JPH11341044A - File transfer protocol system, its method and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide the constitution of a protocol capable of realizing stable file transfer by using an unstable communication environment such as mobile object communication. SOLUTION: Each of devices 1, 2 for executing file transfer is constituted of a control part 11 for controlling a program, a nonvolatile procedure storing area 14 including procedure states consisting of job ID, a file name, a file attribute, the number of transmitting blocks, the number of receiving blocks, etc., and a non-volatile file 15 for storing data. The control part 11 in the device 1 judges whether file transfer (block transfer) is being executed or not by reading out the contents (procedure states) from the area 14 at the time of starting the device 1, and when file transfer is being executed, restarts file transfer (block transfer) by informing of restart processing. In addition, the control part 11 informs the opposite side of the start of file transfer by a file transfer command from a display operation part 18 to execute file transfer.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a file transfer system, a method, and a recording medium, and more particularly, to a file transfer protocol system that requires high reliability in mobile communication.

[0002]

2. Description of the Related Art As an automatic retransmission method for file transfer in the prior art, for example, Japanese Patent Laid-Open No. 7-46289 is disclosed. Although such an automatic retransmission scheme is intended for recovery from a temporary and minor communication failure such as a bit error, the network architecture [for example, ISO
Layer 1 (physical layer) of the OSI (Open System Interconnection Reference Model) recognizes that a fatal communication failure has occurred and immediately stops the file transfer protocol.

[0003] This concept is based on a fixed wired network. In the case of direct use of a dedicated line, layer 1 should always be able to communicate and line disconnection is a fatal obstacle. Was intended to start over.
When a layer 1 failure occurs in the circuit switching network, the network releases the line, so the intention was to stop the file transfer and start over from the beginning.

[0004] As a function of the network side to alleviate the line disconnection state, there is a case in which even if a wireless section is disconnected in a mobile communication network, the line end is not immediately set to a disconnected state. , The line is disconnected.

[0005] As a measure to alleviate the line disconnection on the conventional data communication protocol side, the concept of line disconnection is eliminated by using a connectionless datagram protocol in a lower layer of file transfer, and then the delivery confirmation of the file transfer protocol is confirmed. There are some that can withstand timeouts.

For example, Japanese Patent Application Laid-Open No. H06-243059 discloses a method for recovering from interruption of the file transfer protocol itself. This method is a method specialized for OSI-FTAM. However, since this interruption recovery processing method is realized by explicitly providing interruption and resumption by operator intervention, it cannot cope with a long-term communication failure or node failure.

[0007]

An object of the present invention is to provide a protocol configuration for realizing stable file transfer using an unstable communication environment such as mobile communication.

[0008] The conventional file transfer protocol assuming a wired communication line interprets the interruption of layer 1 as a fatal failure and immediately stops the processing completely.

[0009] However, in mobile communication using wireless communication, layer 1 over a short time to a long time may be caused by various causes.
Frequently occurs. An example due to the radio wave condition may be an effect of passing through a tunnel or the like, and may actually be out of a spatial service area and become disconnected.

[0010] Further, since many mobile communication devices are driven by batteries, they may be interrupted when the electric capacity is insufficient. Alternatively, a combination of both causes may cause a long interruption in file transfer.

The line holding function prepared by the line switching service of the mobile communication network is limited to alleviation of line disconnection in a short time. It is difficult to keep a line for a long time from the viewpoint of accounting processing and resource management.

Therefore, the conventional file transfer protocol designed for the fixed network has insufficient recovery processing for such characteristics of the mobile communication network.

The method of eliminating the concept of line disconnection by using a connectionless datagram protocol on a network architecture requires mounting many lower-order protocol stacks, and has a disadvantage of consuming a large amount of resources.

Further, even on a connectionless network, the conventional file transfer protocol can only withstand a communication failure until a timeout of delivery confirmation. He could not recover in another session.

In Japanese Patent Application Laid-Open No. 06-243059, which is a prior art file transfer interruption recovery process, O
Since it is limited to an implementation method in the category of SI-FTAM, it has no versatility, is heavy in mounting, and cannot be mounted on a portable terminal. Furthermore, since it is compatible only with interruption due to operator intervention, it is ineffective for automatic recovery of interruption due to communication failure and node failure. Specifically, it does not function under a communication failure and a node failure because the communication itself is used for the method of shifting to the suspended state. Furthermore, if the power is interrupted during the interruption, the parameters and the like are erased, and the recovery process cannot be performed.

In these prior arts, interruption of file transfer due to a serious communication failure cannot be automatically recovered, and file transfer has been manually performed again.

An object of the present invention is to automatically recover a file transfer from a catastrophic communication failure without depending on a lower protocol.

Another object of the present invention is to provide a file transfer protocol that can be recovered from a node failure including a power failure by using a reasonable and light mounting method so that it can be mounted on a portable terminal.

[0019]

SUMMARY OF THE INVENTION In order to achieve the above object, the present invention provides a file transfer method for dividing a file into a plurality of blocks when transferring a file between a plurality of devices connected to a communication network. A transfer protocol method, wherein a first device of the plurality of devices is the file transfer source and a second device of the plurality of devices is the file transfer destination. Independently of the distinction, the initiating device sends a start request to the passive device and the passive device responds, so that both devices have a unique job ID before file transfer. A start notification procedure is negotiated, and when the start-up apparatus receives a start response from the passive apparatus, the start apparatus writes its own procedure initial state including a job ID in a nonvolatile storage unit, and the passive apparatus performs the start Required Is received, a start response is returned to the activation side after writing its own procedure initial state to the non-volatile storage unit, and the first device starts transmitting block transfer from the initial state, and the second device returns to the initial state. Starts receiving a block transfer from each time, each time the block transfer is received, the block is stored in a file storage area, and the number of acknowledgments is written in the second nonvolatile storage as a part of the second procedure state, and the block transfer acknowledgment is performed. Each time the first device receives the block transfer acknowledgment from the second device during the block transfer, the acknowledgment number indicating the acknowledgment as part of the first procedure state. 1 is written into the non-volatile storage unit,
When the first device has received all the blocks, the file transfer end notification is transmitted to the second device, and the second device returns a completion acknowledgment,
Device returns an end confirmation notification, erases all the first procedure states from the first nonvolatile storage unit, the second device repeatedly sends an end acknowledgment until receiving an end confirmation notification, Upon receipt of the end confirmation notice, all the second procedure states are deleted from the second nonvolatile storage unit.

Further, according to the present invention, the first apparatus searches the first procedure state stored in the first non-volatile storage unit when the apparatus is started up, and determines whether the first apparatus is in a file transfer state, that is, in a suspended state. For example, a restart request to which the job ID is added is transmitted to the passive unit of the second device. When the second device receives the restart request from the first device, the job ID and the job ID added to the restart request are transmitted. The procedure status is compared with the procedure status in the second nonvolatile storage unit, and if there is no corresponding job, a restart rejection response is returned to the first device. If the corresponding job exists, a restart acknowledgment is sent to the first device. When the first device receives an acknowledgment for the restart request, the first device starts transfer from the next block that has been transferred before starting the device, and when the first device receives a restart rejection response, it receives the first nonvolatile memory. From the sex memory It is characterized by erasing all of the job information.

Further, according to the present invention, the second apparatus searches the second procedure state stored in the second nonvolatile storage unit at the time of starting the apparatus, and determines whether the second procedure is in a file transfer state, that is, in a suspended state. For example, a restart request including the job ID and the received block is transmitted to the passive unit of the first device,
When the first device receives the restart request from the second device, the first device compares the job ID added to the restart request with the procedure state in the first non-volatile storage unit. Return a rejection response to the second device,
If the job is present, a restart acknowledgment is returned to the second device, and the transfer is restarted from the position following the transferred block. If the second device receives an acknowledgment for the restart request, the device is not activated. Starts receiving from the next block that has been transferred to
In which the entire job information is deleted from the non-volatile storage unit.

The device to which the present invention can be applied can be a general computer having a non-volatile storage area. Therefore, a recording medium of a program for realizing the file transfer protocol having the above characteristics can be included in the components. The program recording medium of the file transfer protocol may include only the starting program or only the passive program.

[0023]

Next, embodiments of the present invention will be described with reference to the drawings. FIGS. 3 to 5 show an embodiment in which datagram communication is used in which it is not possible to distinguish between a communication failure and a node failure.

FIG. 2 shows a communication network 3 and an apparatus 1 which is a node connected to the communication network 3, showing an embodiment of the present invention.
Device 2. Note that devices other than the device 1 and the device 2 can be connected to the communication network 3, but are omitted in the embodiment.

Referring to FIG. 2, device 1 or device 2
Is an entity that performs file transfer, and performs program control having a nonvolatile storage area, a human operation display unit, and an input / output device for communication. The communication network 3 is a communication path including a network, and a plurality of devices other than the device 1 or the device 2 can be connected.

FIG. 1 is a block diagram showing the device 1 or 2 in FIG. 2 in detail.

Referring to FIG. 1, an apparatus 1 or 2 provided with a recording medium 21 on which a file transfer protocol program is recorded is provided with a control unit 1 operating under program control.
1, a timer 12 used for monitoring time during transfer, and a nonvolatile storage area 13 connected as an input / output device of the control unit 11.
, A display operation unit 18, a communication unit 17 having an interface with the communication network 3 and communicating with other devices, and a bus 16. This recording medium 21 may be a magnetic disk or another program recording medium.

The control unit 11 includes a central processing unit (not shown) and a memory for storing programs, and operates as a computer under program control. The file transfer protocol program is read from the recording medium 21 into the memory in the control unit 11 and controls the operation of the control unit 11. That is, the control unit 11 executes the program written in the memory in the control unit 11 to execute the file transfer according to the file transfer protocol. In the present embodiment, the file transfer protocol program is loaded from the recording medium 21 to the memory in the control unit 11 and the program is executed, but the program (the ROM, for example) in the control unit 11 is executed in advance. May contain a program.

The display operation section 18 executes a display function (output function) and an input function as means of a man-machine interface. This performs a file transfer start operation and a result display.

The communication unit 17 executes a function of bidirectionally communicating with a party to which a file is to be transferred via the communication network 3.

The non-volatile storage area 13 has a function of continuously storing data even when the power of the apparatus 1 (or the apparatus 2) is turned off, and includes a file 15 used for storing a file;
Status of the control procedure of the present invention (consisting of information indicating transmission source or destination, job ID, file name, file attribute, number of transmission blocks, number of reception blocks, destination address, etc. ) Is stored in the procedure storage area 14. In this case, the recording of the procedure state is classified into the transmission source and the transmission destination for each job ID unit and recorded.

It is to be noted that an apparatus having a file originally requires a medium such as a magnetic disk as a non-volatile storage area as a storage area for the file. Therefore, the embodiment of the present invention is characterized in that the non-volatile storage area 13 indispensable for the file 15 is also used as the procedure storage area 14. Therefore, the present invention is applicable to a general computer, and can implement a protocol by loading a control program with a program medium.

In the conventional file transfer method, the state of the job is only temporarily stored in a volatile storage area (for example, RAM).

Of course, instead of a medium such as a magnetic disk, any nonvolatile memory (eg, flash memory) that can be repeatedly stored may be used.

Next, the overall operation of the present embodiment will be described with reference to FIGS.

The present invention is a procedure developed based on a general procedure used for conventional file transfer.
The part similar to the general procedure is that a file is divided into a plurality of transfer blocks and transmitted. The transferred block is immediately stored in non-volatile storage as part of the file. The transfer source and the transfer destination store up to the transferred and stored block as a check point (a point at which validity is checked in block transfer). The transfer source and the transfer destination ensure the validity of the file transfer by reliably performing this checkpoint. At this time, communication transfer is provided by providing a plurality of blocks (a method of transmitting the next block before replying to block transmission) between the block sent from the transfer source and the block stored at the transfer destination. It is also possible to increase efficiency. In this case, the procedure for managing the checkpoints becomes complicated, but is not related to the essence of the present invention. Therefore, a basic embodiment in which the forward transfer is not performed will be described.

FIG. 3 is a flowchart when the device 1 executes a file transfer protocol program when the transfer source is the device and the device 2 is the transfer destination.

FIG. 4 shows steps B2 and C2 in FIG.
Is a flowchart showing the details of the operation on the activation side or the passive side.

A general device can be a start side or a passive side of a file transfer job, and can be a transfer source or a transfer destination by starting. Basically, only the activated display operation unit 18 is used interactively.

Now, assuming that the device 1 (transfer source) is to be the starting side to transfer a file to the device 2 (transfer destination), first, a file transfer job is started from the display operation unit 18 of the device 1. (Step A1 in FIG. 4), and the process transits to Steps A2, A3, A4, A5, A6, A7, and A8. At this time, the start-up process of the apparatus 1 determines the record data of the new job by starting the transfer (step A1), and sends a request to start the transfer file by the new job to the apparatus 2 by adding job information including the job ID. It transmits and waits for a response (steps A2 to A5 and communication 1). on the other hand,
The device 2 has already been started and is in a request reception waiting state (steps D1 and D2). At this time, device 1
When the notification of the start request is received from the server, it is determined whether the job is new or the job is restarted due to interruption. However, since the job is new, the job recording data is initialized by storing it in the procedure storage area 14 in the nonvolatile storage area 3. (D3-D5). Thereafter, an acknowledgment is transmitted to the device 1 as a response to the start request (step D6 and communication 2). Thereafter, it is determined whether the own apparatus is the transfer destination or the transfer source. However, it is determined from the content of the received job information that the apparatus is the transfer destination, and is recorded in the procedure storage area 14 as the transfer destination job status recording data (step D
7). On the other hand, the device 1 that has received the acknowledgment determines the transfer source or the transfer destination. However, the device 1 determines the transfer source based on the job information, and stores the job state record data in the procedure storage area 14 as the transfer source (step). A6 to A8).

If it is assumed that the device 2 (transfer destination) is the starting side to transfer a file to the device 1 (transfer source), a file transfer job is started from the display operation unit 18 of the device 2. Steps A1, A2, A3, A4
A5, A6, A7, A9, and C3 are transited. Steps D1, D2, D3, D4, D5, D
6, D7 and B3.

A job ID of a common value that uniquely identifies a file transfer job for both nodes during this start process
To determine. The procedure status value includes a job ID, a file name, a file attribute, information indicating whether the apparatus is a transfer source or a transfer destination, the number of delivery confirmations or reception confirmations, and the transmission address of the transfer destination. Are stored in the procedure storage area 14 in the nonvolatile storage area 13 of both nodes as job type information. The recording data of these jobs is stored until the job is completed.

Next, the process proceeds to the transfer process. As the transfer process, transmission of the transfer block (step B3) is started from the transfer source node (device 1) (communication 3 in FIG. 3).
The transfer destination node (device 2) immediately stores the received block as a part of the file 15 in the nonvolatile storage area 13 (Step C3). Immediately thereafter, the transfer destination node (apparatus 2) stores the number of transferred blocks (referred to as the number of blocks received for delivery confirmation) as the latest checkpoint in the procedure storage area 14 in the nonvolatile storage area 3. (Step C3). Subsequently, the transfer destination node (device 2) transmits a block transfer acknowledgment to inform the transfer source node (device 1) that the block transfer was successful (step C5 and communication 4). The number of transferred blocks can be identified by looking at the sequence number assigned to the header of the block at the time of transfer.

At the time of this block reception (step C)
4) If the received block data is invalid due to a transmission error or the reception times out, a retransmission request including the latest checkpoint is sent to the transfer source (step C).
6), block reception is continued (step C3).

When the reception timeout continues, the predetermined waiting time per time is increased (for example, 4, 8, 16, 32, 64, 128, 256, 51).
2, 512 [seconds]...) Repeat retransmission request and reception wait. Furthermore, if the recovery is not made after waiting for the specified number of times (for a considerably long time, for example, 2 hours), the state shifts to the suspended state, and the current program entity ends (steps C10 and C11).
After that, it waits for the restart processing (FIG. 5) from both nodes.

The transfer source node (device 1) waits for a response from the transfer destination node (device 2) to determine whether the block transfer step is successful or not (step B4). Source node (device 1)
Retransmits the block transfer only when a retransmission request is received or when the received data has a transmission error. It does not independently judge the timeout and retransmit.

When the block transfer is successful, the transfer source node (device 1) stores the number of blocks transferred as a check point in the procedure storage area 4 in the nonvolatile storage area 3.

This series of transfer processing is repeated until all blocks are transferred (step B6, step C).
4).

When all the files have been transferred, an end request is sent from the transfer source node (device 1) to the transfer destination node (device 2) (in FIG. 3, step B3 and communication 3 are shared).

Upon receiving this end request, the transfer destination node (apparatus 2) responds to the apparatus 1 with an end (step C5 and communication 4 are used in FIG. 3). Further, an end confirmation notice (communication 5 in FIG. 3) comes. Then, all the records of the job are deleted from the non-volatile storage area 14 and the processing is terminated (step C8,
C9) If the termination confirmation notification does not arrive, the transfer destination node (device 2) responds with termination (step C5 or C11).
And wait for the end notification again. Further, if the end notification is not received even after waiting for the specified number of times (a considerably long time), all records of the job are deleted from the non-volatile storage area as in the case of the normal end (step C8).

Upon receiving the end response, the transfer source node (apparatus 1) returns an end confirmation notice to the transfer destination node (communication 5 in FIG. 3), and deletes all records of the job from the nonvolatile storage area (step B7). The job is ended by displaying an end message on the display operation unit (step A).
2, B8).

Next, the operation of the abnormal processing will be described.

First, timeout processing during transfer is determined only by the transfer destination node, and unless a long-time no response is received, reception retry is continued and the protocol is interrupted (step C).
Don't do 11). Since the transfer destination node basically cannot distinguish between a communication failure and a partner node failure, it waits for a long time and then suspends.

Therefore, at the node to be started, the job is explicitly stopped from the operation display unit, or is implicitly interrupted due to power cutoff or the like. Canceling means erasing the job record from the non-volatile memory and completely stopping it. Interruption means that the job record remains in the non-volatile memory and can be restarted across power cuts of the device. It is. If the failure is not a catastrophic failure such as a communication line disconnection but is a failure that recovers naturally, a retransmission request (step C6) because timeout processing during normal block transfer corresponds to a long wait. While the retry of reception (step C3) is being executed.

In the portable terminal, the function is often stopped by running out of the battery, and this is a kind of node failure. Node failure occurs when the OS of the personal computer is low in reliability and it is desired to recover by reset. Even if the file transfer is interrupted and restarted over such a system down, the result is properly guaranteed.

Regardless of whether the program terminates abnormally or terminates after being explicitly instructed to suspend, the state at the time of suspension of the file transfer job is stored in the procedure storage area 14 in the nonvolatile storage area 13. . Therefore, after the system restart of the node (apparatus) due to power-on or reset (step E1 in FIG. 5), if a job record remains by referring to the procedure storage area 14 in the nonvolatile storage area 13 (step E2) It is determined that there is a suspended job, the program is started automatically, and the job is restarted (step E).
3). Further, when the initiating node (apparatus) explicitly performs an operation of suspending, the job can be restarted by an explicit operation. The restarted node-side program restarts the other node's program as "pre-processing", recovers the state up to the interruption point, and recovers the checkpoint synchronization.

When the restarting node finds the interrupted job information to be the transfer source (apparatus 1) in the procedure storage area 14, the steps E4 to E7 are executed to recover the state B5. At this time, in step E4, a request to restart the job is transmitted to the node of the transfer destination (apparatus 2), and the transfer destination node recovers the state C5 by receiving the restart request. The transfer destination node (apparatus 2) reads the job ID, file name, and the like of the restart request, and compares them with the contents written in the procedure storage area 14 of the own apparatus. If there is an equal job as a result of the comparison, it is determined that the file transfer is to be resumed, an acknowledgment is returned to the file transfer restart request together with the latest number of received confirmations, and block transfer reception is waited. Thereafter, the device 2 performs the same processing as in step C3 and subsequent steps in the flowchart of FIG. Further, upon receiving an acknowledgment of the restart request from the device 2, the device 1 starts transfer from the block next to the block that was completed last time. The subsequent processing is the same as the processing from step B3 in the flowchart of FIG. If the result of the comparison between the restart request and the procedure storage area 14 in the apparatus 2 is not the same, a rejection response is returned to the apparatus 1 so that the apparatus 1 deletes the record of the transfer source job from the procedure storage area 14. .

When the restarting node finds the interrupted job information to be the transfer destination (apparatus 2) in the procedure storage area 14, the steps E8 to E11 are executed to recover the state C5. At this time, in step E8, a request to restart the job to which the number of received blocks has been added is transmitted to the node of the transfer source (apparatus 1), and the transfer source node receives the restart request and returns to state B5.
To recover. The transfer source node (apparatus 1) reads the job ID, file name, and the like of the restart request and compares them with the contents written in the procedure storage area 14 of the own apparatus. If there is an equal job as a result of the comparison, it is determined that the file transfer is to be resumed, and an affirmative response to the file transfer restart request is returned. Apparatus 1
Thereafter, the same processing as in step B3 and subsequent steps in the flowchart of FIG. 3 is performed. In addition, upon receiving the restart acknowledgment from the device 1, the device 2 starts receiving from the block next to the last interrupted block. The processing is the same as the processing from step C3 in the flowchart of FIG. If the result of the comparison between the restart request and the procedure storage area 14 in the apparatus 1 is not the same, the apparatus 2 sends a restart rejection response to the apparatus 2 so that the apparatus 2 deletes the record of the transfer source job from the procedure storage area 14. I do.

Actually, as the operation of the partner (passive node) at the time of suspension / resumption, the operation according to the flowchart of FIG. 4 (steps D3, D8, D9, D10, D11, D
7), but there is a possibility that the entity running before communication is basically waiting for the program of the partner node operated by communication 1 to continue. In this state, the execution entity existing before the interruption is terminated, and the new execution entity is made effective (step D9 in FIG. 4). This is for performing the recovery processing of the circuit switching type communication line in the "preprocessing" of the new execution entity. The passive node recovers its state up to the point of interruption,
Perform the rest of the file transfer after resynchronization. That is, if it is the transfer source node, the state, step B5 is recovered, and the processing is executed from step B3. If it is the transfer destination node, the state is restored, and step C5 is recovered and executed from step C3.

If the transfer source node (apparatus 1) cannot communicate with the transfer destination node (apparatus 2) at the "pre-processing" stage of the restarted job (communication 1 and communication 2 in FIG. 5). In this case, if a predetermined number of retries of steps E4 and E5 or steps E8 and E9 are attempted (retry is performed at a predetermined time interval) and recovery is not possible, the job recording and the file being transferred are transferred from the non-volatile storage area 13 to the non-volatile storage area 13. Delete (step E12) and end the job (step E12).
13).

The time until the job is abandoned after the retry performed in the abnormal processing is assumed to be several hours to several tens of hours, so that the chance of recovery is increased. If the partner node (device) is activated, it can always wait in the passive-side start processing step D2 and start the recovery procedure, so that a long-time retry request corresponding to the power failure of the partner node (device) is retried. That's why.

As described above, according to the present invention, unless the procedure storage area 14 is destroyed, the file transfer protocol itself (operation by the program stored in the memory of the control unit 11) does not depend on the protocol of the other layers of the network architecture. ) Stores the steps of the procedure in the non-volatile storage area 13, and after the control unit 11 executes a program using the steps of the procedure to switch to another data link over a long-time line disconnection. However, it is possible to surely realize the file transfer.

In the embodiment described above, the lower protocol to be used has been described on the assumption that there is no distinction between a communication failure and a partner node failure. If a lower-level protocol that can detect a line disconnection is used, the file transfer protocol is immediately suspended when the disconnection of the communication line is detected, or the program execution is terminated once, and a new communication line is established. Alternatively, the suspension restart processing may be started. This procedure attempts to recover from the disconnection of the communication line by switching to another communication line.

[0064]

The first effect is that any lower-level protocol can be used. The recovery process of the present invention functions independently and does not depend on the operation of the lower protocol. However, there is room for optimization by the lower protocol.

The second effect is that even if there is a long interruption in the middle of the file transfer process, the file can be restarted without interruption and recovered. The reason is that the attribute and the checkpoint (procedure state) of the file transfer job are held in the non-volatile storage area, so that the restart processing can be performed again even after the end of the program or the power-off of the apparatus.

A third effect is that congestion of transmission data related to harmful retransmission can be prevented. The reason is that the determination of the timeout of the block transfer before the interruption is performed only by the transfer destination node, and the transfer source follows the determination, and the timeout of the transfer destination node at that time is repeated. In this case, useless retransmission requests are not generated by increasing the determination time.
Communication is not used at all while both nodes are suspended, and even if communication is not possible at the time of suspension restart, wasteful restart requests can be made by increasing the retry interval of restart processing to several minutes. Can be lowered.

A fourth effect is that a serious communication failure and a node failure can be recovered by the same procedure without distinction. The reason is that without using the second effect to separate both cases,
This is for recovering from a serious failure during the interruption restart processing. If a serious communication failure occurs, it is possible to recover by re-securing the communication line at the beginning of the interruption restart processing. If a node failure has occurred, the restart processing should be executed after the node failure has been recovered, so that recovery is possible.

A fifth effect is that not only the application of mobile communication but also the automatic transfer of a file transfer over a communication failure or a node failure of a general information processing apparatus can be automatically recovered and executed. The reason is that a general-purpose computer can be included in the configuration to which the present invention is applied, and the above-described first to fourth effects are obtained.

[Brief description of the drawings]

FIG. 1 is a block diagram showing the inside of the apparatus of FIG. 2 in detail.

FIG. 2 is a block diagram showing an embodiment of the present invention.

FIG. 3 is a flowchart at the time of block transfer in the embodiment of the present invention.

FIG. 4 is a flowchart relating to pre-processing of FIG. 3;

FIG. 5 is a flowchart relating to a restart process.

[Explanation of symbols]

 Reference numerals 1 and 2 3 Communication network 11 Control unit 12 Timer 13 Non-volatile storage area 14 Procedure storage area 15 File 16 Bus 17 Communication unit 18 Display operation unit 21 Recording medium

Claims (16)

[Claims] In a file transfer protocol system in which a file is divided into a plurality of blocks and transferred when a file is transferred between a plurality of devices connected to a communication network, a first device among the plurality of devices is used. Is the file transfer source and the second device of the plurality of devices is the file transfer destination, the first and second devices both have a unique job ID before performing the file transfer. The first device has a first procedure status including a job ID when a response to the start notification procedure is received from the second device. First state storage means for writing to the storage unit, block transfer start means for starting the block transfer when a response to the start notification means is received from the second device, Delivery confirmation storage means for writing the number of delivery confirmations indicating delivery confirmation to the first storage unit as a part of the first procedure state each time reception of the block is confirmed from the second device during transmission; Transfer source erasing means for erasing all the first procedure states from the first storage unit when all the blocks have been transferred, and the second apparatus includes a job ID. A second state storage unit that writes its own second procedure state to a non-volatile second storage unit before performing the file transfer; and a reception confirmation number indicating that the block has been received each time the block is received. Receiving confirmation storage means for writing to the second storage section as a part of the second procedure state, reply means for returning to the first device that the block has been received each time the block is received, When the reception is completed, the first Receiving means for waiting for an end notification from the device, and storage erasing means for erasing all the second procedure states from the second nonvolatile storage unit after receiving the end notification. File transfer protocol method to use. 2. The second device, if it does not receive the next block to be received next within a predetermined time,
A retransmission request is made to the first device by the reply means, and when the timeout continues, the retransmission request is repeated until the predetermined number of times is reached while increasing the predetermined waiting time per one time. The file transfer protocol system according to claim 1. 3. The first or second non-volatile storage unit stores the job ID, a file name, a file attribute, information indicating whether the apparatus is a transfer source or a transfer destination, 2. The communication system according to claim 1, wherein a procedure status including the number of acknowledgments and a transmission address of a transfer partner is provided.
File transfer protocol method. 4. The first device searches the procedure status stored in a first non-volatile storage unit when the device is started up, and if a file is being transferred, the job being transferred to the second device. A restart requesting means for transmitting a restart request to which an ID has been added, and when an acknowledgment to the restart request is received from the second device, the suspended state is restored and the transfer is completed before the device is started. Having block transfer restart means for starting transfer from a block, and when receiving a negative response to the restart request from the second device,
3. The file transfer protocol system according to claim 1, further comprising erasing all records of the job from the first nonvolatile storage unit and ending the job. 5. The second device, upon receiving the restart request from the first device, compares the job ID added to the restart request with the procedure state in a second nonvolatile storage unit. And returning a restart rejection response to the first device if the job does not exist, and returning a restart acknowledgment to the first device if the job exists.
The described file transfer protocol method. 6. The second device retrieves the procedure status stored in a second nonvolatile storage unit when the device is started up, and if a file transfer is being performed, recovers an interrupted state, and
4. The file transfer protocol system according to claim 1, wherein the file transfer request is sent to the first device. 7. The method according to claim 7, wherein the first device, when in a normal job reception waiting state, receives the restart request of the second device and hands over the job status to another execution entity. Item 5. The file transfer protocol system according to item 1 or 4. 8. The apparatus according to claim 1, wherein both the first and second apparatuses receive a request to start a new job or restart the interrupted job with each other, and when requested to resume the interrupted job, the first or second apparatus determines whether the current entity is executing. 8. The file transfer method according to claim 1, wherein the job is taken over. 9. A file transfer protocol method for dividing and transferring a file into a plurality of blocks when transferring a file between a plurality of devices connected to a communication network, the method comprising: When a device is the file transfer source and a second device of the plurality of devices is the file transfer destination, the device on the job starting side is a device on the passive side independently of the distinction between the transfer source and the transfer destination. Send a start request to
When the passive device responds, both devices exchange a start notification procedure for determining a job ID having a unique value before performing file transfer, and the activation device responds to the start response from the passive device. Job I when receiving
D writes its own procedure initial state to a non-volatile storage unit, and when the passive-side device receives the start request, writes its own procedure initial state to a non-volatile storage unit, and then sends a start response to the boot-side unit. Returning to the device, the first device starts transmitting a block transfer from the initial state, and the second device starts receiving the block transfer from the initial state, and stores the block in the file storage area each time it receives the block transfer. Writing the number of acknowledgments to the second non-volatile storage unit as a part of the second procedure state, returning a block transfer acknowledgment, and allowing the first device to execute the block transfer from the second device during the block transfer. Each time a block transfer acknowledgment is received, the acknowledgment number indicating the acknowledgment is regarded as a part of the first procedure state.
And when the first device has received all of the blocks, sends a file transfer end notification to the second device, and the second device sends an end acknowledgment Is returned, the first device returns a termination confirmation notice, all the first procedure states are deleted from the first nonvolatile storage unit, and the second device terminates until receiving the termination confirmation notice. A file transfer protocol method, wherein an acknowledgment is repeatedly sent, and upon receipt of an end confirmation notice, all the second procedure states are deleted from the second nonvolatile storage unit. 10. The first device searches the first procedure status stored in the first nonvolatile storage unit when the first device starts up, and if there is a file transfer status, that is, if there is an interrupted status, the first device returns to the first procedure status. The second device transmits a restart request to which the job ID is added to the passive unit of the second device. When the second device receives the restart request from the first device, the second device outputs the job ID added to the restart request and the second request. Comparing the procedure status in the nonvolatile storage unit, if there is no corresponding job, return a refusal rejection response to the first device, if there is a corresponding job, the restart acknowledgment added the number of received blocks, the When the first device receives an acknowledgment for the restart request, the transfer is started from the next block that has been transferred before the device is started, and resumption is rejected. response 10. The file transfer protocol method according to claim 9, wherein when receiving the job information, all of the corresponding job information is deleted from the first nonvolatile storage unit. 11. The apparatus according to claim 1, wherein said second device is connected to a second
The second procedure status stored in the non-volatile storage unit is searched, and if there is a file transfer status, that is, an interrupted status, the job ID and the number of received blocks are added to the passive unit of the first device. When the first device receives the restart request from the second device, the first device compares the job ID added to the restart request with the procedure state in the first nonvolatile storage unit. If the corresponding job does not exist, a return rejection response is returned to the second device. If the corresponding job exists, a restart acknowledgment is returned to the second device. When the second device receives an acknowledgment for the restart request, the second device starts reception from the next block that has been transferred before starting the device, and when the second device receives a restart rejection response, it receives a response from the second nonvolatile storage unit. Applicable File Transfer Protocol method of claim 9, wherein the erasing of all of the probe information. 12. The first or second nonvolatile storage unit stores the job ID, a file name, a file attribute, information indicating whether the apparatus is a transfer source or a transfer destination, the number of delivery confirmations, 12. The file transfer protocol method according to claim 9, further comprising a procedure state including the number of acknowledgments and a transmission address of a transfer partner. 13. When a file transfer source is used,
A start notification process for notifying the start of the file transfer by adding a procedure state including a job ID to the transfer destination of the file transfer before performing the file transfer, and a transfer source state before performing the file transfer. A transfer source state storage process for writing to a nonvolatile nonvolatile storage unit as a procedure state;
A block transfer start process for starting the block transfer when a response to the start notification process is received from the transfer destination, and a delivery indicating a delivery acknowledgment every time reception of the block is confirmed from the transfer destination during the block transfer. A delivery confirmation storage process for writing the number of confirmations to the nonvolatile storage unit as a part of the transfer source procedure state, and transferring the transfer source procedure state from the nonvolatile storage unit when all the blocks have been transferred. In order to cause the processor to execute a transfer source erasing process to be erased and to transfer the procedure status received by the start notification process to a non-volatile storage unit as a transfer destination procedure status when the transfer destination is the file transfer. The destination state storage process and the number of acknowledgments indicating that the block has been received each time the block is received are stored in the destination procedure state as 1
Reception confirmation storage processing to be written into the non-volatile storage unit as a unit, reply means for returning to the transmission source that the block has been received each time the block is received, and all the transfer when all the blocks have been received A file transfer protocol recording medium recording a program for causing a processor to execute a transfer source erasure process for erasing a pre-procedure state from the nonvolatile storage unit. 14. A retrieval process for retrieving the procedure status stored in the non-volatile storage unit at the time of device startup, and if the file is being transferred during the retrieval process, the transfer source procedure status is transmitted to the transfer destination. A restart request process for transmitting a restart request with an affixed thereto, a restart request process for recovering an interrupted state if the file is being transferred in the search process and transmitting a restart request to the transfer source, and affirming the restart request. 14. The file transfer protocol according to claim 13, wherein, when a response is received from the transfer destination, the processor causes the processor to execute a block transfer restart process that starts transfer from the next block that has been transferred before starting the apparatus. recoding media. 15. Receiving the restart request, comparing the transfer source procedure state added to the restart request with the transfer destination procedure state stored in the nonvolatile storage unit, and if the job IDs match. 15. The file transfer protocol recording medium according to claim 14, wherein the processor is made to execute returning the affirmative response to the transfer source and returning a reject response to the transfer source if the job ID does not match. 16. A processor according to claim 14, wherein when a rejection response to said restart request is received from said transfer destination, said processor deletes said transfer source procedure state from said nonvolatile storage unit. The file transfer protocol recording medium described in the above.