JPWO2018235180A1 - Distributed processing system - Google Patents

Abstract

INDUSTRIAL APPLICABILITY The present invention realizes a multitask for executing tasks in a time-division manner with a simpler configuration in a distributed processing system. In a distributed processing system in which received data is relayed between nodes 1 to transfer the data transmitted by each node 1 to all the other nodes 1, each node 1 responds to a task switching command and selects a relay target. The task that sends the data is switched to the task of the switching destination specified by the task switching command. In addition, the saved data of the task at the switching destination is restored and the relay of the data is restarted. In each node 1, subtasks 101 of a plurality of tasks are executed by multitasking, and each subtask 101 processes data of the corresponding task received from another node 1. The subtask 101 waiting for the task data is put in a standby state and the execution time is reduced, and the execution time is allocated to the subtask 101 having a process to be executed correspondingly.

Description

  The present invention relates to a task switching technique in a distributed processing system.

  As a distributed processing system, there is known a distributed processing system in which a subtask obtained by dividing a task is assigned to a plurality of nodes and the subtasks assigned in each node are processed in parallel (for example, Patent Document 1).

  In addition, as a distributed processing system, each node is connected in a ring shape by bidirectional channels to form a ring network, and data is sequentially relayed along the ring at each node, so that A distributed processing system is also known in which data sent from a node to a ring is transferred to all other nodes (for example, Patent Document 2).

JP, 2013-23499, A International Publication No. 2005/073880

The following advantages can be enjoyed by operating a computer system with multitasking.
(1) It is possible to improve the usage efficiency of the computer system by executing another task when waiting for processing such as I / O waiting.
(2) Since there is no need for queuing, even if a short process occurs after a long process is input, the short process is completed in a short time.
(3) The trouble task can be easily stopped.
On the other hand, in the above distributed processing system, when multitasking is realized by executing tasks in time division, the following problems occur.
That is, in this case, since the subtasks of the task are executed asynchronously in each node, there is no inconsistency between the processing of the subtasks in each node, and it is necessary to switch the tasks so that the tasks can be performed efficiently. , Central control of the state of the subtasks of each node, and complicated control for switching the subtasks assigned to each node according to the state of the subtasks of each node is required, and the overhead for the control occurs.

  Therefore, it is an object of the present invention to realize a multitask for executing tasks in a time-sharing manner with a simpler configuration in a distributed processing system.

  In order to achieve the above object, the present invention provides a control for simultaneously transmitting a task switching instruction accompanied by an identifier of a task to be switched to each node to a distributed processing system in which a plurality of nodes are connected by a network and which executes a plurality of tasks. With each unit, a subtask that is a part of each of a plurality of tasks executed by the distributed processing system, a subtask execution unit that executes a multitask, and a subtask of the task of each node are the originators. In order for the data of the task of which the subtask is a part to propagate through the network to all the nodes other than the node where the subtask that is the source of the data of the task is being executed, the network and the task A data transceiver that transmits and receives data, and a task data received by the data transceiver from the network. A data capturing unit that transfers the data to a subtask of the task, and data that propagates through the network in response to the task switching instruction is switched to the data of the task indicated by the identifier associated with the task switching instruction. And a propagation data switching unit that controls the operation of transmitting and receiving the data of the task of the data transmitting and receiving unit.

  Further, in order to achieve the above object, the present invention provides a distributed processing system including a plurality of nodes, which executes a plurality of tasks, with a control for simultaneously transmitting a task switching instruction accompanied by an identifier of a task to be switched to each node. With each unit, a subtask execution unit that executes a subtask, which is a part of each of the plurality of tasks executed by the distributed processing system, and a subtask execution unit that executes the multitask in each node, A sending part that sends to another node as the data of the task that the subtask is a part of, a data fetching part that transfers the task data received from another node to the subtask of the task, and a task received from another node Data to other nodes so that the data is sequentially transferred from the source node of the data to all other nodes. It is provided with a data transfer unit for. Here, in response to the reception of the task switching instruction, the transmission unit stops transmission of data of tasks other than the task indicated by the identifier accompanying the task switching instruction to another node, and the relay unit, In response to receiving the task switching command, relaying the data of the task not relayed to the other node at the time when it is received from the other node, until receiving the task switching command accompanied by the task identifier of the data. put off.

  Here, more specifically, for example, the subtask includes each of the transferred data in the processing target data, performs predetermined processing on each processing target data, and waits for the processing target data. If it becomes a state, the subtask execution unit is assumed to transition to the standby state, the subtask execution unit decreases the execution time of the subtask transitioned to the standby state, and increases the execution time of the subtasks not in other transition states. Good.

  In addition, in such a distributed processing system, in the relay unit, in response to the reception of the task switching command, the data of the task that is not relayed to the other node at the time when it is received from the other node The save data may be saved, and the saved save data of the task indicated by the identifier associated with the task switching instruction may be relayed to the other node.

  Further, in such a distributed processing system, in the data fetching unit, in response to the task switching command, transfer of the data not transferred to the subtask at the time, which is received from another node, to the subtask is performed. It may be arranged to defer until the task switching instruction with the task identifier of the data is received.

  Further, in order to achieve the above object, the present invention provides a distributed processing system including a plurality of nodes, which executes a plurality of tasks, with a control for simultaneously transmitting a task switching instruction accompanied by an identifier of a task to be switched to each node. With each unit, a subtask execution unit that executes a subtask, which is a part of each of the plurality of tasks executed by the distributed processing system, and a subtask execution unit that executes the multitask in each node, A transmitting unit that transmits to another node as data of a task that is a subtask, and the data of a task received from another node are transferred to the subtask of the task, and the data is transmitted from the node that is the source of the data. A data transfer unit that relays to other nodes is provided so that the data is sequentially transferred to all other nodes. Here, in response to the reception of the task switching instruction, the transmission unit stops transmission of data of tasks other than the task indicated by the identifier accompanying the task switching instruction to another node, and the data transfer unit In response to receiving the task switching instruction, the data received from another node at that time is saved as save data for the task, and the saved identifier associated with the task switching instruction is indicated. Transfer of task save data to the subtask and relay to the other node.

  Further, each of the above distributed processing systems may have a plurality of rings in which n (where n is an arbitrary natural number of 2 or more) nodes are connected in a ring shape by channels. Here, the number of nodes included in each ring is equal, and each node of each ring is included in the same ring as a node of each ring other than the ring including the node in a channel whose transmission side is the node. Each node is connected so as to be connected to a different node on another ring by a channel whose transmission side is the node. Then, the data transfer means of each node is configured such that the node receives data of a task received from a channel connected to a node whose connection order in the ring including the node is the order before the node. A channel that connects to a node whose connection order of the included ring in the ring is next to that of the node, and a node of a ring other than the ring in which the node is included and which is a channel that has the node as a transmission side The data of the task received from the other node is relayed to the other node by transmitting to the channel connected to the node. In addition, the transmission unit, the channel that connects the data of the task, which is the source of the subtask of the task, to the node whose connection order in the ring including the node is the next order of the node, The data of the task is transmitted to another node by transmitting to a channel connected to a node of a ring other than the ring in which the node is included and the node is the transmission side.

  According to the distributed processing system as described above, the data of the task of which the subtask of the task of each node is the source is other than the node where the subtask of the source of the data of the task is executed. In a distributed processing system configured to propagate to all other nodes via the network, the control unit broadcasts a task switching instruction accompanied by an identifier of a task to be switched to each node, The data propagating through the network can be switched to the data of the task to which the task switching instruction is switched.

  According to general multitask control, in each node, a subtask of a task whose data cannot be transferred from other nodes loses or decreases its execution time because there is no data to be processed, while other subtasks The subtasks of the task that have been transferred the data from the node sequentially generate data to be processed, so that the execution time increases.

Therefore, in each node, the task to which the task switching instruction is switched, that is, the subtask of the task in which data is transferred from another node in each node, is predominantly executed, and the distributed processing is performed. The tasks performed in the system switch.
Therefore, according to the distributed processing system as described above, it is possible to realize a multitask that executes tasks in a time-division manner with a simple configuration that does not require centralized management of the subtask state of each node.

  As described above, according to the present invention, in a distributed processing system, it is possible to realize a multitask that executes tasks in a time-sharing manner with a simpler configuration.

It is a figure which shows the distributed system which concerns on embodiment of this invention. It is a block diagram which shows the structure of the node which concerns on embodiment of this invention. It is a block diagram which shows the function of the node which concerns on embodiment of this invention. It is a block diagram which shows the structure of the task data transfer process part which concerns on embodiment of this invention. It is a figure which shows the mode of data transfer of the distributed system which concerns on embodiment of this invention. It is a flow chart which shows transfer data change processing concerning an embodiment of the present invention. It is a figure which shows the process example of the transfer data switching process which concerns on 1st Embodiment of this invention. It is a figure which shows the process example of the transfer data switching process which concerns on 1st Embodiment of this invention. It is a figure which shows the process example of the transfer data switching process which concerns on 1st Embodiment of this invention.

Hereinafter, an embodiment of a distributed system according to the present invention will be described.
FIG. 1 shows the configuration of the distributed system according to this embodiment.
As illustrated, the distributed system has one controller (CNT) and a plurality of rings (Ri), and each ring i includes the same number of nodes (Nij). Note that Ri represents the i-th ring, and Nij represents the j-th node of the i-th ring.

Here, FIG. 1a shows a case where the number of rings is two and the number of nodes in the ring is three, and FIG. 1b shows a case where the number of rings is three and the number of nodes in the ring is four.
Then, as shown in the figure, in each ring, a ring-type network is formed in which nodes included in the ring are sequentially connected in a ring shape by channels.
The j-th node of the i-th ring is connected to the j-th node of each ring other than the i-th ring by bidirectional channels.
On the other hand, the controller (CNT) is directly connected to each node (Nij) of each ring (Ri).
The connection between the controller (CNT) and each node (Nij) may be such that the node (Nij) and the controller (CNT) are connected by a channel provided for each node (Nij), or a bus type. You may make it connect a controller (CNT) and each node (Nij) with a channel.

Next, FIG. 2 shows the configuration of each node.
As illustrated, the node 1 includes a storage 11 and a processor 12. The processor 12 has a configuration as a computer including a CPU and its peripheral devices, and can perform data processing independently of other nodes 1 using the storage 11.

  The node 1 also includes a channel interface 13 that is connected to the node 1 and that is provided for each channel and that uses the channel to perform data transmission with another node 1.

Next, FIG. 3 shows a configuration of functional blocks of the node 1 (Nij) formed by execution of software of the processor 12.
A task part obtained by dividing a task performed by the distributed processing system is called a subtask, and the node 1 is assigned with subtasks 101 of a plurality of tasks. Each sub-task 101 is executed in a time-division manner in the form of multi-task under the control of a multi-task control unit 1021 which provides a multi-task environment by scheduling the sub-task 101 and managing resources on the operating system 102. .

  Next, each subtask 101 has its own data space 1011 and performs processing using its own data space 1011. Then, each data space 1011 is provided with a transmission queue SQ for storing data to be transmitted to another node and a reception queue RQ for storing data received from another node.

  Then, the operating system 102 transfers the data stored in the transmission queue SQ of each subtask 101 to another node 1 and stores the data received from the other node 1 in the reception queue RQ. The processing unit 1022 is provided.

Next, FIG. 4 shows the configuration of the task data transfer processing unit 1022 of the operating system 102.
As illustrated, the task data transfer processing unit 1022 controls the reception buffer 41, the transmission buffer 42, the save buffer 43, the duplicate transfer suppression unit CHK44, the reception data transfer unit 45, the transmission data transfer unit 46, and the above-mentioned transfer control units. The unit 47 is provided.

  The reception buffer 41 includes a buffer RS (r) in which the data received by the channel interface 13 from the node 1 in the ring which is included in the same ring as the own node 1 and whose order on the ring is one before is stored. There is.

Here, the order on the ring is the order of the node 1 which is passed when the channel data transfer direction is advanced along the ring.
In addition, the reception buffer 41 stores the data received by the channel interface 13 from the node 1 having the same order in the ring of another ring different from the ring including the own node 1, and has the same number as that of the other ring. It has a buffer Ri (r).

  Next, the transmission buffer 42 includes a buffer RS (s) that stores the data to be transmitted via the channel interface 13 to the node 1 which is included in the same ring as the own node 1 and is one order later on the ring. I have it. Further, the transmission buffer 42 stores the data to be transmitted via the channel interface 13 to the node 1 having the same order in the ring of another ring different from the ring including the own node 1, and the same number as the number of the other ring. Buffer Ri (s).

Then, the transfer control unit 47 normally controls each unit so that the following operations are performed.
That is, the transfer control unit 47 sets the task, which is the execution target task, in the transmission data transfer unit 46 as the transmission target task. The task to be executed will be described later.
The transmission data transfer unit 46, to which the transmission target task is set, transfers the data stored in the transmission queue SQ of the data space 1011 of the subtask 101 of the transmission target task to the buffer RS (s) of the transmission buffer 42 and each buffer Ri (s). ), The node 1 in the same ring that is included in the same ring as the own node 1 is one order later in the ring, and the other rings that are different from the ring in which the own node 1 is included have the same order in the ring. Send to each of node 1.

The data stored in the transmission queue SQ of the data space 1011 of the subtask 101 is added with a task identifier indicating the identification of the task to which the subtask 101 belongs.
Further, the transfer control unit 47 receives the data stored in the buffer RS (r) of the reception buffer 41, which is received from the node 1 which is included in the same ring as the own node 1 and whose order on the ring is one before. , To the duplicate transfer inhibiting unit CHK44. The duplicate transfer inhibiting unit CHK44 checks whether or not the transferred data is the source data of the own node 1, and outputs the data only when the own node 1 is not the source data. If the source data, the data is discarded.

  Then, the data output from the duplicate transfer inhibiting unit CHK44 is transferred to the buffer RS (s) and each buffer Ri (s) of the transmission buffer 42, and the data is included in the same ring as the own node 1. It is transmitted to each of the node 1 that is one order later in the ring and each node 1 that has the same order in the rings of the other rings different from the ring in which the own node 1 is included.

  The data output from the duplicate transfer inhibiting unit CHK44 is also transferred to the received data transfer unit 45, and the received data transfer unit 45 uses the transferred data as the subtask 101 of the task having the task identifier added to the data. The data space 1011 is stored in the reception queue RQ.

  Further, the data received from each of the nodes 1 having the same order in the rings of the other rings different from the ring including the own node 1 and stored in each Ri (r) is transferred to the reception data transfer unit 45, The reception data transfer unit 45 stores the transferred data in the reception queue RQ of the data space 1011 of the subtask 101 of the task having the task identifier added to the data.

  Hereinafter, the data transfer operation between the nodes 1 realized by the operation of the task data transfer processing unit 1022 will be described by taking a distributed system having two rings and three nodes in the ring as an example.

Now, as shown in FIG. 5a, in the node N11 of the ring R1, when the task A is the task to be executed, the transmission data transmitted by the subtask 101 of the task A is transferred from the node N11 to the nodes N12 and N21, It is sent to the subtask 101 of the task A of the node N12 and the node N21. The data sent to the node N12 is relayed by the node N12 to the nodes N13 and N22 and sent to the subtask 101 of the task A of the nodes N13 and N22. Further, the data sent to the node N13 is relayed to the node N23 by the node N13 and sent to the subtask 101 of the task A of the node N23.
As a result, when the task A is the task to be executed, the transmission data transmitted by the subtask 101 of the task A of the node N11 is sequentially propagated to the nodes 1 one by one as described above, and all the nodes other than the own node N11. Is transferred to the subtask 101 of task A.

Similarly, as shown in FIG. 5B, in the node N12 of the ring R1, when the task B is the task to be executed, the transmission data transmitted by the subtask 101 of the task B is transmitted from the node N12 to the node N13 as illustrated. Are transferred to the node N22 and sent to the subtask 101 of the task B of the nodes N13 and N22. The data sent to the node N13 is relayed by the node N13 to the nodes N11 and N23 and sent to the subtask 101 of the task B of the nodes N11 and N23. The data sent to the node N11 is relayed to the node N21 by the node N11 and sent to the subtask 101 of the task B of the node N21.
As a result, when the task B is the execution target task, the transmission data of the subtask 101 of the task B of the node N12 is sequentially propagated to the nodes 1 one by one as described above, and is transmitted to all the nodes 1 other than the own node N12. It is transferred to the subtask 101 of task B.

  Hereinafter, similarly, the transmission data transmitted by the subtask 101 of the arbitrary task X of the arbitrary node 1 (Nij) of the arbitrary ring, when the task X is the execution target task, It is transferred to the subtask 101 of the task X of all the nodes 1.

  Therefore, when the task X is the task to be executed, the sub-task 101 of the task X of each node 1 of each ring transmits data, and the sub-task 101 of the task X of each node 1 of each ring receives another data. The data transmitted by the subtasks 101 of the task X of all the nodes 1 are transferred.

The operation of switching tasks in such a distributed processing system will be described below.
First, the subtask 101 of each task of each node 1 according to this embodiment performs the following processes P1 and P2.
Process P1; predetermined calculation is performed, data representing the calculation result is stored in the transmission queue SQ, and is transmitted to the subtask 101 of the same task of another node.
Process P2: The data received from the subtask 101 of the same task of another node stored in the reception queue RQ is taken out, and the operation result represented by the data is integrated with the operation result of the process P1.

  However, the calculation of the process P2 is performed for the data received from the subtask 101 of the same task of all other nodes. In addition, at least the operation of the process P2 has no order dependency, and the same result is finally obtained no matter which order the data received from the subtask 101 of the same task of another node 1 is processed. Is.

  Further, when there is no processing to be performed in each subtask 101, that is, the processing P1 has been completed and the data to be processed next in processing P2, which is received from the subtask 101 of the same task of another node, is the reception queue RQ. If it does not exist in the queue, then it shifts to the standby state until the data stored in the reception queue RQ is received.

  Then, the multitask control unit 1021 of the operating system 102 of each node 1 does not give the execution time (CPU time) to the subtask 101 in the standby state, or reduces the execution time, and the standby state is not in the corresponding state. The execution time (CPU time) of the other subtask 101 is increased.

Now, returning to FIG. 1, the controller (CNT) switches tasks of the distributed processing system according to a predetermined schedule.
Further, the task switching is performed by the controller (CNT) simultaneously sending a task switching command designating the task identifier of the task to be switched to each node 1 (Nij) of each ring when switching the task.

On the other hand, the transfer control unit 47 of the task data transfer processing unit 1022 of each node 1 (Nij) performs the transfer data switching process shown in FIG. 6 in order to switch the task.
As shown in the figure, the transfer control unit 47 monitors the occurrence of the reception of the task switching instruction transmitted by the controller (CNT) (step 602), and when the reception of the task switching instruction occurs, the transfer target task The task is switched to the switching destination task indicated by the task identifier of the task switching instruction, and the task that is the execution target task is set in the transmission data transfer unit 46 as the transmission target task (step 604).

  When the transmission target task is switched in step 604, the data stored in the transmission queue SQ of the data space 1011 of the subtask 101 of the transmission target task after the switching is transferred to the same ring as the own node 1 as described above. Node 1 which is one order later on the ring that is included and each of the nodes 1 whose ring order of the other rings different from the ring in which the own node 1 is included are transmitted to each node 1, The data stored in the transmission queue SQ of the data space 1011 of the subtask 101 of the transmission target task before switching is not transmitted to another node.

  Further, next, the transfer destination of the data stored in the buffer RS (r) of the reception buffer 41 is changed from the duplicate transfer inhibiting unit CHK44 to the save buffer 43, and the save buffer 43 is changed to the buffer RS (r). The transferred data is stored as the save data of the task indicated by the task identifier added to the data, and the save data of the task of the execution target task after switching in step 604, which is stored in the save buffer 43, is also stored. , To the duplicate transfer inhibiting unit CHK44 (step 606).

The duplicate transfer inhibiting unit CHK44 processes the data transferred from the save buffer 43 in the same manner as the data transferred from the buffer RS (r) of the reception buffer 41.
That is, the duplicate transfer inhibiting unit CHK44 checks whether or not the data transferred from the save buffer 43 is the source data of the own node, and only when the own node is not the source data, the data is transferred to the transmission buffer. 42 of the buffer RS (s) and each of the buffers Ri (s), and the node 1 in the same ring as the own node 1 is one order later on the ring and the ring including the own node 1 While transmitting to each of the nodes 1 having the same order in the ring of each of the other rings different from that, the data is transferred to the reception data transfer unit 45, and the subtask 101 of the task with the task identifier added to the data is transferred. The data is stored in the reception queue RQ of the data space 1011.

Then, the transfer destination of the data stored in the buffer RS (r) of the reception buffer 41 is returned to the duplicate transfer inhibiting unit CHK44 (step 608), and the process returns to the monitoring of step 602.
The transfer data transfer process performed by the transfer control unit 47 of the task data transfer processing unit 1022 has been described above.
Here, processing examples of such transfer data switching processing are shown in FIGS.
In each figure, Dk (Nij) represents data whose source is the subtask 101 of the task k of the jth node 1 of the ith ring.
Now, as shown in FIG. 7a, task A is set as a task to be executed, and the data of task A originating from subtask 101 of task A of each node 1 causes node 1 to pass through node 1 as described above. FIG. 7b shows a task switching instruction sent from the controller (CNT) to each node 1 (Nij) in the middle of the data propagation of the task A while the data is being propagated one by one. As described above, in each node 1, the data of the task A that has not been transferred to the other nodes and the subtasks 101 being received at that time is saved in the save buffer 43 as the save data of the task A, and the data of the task A is propagated. Will stop.

  Further, in each node 1, the task B is set as the execution target task, the data of the task B originating from the subtask 101 of the task B of each node 1 is transmitted, and the data of the task B is shown in FIGS. 7b and 8a. As shown, the data of the task B is propagated through the nodes 1 one by one, and is transferred to the subtask 101 of the task B of each node 1.

  Then, in the course of the propagation of the data of the task B shown in FIG. 8A, when the task switching instruction having the task A as the switching destination is transmitted from the controller (CNT) to each node 1 (Nij), as shown in FIG. 8B. Further, in each node 1, the data of the task B that has not been transferred to the other node and the subtask 101 that are being received at that time is saved in the save buffer 43 as the save data of the task B, and the data of the task B is not propagated. Stop.

  Also, the saved data of the task A that has been saved is transmitted from each node 1, whereby the propagation of the data of the task A is restarted and the data of the task A is changed as shown in FIGS. 8B and 9. It is transferred to the subtask 101 of task A of each node 1.

  As described above, according to the present embodiment, the controller (CNT) transmits the task switching command to each node 1 (Nij) to transfer the data between the nodes 1 (Nij) of the distributed processing system. You can switch the tasks to be performed.

  Then, in each node 1, the subtask 101 of the task that has not been able to transfer the data from the other node 1 is in the standby state because there is no data to be processed in the above-described process P2, while the data is transferred from the other node 1. The subtask 101 of the task thus started generates data to be processed in the above-described processing P2.

  Therefore, in each node 1, the execution of the task to the switching destination of the task switching command transmitted by the controller (CNT), that is, the subtask 101 of the task in which data is transferred from another node 1 in each node 1 is executed. It becomes dominant, and the tasks executed in the distributed processing system are switched.

That is, the task switching in the distributed processing system is realized by transmitting the task switching command from the controller (CNT).
Therefore, according to the present embodiment, it is possible to realize a multitask that executes tasks in a time-sharing manner with a simple configuration that does not require centralized management of the state of subtasks of each node.
It should be noted that FIGS. 7, 8 and 9 show the case where each node 1 operates in synchronization, but in reality, the operation of each node 1 may be asynchronous. Even in this case, when the task switching command is transmitted from the controller (CNT) to each node 1 (Nij), the data of the task of the switching destination specified by the task switching command is immediately predominant between the nodes 1 (Nij). Will be transferred to. Therefore, in each node 1, the subtask 101 of the task of the switching destination designated by the task switching instruction is quickly and dominantly executed, and the task executed in the distributed processing system is switched.

  By the way, in the above embodiment, the controller (CNT) transmits the pend command specifying the task identifier and the stop command specifying the task identifier to each node 1 (Nij) in addition to the task switching command. May be.

  The pend instruction is an instruction to temporarily stop the task, and when each node 1 receives the pend instruction, it suspends the transfer of the data to which the task identifier specified by the pend instruction is added and transfers the data. In a resumable form.

  The stop instruction is an instruction to stop the task, and when each node 1 receives the stop instruction, it clears the transmission queue SQ and the reception queue RQ of the subtask 101 of the task indicated by the task identifier specified by the stop instruction, and thereafter When the data with the task identifier specified by the stop command is received, the data is discarded.

  1 ... Node, 11 ... Storage, 12 ... Processor, 13 ... Channel interface, 41 ... Reception buffer, 42 ... Transmission buffer, 43 ... Saving buffer, 44 ... Duplicate transfer inhibiting unit CHK, 45 ... Reception data transfer unit, 46 ... Transmission Data transfer unit, 47 ... Transfer control unit, 101 ... Subtask, 102 ... Operating system, 1011 ... Data space, 1021 ... Multitask control unit, 1022 ... Task data transfer processing unit.

Claims (7)

A distributed processing system that executes multiple tasks by connecting multiple nodes in a network,
Each node is provided with a control unit that broadcasts a task switching instruction accompanied by the task identifier of the switching destination,
Each node is
A subtask execution unit that executes a subtask, which is a part of each of a plurality of tasks executed by the distributed processing system, by a multitask,
The subtask of the task of each node is the source, and the data of the task of which the subtask is a part is transmitted via the network to all the nodes other than the node where the subtask of the source of the data of the task is executed. Data transmission / reception unit that transmits / receives data of the task to / from the network,
A data acquisition unit that transfers the task data received from the network by the data transmission / reception unit to a subtask of the task,
In response to the task switching instruction, the data transmission / reception operation of the data transmitting / receiving unit is performed so that the data propagating through the network is switched to the task data indicated by the identifier associated with the task switching instruction. A distributed processing system comprising: a propagation data switching unit for controlling. A distributed processing system including a plurality of nodes, which performs a plurality of tasks,
Each node is provided with a control unit that broadcasts a task switching instruction accompanied by the task identifier of the switching destination,
Each node is
A subtask execution unit that executes a subtask, which is a part of each of a plurality of tasks executed by the distributed processing system, by a multitask,
A transmission unit that transmits data whose source is a subtask of a task, to another node as data of the task of which the subtask is a part,
A data capturing unit that transfers the task data received from another node to a subtask of the task,
The data of the task received from another node has a data transfer unit that relays the data to another node so that the data is sequentially transferred from the node that is the source of the data to all other nodes,
The transmission unit, in response to receiving the task switching instruction, stops transmission of data of tasks other than the task indicated by the identifier associated with the task switching instruction to another node,
The relay unit, in response to receiving the task switching command, relays data of a task that has not been relayed to another node at the time when it is received from another node, and switches the task with the task identifier of the data. A distributed processing system characterized by postponing until an instruction is received. The distributed processing system according to claim 2, wherein
The subtask includes each of the transferred data in the data to be processed, performs a predetermined process on each data to be processed, and when the data to be processed is in a waiting state, enters the standby state. Transition,
A distributed processing system, wherein the subtask execution unit reduces the execution time of a subtask that has transited to a standby state and increases the execution time of a subtask that is not in another transition state. The distributed processing system according to claim 2 or 3, wherein
In response to receiving the task switching command, the relay unit saves the data of the task that is not relayed to the other node at the time when it is received from the other node as the save data of the task and is saved. The distributed processing system is characterized in that the save data of the task indicated by the identifier associated with the task switching instruction is relayed to the other node. The distributed processing system according to claim 2, 3 or 4, wherein
In response to the task switching command, the data capturing unit transfers the data, which is received from another node and is not transferred to the subtask at that time point, to the subtask, and switches the task with the task identifier of the data. A distributed processing system characterized by postponing until an instruction is received. A distributed processing system including a plurality of nodes, which performs a plurality of tasks,
Each node is provided with a control unit that broadcasts a task switching instruction accompanied by the task identifier of the switching destination,
Each node is
A subtask execution unit that executes a subtask, which is a part of each of a plurality of tasks executed by the distributed processing system, by a multitask,
A transmission unit that transmits data whose source is a subtask of a task, to another node as data of the task of which the subtask is a part,
Transfers the task data received from other nodes to the subtasks of the task, and relays the data to other nodes so that the data is sequentially transferred from the source node of the data to all other nodes. And a data transfer unit that
The transmission unit, in response to receiving the task switching instruction, stops transmission of data of tasks other than the task indicated by the identifier associated with the task switching instruction to another node,
In response to receiving the task switching instruction, the data transfer unit saves the data received from another node at that time as the save data of the task, and the saved task switching instruction. The distributed processing system is characterized in that the save data of the task indicated by the identifier is transferred to the subtask and relayed to the other node. The distributed processing system according to claim 2, 3, 4, 5 or 6, wherein
n (where n is an arbitrary natural number of 2 or more) has a plurality of rings in which the nodes are connected in a ring shape by a channel,
The number of nodes in each ring is equal,
Each node of each ring is a channel having the node as a transmission side, and a node of each ring other than the ring including the node, each node included in the same ring is a channel having the node as a transmission side, Are connected so that they are connected to different nodes on other rings,
The data transfer means of each of the nodes includes the data of the task received from the channel connected to the node of the order in which the ring including the node is connected in the ring in the order before the node includes the node. A channel that connects to a node that has the next order of connection of the ring in the ring, and a channel that has the node as a sender and that is a ring other than the ring in which the node is included. By relaying the data of the task received from the other node to another node,
The transmitting unit connects the node with a channel that connects the data of the task whose source is a subtask of the task, to a node whose connection order in the ring including the node is next to that node. A distribution characterized in that the data of the task is transmitted to another node by transmitting to a channel connected to a node of a ring other than the ring in which the node is included, which is a transmission side. Processing system.

Images