming if the distribution completion level is low, the file currently being transferred is discarded, and the rescue distribution is newly performed by the FTP method with a high distribution speed. Is advantageous. Therefore, in such a case, the rescue server control unit 36 changes the DL mode for the delivery delay client from the normal “P2P mode” to the “FTP rescue mode” (step S64), and controls the FTP rescue server 12 to The data file corresponding to the file currently being transferred is directly distributed to the distribution delay client (step S68). In this case, the file distributed halfway in the normal “P2P mode” is discarded.

  On the other hand, if the degree of distribution completion exceeds a predetermined threshold (YES in step S63), it can be said that it is wasteful in time to discard the currently transferred file and newly deliver the file. Therefore, in this case, the rescue server control unit 36 changes the DL mode for the delivery delay client from the normal “P2P mode” to the “P2P rescue mode” (step S65), and controls the P2P rescue server 13 to deliver the delivery delay. The undelivered sub-piece data file among the files currently being transferred is directly distributed to the client (step S66).

  Thereafter, it is confirmed whether or not the distribution is completed for the file being transferred (step S67). When the distribution is completed (YES in step S67), the rescue server control unit 36 changes the rescue mode to “FTP rescue mode” (step S64). This is because it is advantageous in time to select the “FTP rescue mode” with high transfer efficiency for the remaining files. The remaining files are delivered directly from the FTP rescue server 12 (step S68). If there are no remaining files, steps S67 to S68 are skipped.

  FIG. 18 is a diagram schematically showing a mode of rescue delivery in the composite rescue mode. Assume that file 1, file 2, or file 3 exists as a data file to be distributed, and that only a halfway transfer of file 1 has been completed for a specific client at the distribution check date and time. If the distribution completion degree of the file 1 exceeds a threshold value (for example, 50%), as shown in the figure, after the normal P2P communication is performed up to the sub-piece with good separation, the remaining of the file 1 The sub-piece is directly distributed from the P2P rescue server 13 in a form that monopolizes the band. Thereafter, the file 2 and the file 3 are relieved and distributed by the FTP rescue server 12 with high distribution efficiency. When the distribution completion degree of the file 1 is below the threshold value, all of the file 1, the file 2 or the file 3 is rescued and distributed by the FTP rescue server 12.

  When the capacity that can be distributed for a specific client is larger than the remaining transfer capacity (YES in step S62), no special distribution is performed for that client. Thereafter, it is confirmed whether or not the distribution of the data file is 100% complete (step S69). If the distribution is 100% complete (YES in step S69), the distribution is completed to the terminal and client of the system administrator. Is sent from the distribution management unit 112 (step S70). Thereafter, the server disconnection return processing function by the network connection control unit 37 is stopped (step S71), and the distribution processing is completed.

  As described above, in order to express the present invention, the present invention has been described appropriately and sufficiently through the embodiments with reference to the drawings. However, those skilled in the art can easily change and / or improve the above-described embodiments. It should be recognized that this can be done. Accordingly, unless the modifications or improvements implemented by those skilled in the art are at a level that departs from the scope of the claims recited in the claims, the modifications or improvements are not limited to the scope of the claims. To be construed as inclusive.

It is a figure which shows typically the hardware constitutions of the data file delivery system S which concerns on embodiment of this invention. 2 is a block diagram showing a functional configuration of a data file distribution system S. FIG. It is a figure which shows a subpiece transfer system typically. It is a figure which shows typically the system of the file transfer from a delivery origin server etc. to a client. It is a figure which shows typically the transfer system in the case of carrying out relief delivery of the data file from an FTP rescue server to a delivery delay client. It is a figure which shows typically the transfer system in the case of carrying out relief delivery of a data file from a P2P rescue server to a delivery delay client. It is a table format figure which shows an example of the table managed in a delivery management table. It is a figure of a table format which shows an example of the table in which delivery history information was stored. 5 is a flowchart showing an overall schematic operation of data file distribution processing by the data file distribution system S. It is a flowchart which shows the detail of the delivery setting process of step S1 shown in FIG. It is a flowchart which shows the outline of the delivery control process of step S2 shown in FIG. It is a flowchart which shows the detail of the server leaving recovery process of step S23 shown in FIG. It is a flowchart which shows the detail of the relief process by FTP rescue mode. It is a figure which shows typically the aspect of the relief delivery by FTP rescue mode. It is a flowchart which shows the detail of the relief process by P2P rescue mode. It is a figure which shows typically the aspect of the relief delivery by P2P rescue mode. It is a flowchart which shows the detail of the relief process by composite rescue mode. It is a figure which shows typically the aspect of the relief delivery by composite rescue mode.

Explanation of symbols

1 Server device 11 Distribution source server (first server)
12 FTP rescue server (second server)
13 P2P rescue server (third server)
21A to 21C Seeder client 22A to 22C Client 31 Distribution control unit 32 Management information communication unit 33 Distribution management table 34 Distribution history storage unit 35 Distribution prediction unit 36 Rescue server control unit 37 Network connection control unit 111 Content database 112 Distribution management unit S Data file distribution system

Claims (11)

A data file distribution system including a server device that is a data file distribution source and a plurality of clients, wherein the server device and the plurality of clients are connected to each other via a network and can perform data file distribution by a peer-to-peer method. And
The server apparatus includes a distribution management unit that acquires information on a distribution status of a data file to each client and predicts a distribution completion time to each client based on the distribution status.   2. The distribution management unit according to claim 1, wherein any one of the plurality of clients is treated as a seeder that performs a server function, and the data file is distributed from the seeder to another client. Data file delivery system. The server device is
A first server having the data file and capable of delivering the data file to a client determined by a predetermined condition among the plurality of clients;
A second server that holds the data file and that can directly deliver the data file in file units to a specific client;
The distribution management unit causes the second server to directly distribute the data file to the specified client when a client whose data file distribution is predicted to be delayed from a predetermined time is specified. The data file distribution system according to claim 1 or 2, wherein The server device is
A first server having the data file and capable of delivering the data file to a client determined by a predetermined condition among the plurality of clients;
A third server that holds the data file and can directly deliver one data file to a specific client in units of subdivided small files;
The distribution management means causes the third server to directly distribute the data file to the specified client when a client whose data file distribution is predicted to be delayed from a predetermined time is specified. The data file distribution system according to claim 1 or 2, wherein In the case where the delivery to the specified client has been completed for some of the small files that constitute one data file,
5. The data file distribution system according to claim 4, wherein the remaining small files that have not been distributed are distributed from the third server to the specified client. The server device is
A first server having the data file and capable of delivering the data file to a client determined by a predetermined condition among the plurality of clients;
A second server that holds the data file and can directly deliver the data file in file units to a specific client;
A third server that holds the data file and can directly deliver one data file to a specific client in units of subdivided small files;
When the delivery management unit identifies a client that is expected to deliver the data file later than a predetermined time, the delivery management unit checks the degree of delivery of the data file to the client,
When the distribution completion degree is a first value lower than a predetermined threshold, the distribution management means directly distributes the data file in file units from the second server to the specified client,
When the distribution completion degree is a second value that exceeds a predetermined threshold, the distribution management unit directly distributes the data file from the third server to the specified client in units of small files. The data file distribution system according to claim 1 or 2, characterized in that In the case where the data file to be distributed includes a first data file and a second data file to be distributed following the first file,
When the distribution completion degree of the first data file is the first value, the distribution management means files the first and second data files from the second server to the specified client. Delivered directly in units,
When the distribution completion degree of the first data file is the second value, the distribution management means transfers the first data file from the third server to the specified client in units of small files. 7. The data file distribution system according to claim 6, wherein after the direct distribution, the second data file is directly distributed from the second server in file units. The distribution management means includes a function of detaching the server device from the network and returning to the network after the detachment,
8. The distribution management unit according to claim 1, wherein the server device controls the detachment of the server device from the network or the return to the network according to a state of formation of a data distribution environment between clients. The data file distribution system described in 1. The distribution management means treats an arbitrary client of the plurality of clients as a seeder that performs a server function, and causes the other client to distribute a data file from the seeder,
The server device is detached from the network when the number of seeders is greater than a predetermined number, and then the server device is removed from the network when the number of seeders is less than a predetermined number. The data file delivery system according to claim 8, wherein The server device includes a single server that holds the data file, and the single server distributes the data file to a client that is determined by a predetermined condition among the plurality of clients. 1 functional part,
A second function unit that directly distributes the data file to a specific client in units of files and / or a third function that directly distributes one data file to a specific client in units of subdivided small files And comprising
The distribution management means maintains the state in which the second function unit and / or the third function unit are connected to the network while the first function unit is disconnected from the network at the time of leaving. The data file distribution system according to claim 8.   The data file delivery system according to any one of claims 1 to 10, further comprising a history information database in which data file delivery record information for each client is stored.

Images