JP3007324B2 - Distributed processing system and method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a distributed processing system in which a plurality of computer nodes are connected via a network, and a plurality of processes operating on the plurality of computer nodes perform processing while performing message communication between the processes. And its method.

[0002]

2. Description of the Related Art There is a distributed processing system in which a process that operates on a server device is transmitted and called by a process that operates on a client device disposed at a remote place. In such a distributed processing system, for example, there is a case where it is desired to exchange or change a process operating on a server device from a predetermined time. For example, there may be a case where it is desired to change the process A that operates on the server device and performs the processing of the consumption tax of 3% to the process B that performs the processing of the consumption tax of 5% at 00:00:04 on April 1. In this case, for example, April 1, 0
At 00:00, the server device is stopped or the process A
There is a method of rejecting the reception of the message to the process A, and in the meantime, changing the process A for processing the consumption tax 3% to the process B for processing the consumption tax 5%.

[0003]

However, according to this method, for example, a message is output from the client device immediately before 0:00:00 on April 1, and after 0: 00: 00: 00 on April 1,
For the call request received by the server device, the process B processes the consumption tax of 5%. Further, when switching from the process A to the process B, the server must be stopped, and confusion is expected.

In addition, the server is stopped at, for example, 10:00:00 on March 31 before 00:00:00 on April 1, and the process A is included in the message, for example. 3% consumption tax according to the output time of the message
C that selectively performs the processing of the consumption tax and the processing of the consumption tax of 5%
There is a way to replace. According to this method, for example, a message that is output from the client device immediately before 00:00:00 on April 1 and that is received by the server device after 00: 00: 00: 00 on April 1 is accurately output. 3% processing before consumption can be performed. However, even with this method, when switching from process A to process C, the server device must be stopped, and confusion is expected.

[0005] Normally, a process operating on a server device is used by successively switching to a new process due to factors such as a change in social conditions and provision of a new service. This is inconvenient. On the other hand, it is effective to incorporate processing required in the future into the process in advance and switch processing based on predetermined conditions, but it is possible to accurately predict the processing required in the future. It is often difficult to do so.

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned problems of the related art, and when switching a running process to another process under a predetermined switching condition, a message can be processed by an appropriate process according to the switching condition. It is an object to provide a distributed processing system and a method thereof. In addition, the present invention does not stop the operation of the system,
An object of the present invention is to provide a distributed processing system capable of switching a running process to another process and a method thereof.

[0007]

According to the distributed processing system of the present invention, a plurality of computer nodes are connected via a network, and a plurality of processes operating on the plurality of computer nodes perform message communication between the plurality of processes. A distributed processing system for performing processing, comprising: a first process for outputting a message including predetermined attribute information in order to invoke a process of another process; and a second process for performing a process according to the input message. And a management process for generating the third process, which is the new process, based on a process replacement start request instructing replacement of the second process with a new process when a predetermined condition is satisfied. And the first process comprises:
The message is output to the second process, and the third process, when generated by the management process, outputs a message transfer information setting request to the second process, and performs processing according to the input message. The second process sets a message transfer condition and reference information of the third process which is a transfer destination process of the message in response to the message transfer information setting request, and sets the attribute of the input message. When the information does not satisfy the message transfer condition, a process according to the input message is performed, and when the message transfer condition is satisfied, the input message is sent to the third process using the reference information. Forward.

In the distributed processing system according to the first aspect of the present invention, there are a first process and a second process, which perform processing while performing message communication. Then, based on a process replacement start request instructing replacement of the second process with a new process when a predetermined condition is satisfied, a third process, which is the new process, is generated by the management process. You.
Then, a message transfer information setting request is output from the third process to the second process. The second process sets, in response to the message transfer information setting request, a message transfer condition and reference information for referring to the third process, which is a process to which the message is transferred. Thereafter, the second process performs a process according to the input message when the attribute information of the input message does not satisfy the message transfer condition, and performs the input when the attribute information of the input message satisfies the message transfer condition. A message is forwarded to the third process using the reference information. The third process is
Processing is performed in accordance with the message transferred from the second process.

In the distributed processing system according to the present invention, preferably, the management process includes location information of a program module, information for specifying the second process, and location information of a computer node for arranging a new process. , The third process is generated by reading the program module into the computer node and executing the same.

Further, in the distributed processing system according to the present invention, preferably, when the third process is generated by the management process, a remote identifier of an object existing in the third process, Generates a local reference table indicating a correspondence between a local identifier which is an identifier of the object in the process, a message transfer condition, and a reference to a message transfer destination object, and generates a message transfer condition of the local reference table and a message. Perform export processing to unreference the transfer destination object.

Further, in the distributed processing system according to the present invention, preferably, the third process is a remote identifier which is an identifier of the object on the network with respect to an object of another process making a call,
Import processing is performed to generate a remote reference table indicating the correspondence between the process identifier of the process to which the object belongs and the location information of the computer node where the process exists.

In the distributed processing system according to the present invention, preferably, the management process has a process arrangement table indicating a correspondence between a process identifier of the process and location information of the computer node on which the process operates. ,
The third process queries the management process to obtain the process identifier and the location information of the computer node for each object in the remote lookup table.

In the distributed processing system according to the present invention, preferably, the third process, when the second process exists, sends a message transfer information setting request including the local reference table and a message transfer condition. The second
And the second process transfers the input message to the third process when the attribute information of the input message satisfies the predetermined condition. Using the local reference table and the message transfer condition of the third process included in the setting request, update the local reference table and the remote reference table of the second process.

Further, in the distributed processing system according to the present invention, preferably, the third process inputs the updated local reference table from the second process, and executes the local process of the input second process. The local reference table and the remote reference table of the third process are updated using the reference table and the message transfer condition.

In the distributed processing system according to the present invention, preferably, when the message input by the second process does not include any of the messages that do not satisfy the message transfer condition, the message is not included in the first process. The remote identifier of the concurrent object entry of the remote reference table of the process having a reference to said second process,
Change to the remote identifier of the concurrent object of the third process.

Also, in the distributed processing method according to the present invention, a plurality of computer nodes are connected via a network, and a plurality of processes operating on the plurality of computer nodes perform processing while performing message communication between the plurality of processes. A processing method, wherein a message is output from a first process to a second process, and when a predetermined condition is satisfied, a process replacement start request instructing replacement of the second process with a new process is provided. Generating a third process, which is the new process, based on the input message, and outputs a message transfer information setting request to the second process when the third process is generated. Processing, the second process comprises:
In response to the message transfer information setting request, a message transfer condition and reference information of the third process which is a transfer destination process of the message are set, and the attribute information of the input message does not satisfy the message transfer condition. At this time, processing according to the input message is performed, and when the message transfer condition is satisfied, the input message is transferred to the third process using the reference information.

In the distributed processing method according to the present invention, preferably, the process replacement start request includes a location information of a program module, information for specifying the second process, and a computer node for allocating a new process. Location information, and the program module is read into the computer node and executed to execute the third module.
Spawn a process.

[0018]

[0019]

[0020]

[0021]

[0022]

[0023]

[0024]

[0025]

[0026]

[0027]

[0028]

[0029]

[0030]

[0031]

[0032]

[0033]

[0034]

[0035]

[0036]

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described.
The distributed processing system and its method will be described. Figure
1 is a configuration diagram of the distributed processing system 100 of the present embodiment.
is there. As shown in FIG. 1, the distributed processing system 100
Via the network 8, the computer node 10₁,
10_Two, 10_ThreeIs connected. Network 8
For example, ISDN (Integrated Services Digital Network
rk) Communication line, cable television (CATV) communication
Line and Asynchronous Transfer Mode (ATM)
fer Mode) A digital communication line such as a communication line.
The network 8 usually includes a computer node
10₁, 10_Two, 10 _ThreeMany other computers
Card is connected. Computer node 10₁, 1
0_Two, 10_ThreeHas the node name "node
eA "," nodeB ", and" nodeC "
You.

[0037] In the computer node 10 _1, for example,
By not shown CPU executing a program module stored in the internal memory, the management process 12 ₁
And process 16 ₁ is operating. Process 16 ₁
Has a process name “XonA” and a process ID “10” as a process identifier. In the computer node 10 _2, by the CPU executing a program module stored in the internal memory, the management process 12 ₂ and process 16 ₃ running.

[0038] Here, the process 16 _3, for example, a process of performing consumption tax 3% processing, is generated by the CPU executing a program module 14 _1. In the present embodiment, the CPU 16 executes the program module 14 ₂ to execute the process 16 _{4 for} processing the consumption tax of 5%.
There is generated in the future, the process of the consumption tax by consumption tax processing message replaces the process 16 ₃ to process 16 _4. It should be noted, called the process 16 ₃ with the existing process,
A process ID “20” is assigned. In addition, the process 16 ₄ called the new process, subjecting the process ID "21". The existing process 16 _3, and a new process 16 _4, the same process name "YonB" is attached.

In the computer node 10 ₃ , the CPU executes the program modules stored in the internal memory so that the management process 12 ₃ and the process 16
₂ is working. In addition, the process 16 _2, process name "XonC", are assigned and the process ID "30".

Hereinafter, the distributed processing system 100 shown in FIG.
At the computer node via the network 8
10_TwoIs called by a message entered in
Process, process 16_ThreeTo process 16_FourSwitch to
The following describes an outline of the obtaining process. First, the computer
Node 10_TwoAbove, process 1 to process 3% consumption tax
6 _ThreeIs working.

Next, the master management process 12
_3, accepts a process replacement start request 70 for instructing to replace the existing process 16 ₃ operating on the computer node 10 ₂ in the new process 16 _4.

Next, the management process 12 ₂ from the management process 12 _3, the process generation request 71 is output. Next, the management process 12 _2, the process generates 72 takes place, the new process 16 ₄ is generated.

[0043] Then, the existing processes 16 ₃ and novel processes 16 _4, transfer conditions and transfer destination address of the message is set. At this time, the existing process 16 ₃
"April 1 0: 00: 00: 00"
Seconds after "is set, the address of the new process 16 ₄ is set as the destination address. Also, the new process 16 ₄ has “0 April 1
At 0 minutes before 0 seconds "is set, the address of the existing processes 16 ₃ is set as the destination address.

[0044] As a result, the messages that are input to the new process 16 ₄ before than 0 minutes 0 seconds at April 1, 0, is transferred from the new process 16 ₄ to the existing process 16 _3,
It is processed in existing processes 16 _3. Meanwhile, the message input to existing processes 16 ₃ to 1 April 00:00:00 later, are transferred from the existing process 16 ₃ new process 16 _4, are treated with a new process 16 _4.

Hereinafter, the components of the distributed processing system 100
Will be described in detail.Computer node [Overview of Computer Node] FIG.
Computer node 10 of processing system 100_Two, 10
_ThreeFIG. 2 is a diagram showing an outline of an internal configuration of the device. In FIG. 2,
For simplicity of explanation, the computer node shown in FIG.
C10₁Is omitted. As shown in FIG.
Data node 10_TwoIs a process 1 as a general process
6_TwoAnd management process 12_ThreeHaving. Also, compilation
Computer node 10 _ThreeIs a process as a general process
16_ThreeAnd management process 12_TwoHaving. General processing
Application processes within the application and management processes.
Distributed objects and distributed functions as program modules
Number or global shared variables exist and their application
Message module sends and receives messages
There is a message passing function that supports communication
I do. Note that fractional objects, distribution functions, global shared variables
Concurrent object (Concurrent Obje)
cts). Here, an object is its own internal data.
It is a program with data and procedures.

Although a plurality of concurrent objects can exist in one process, only one message passing function exists in one process. Also, one computer node can have a plurality of general processes, but only one management process. In the distributed processing system 100, a plurality of processes cooperate by a message passing function.

[0047] [General Process Detail] FIG 3 is a diagram for explaining a message passing function of the process 16 ₂ and 16 _3. As shown in FIG.
6 ₂ and 16 _3, as a message passing function, an application programming function (Applicati
on Programming Features) 252, Message Passing Operator 253,
Reference Resolver 254, Communication Media Interface (Communi
cation Media Interface) 255, Message Evaluation Scheduler 257, Message Dispatcher 258,
Development Tools 260, System Initializer 261 and Operating System Interface (Operating Sys.
tem Interface) 262.

Application programming function 2
52: Application programming function 252
Has a function of receiving a message transmission request from a concurrent object and a function of calling a function of a concurrent object such as a method or a function according to an incoming message. Further, the application programming function 252 includes a part that takes an implementation form depending on a processing system of a programming language in which a concurrent object is implemented, and a part that takes a common implementation form without depending on the programming language. . The application programming function 252 also performs a process of converting argument data included in the message between a common expression and an internal expression. That is, the application programming function 252 provides the AP associated with message sending and receiving.
Providing I (Application Programming Interface)

Reference resolver (reference resolver)
254: The reference resolver 254 globally manages references from and references to concurrent objects, their functions and methods, which exist locally (in the same process) or remotely (in a different process).

FIG. 4 shows a reference resolver 254.
FIG. As shown in FIG. 4, the reference resolver 254 includes an export processing unit 31, an import processing unit 32, a remote I
It comprises a D notification unit 33, a transfer information setting unit 34, a transmission reference resolution unit 35, a reception reference resolution unit 36, and a reference table accessed from these units. The reference table includes a local reference table 37, a remote reference table 38, a process reference table 39, and a node reference table 40.

For example, as shown in FIG. 28, the local reference table 37 has a globally effective remote ID (remote identifier) for a concurrent object in its own process and a local ID (local identifier) effective only in its own process.
And stores the names of the concurrent objects in the own process, the remote ID, and the local ID. In addition, the local reference table 37 further includes, for a parallel object that needs to be transferred to a parallel object of another process, a reference to the entry of the transfer destination parallel object in the remote reference table 38 and a message transfer condition. Store.

For example, as shown in FIG. 28, the remote reference table 38 includes the names of the concurrent objects existing in other processes, the remote IDs of the concurrent objects, and the process IDs of the processes in the process reference table 39 in which the concurrent objects exist. Stores a reference to the entry.

The process reference table 39 stores a process name, a process ID, and a reference to an entry of a node where the process exists in the node reference table 40, as shown in FIG. 28, for example.

The node reference table 40 stores a node name, a node network address, and a communication medium as shown in FIG. 28, for example.

The export processing unit 31 performs an export process to generate a local reference table 37 as described later. As a result, the concurrent object of the own process can be referenced from another process. In the export process, for the concurrent objects that exist in the own process,
The remote ID and the local ID are assigned to be unique, and the local reference table 37 is created. In the export process, the message transfer condition in the local reference table 37 and the reference to the entry of the transfer destination parallel object in the remote reference table 38 are not set and are left blank.
Incidentally, the export processing is executed by the initial processing of new processes 16 ₄ shown in FIG. 14 (step S31).

The import processing unit 32 is provided in FIG.
To generate and update the remote reference table 38.

When a remote ID notification request including a list of concurrent object names requiring a remote ID is input from another process, the remote ID notifying unit 33 refers to the local reference table 37 of the own process, for example. The remote ID of the concurrent object name included in the remote ID is obtained, and the remote ID is notified to the other process that issued the remote ID notification request. The remote ID notifying unit 33 includes, for example,
In step S56 of the import processing in the import processing unit 32 shown in FIG. 20, which will be described later, the local reference table 3 refers to the remote ID corresponding to the concurrent object name included in the remote ID notification request, which is input from another process.
7 and outputs the remote ID to the other process that has output the remote ID notification request. The remote ID notifying unit 33 is operable after the local reference table 37 is created.

The transfer information setting unit 34 performs the processing shown in FIG. 27, which will be described later, and implements a message transfer between parallel objects having the same name belonging to different processes. Further, an entry is added to the remote reference table 38, a reference is made to the entry of the transfer destination parallel object existing in the local reference table 37, and a message transfer condition is set.

When transmitting a message to a concurrent object existing in another process, the transmission time reference resolving unit 35
The remote reference table 38, the process reference table 39, and the node reference table 40 are sequentially referred to, the remote ID of the concurrent object, the process ID of the process in which the concurrent object exists, and the communication of the node in which the process exists. Get the media and its network address.

When the internal concurrent object is called from another process, the reception-time reference resolving unit 36 refers to the local reference table 37 and obtains the local ID from the remote ID included in the message of the call. . In addition, the reception-time reference resolving unit 36 determines that the input message is
If it is to be transferred to a concurrent object of another process, the remote ID of the transfer destination is obtained. In addition,
The processing in the reception reference resolving unit 36 is described later with reference to FIG.
Is shown in

Message passing operator 25
3: The message passing operator 253 sends a message input from the application programming function 252 to the message sender (M).
essage Sender) 256 and Reference Resolver 2
Communication media interface 2
A message dispatcher (Message Dispatcher) 258 that performs pre-processing of processing in the application programming function 252, such as converting a remote ID of a destination object included in a message received from the network 8 via the network 55 into a local ID. The process of the message dispatcher 258 is shown in FIG. 32 described later.

Message evaluation scheduler 257: When a plurality of messages arrive from the network 8 via the communication media interface 255, the message evaluation scheduler 257 determines the processing order of the plurality of messages.

Communication media interface 255: The communication media interface 255 uses the destination information (remote ID and network address) input from the message sender 256 and the argument data converted into a common expression via the network 8 A message which is a byte string such as a packet in a format suitable for output is generated, and the generated message is output to the network 8. Also, the communication media interface 255
Outputs a message consisting of a byte string received via the network 8 to the message dispatcher 258.

[Overview of Management Process] In the distributed processing system 100, the management processes 12 ₁ , 12 ₁ are set in advance in order to distribute the processes to one or more computer nodes.
₂ and 12 ₃ are activated. Management processes 12 ₁ , 12
_2, 12 ₃ is a kind of process, the process ID of the fixed assigned. The management process of each computer node is based on the given process allocation request,
The general processes described above are arranged and replaced in cooperation with each other. Here, as shown in FIG.
The management process 12 ₃ which is given the process replacement start request 70 indicating that the process 16 ₃ of “nB” is to be replaced.
Is called a master. In addition, the process generation request 7
The management process 12 ₂ for receiving a 1 is referred to as a slave. Hereinafter, it referred to the management process 12 ₃ and the master management process 12 _3, the management process 12 ₂ slave management process 12
Call it ₂ .

[0065] [Management Process Details] Fig 5 is a block diagram of a master management process 12 _3. The management process 12 ₁ and the slave management process 12 ₂ has the same configuration as the master management process 12 _3. As shown in FIG. 5, the master management process 12 ₃ has a message passing function 51, the process placement function 52, the process generating function 53 and the process arrangement table 54.

The message passing function 51 is basically the same as the message passing function of the general process shown in FIG.

The process placement function 52 receives the process replacement start request 70 shown in FIG. 9, for example, starts the process placement processing, and operates on the master management process. The process placement function 52 includes a process placement unit 55, a process information setting unit 56,
The process information notification unit 57 includes a process arrangement table 54.
To access.

The process placement unit 55 inputs, for example, a process replacement start request 70 shown in FIG.
Is determined in the process management table 58 and the node management table 59, and if they are not registered, their entries are added. The process of the process placement unit 55 is shown in FIG. 10 described later.

[0069] process information setting unit 56, for example, from a new process 16 _4, enter the process information setting request 73 shown in FIG. 16, the process management table 58 and the node management table 59 shown in FIG. 6, existing processes 16 ₃ The entry corresponding to the process name “YonB” of the new process 16
₄ to rewrite the contents shown in FIG. 19 corresponding. Further, process information setting unit 56, the new process 16 _4, FIG. 1
And outputs the process information setting request response 74 including process information for an existing process 16 ₃ shown in 8. The process of the process information setting unit 56 is shown in FIG. 17 described later.

[0070] Process information notification unit 57 is, for example, from a new process 16 _4, enter the process information notification request containing a list of process name as a key process name included in the request, the process management table 58 and node management Referring to Table 59, process I corresponding to the process name
D, node name, to obtain process information and a network address and communication media, and outputs the process information in the new process 16 _4. The processing process information notifying unit 57 is shown in FIG. 23 to be described later, for example, import processing unit of the new process 16 ₄ shown in FIG. 20 32
Is called in the import process (step S52).

[0071] The process generation function 53, for example, operates on the slave management process 12 ₂ performs the processing shown in FIG. 12 from the master management process 12 ₃ enter process generation request 71 shown in FIG. 11.

The process allocation table 54 is accessed from the process allocation function 52 and includes a process management table 58 and a node management table 59. Figure 6 is a diagram for explaining the process management table 58 of the master management process 12 ₃ shown in FIG. As shown in FIG. 6, the process management table 58
It manages process names, process IDs, and computer node names where the processes are located. Also,
The node management table 59 includes a node name of a computer node on which a process is running, a network address of the computer node, and a communication medium.

Hereinafter, the operation of the distributed processing system 100 will be described. Here operation until generating a new process performed in the distributed processing system 100 shown in FIG. 1, when an existing process 16 ₃ for sales tax 3% of the processing is in operation, the consumption tax 5% of the treated new The process 16 ₄ is generated, and a request to replace the existing process 16 ₃ with the new process 16 ₄ at 00: 00: 0 on April 1 is sent to the process replacement request before 00:00:04 on April 1. to be described the operation when the master management process 12 ₃ has accepted.

FIG. 7 shows the distributed processing system 100 described above.
FIG. 7 is a diagram for explaining a request and information flow in the operation of FIG. FIG. 8 is a flowchart for explaining the operation of the distributed processing system 100 described above. Outline of Operations Up to Generation of New Process Hereinafter, an outline of operations of the distributed processing system 100 will be described mainly with reference to FIGS. 7 and 8. Step S1: The process replacement start request 70 containing the information shown in FIG. 9 is input to the process placement portion 55 of the process arrangement features 52 of the master management process 12 _3. Step S2: The process shown in FIG. 10 is performed in the process placement unit 55, and a process generation request 71 including the information shown in FIG. 11 is generated. Step S3: Process generation request 71 generated in step S2 is a process placement unit 55 is output to the process generating function 53 of the slave management process 12 _2.

[0075] Step S4, S5: In the process generating function 53 of the slave management process 12 _2, the processing shown in FIG. 12 is performed, the generation process of the new process 16 ₄ 7
2 is executed. The new process 16 _4, a process is provided information shown in FIG. 13.

[0076] Step S6: In the novel process 16 _4, the initial processing shown in FIG. 14 is performed.

Details of Operation Up to Generation of New Process Hereinafter, each process shown in FIG. 8 will be described in detail. [Processing in the master management process 12 ₃ Process placement section 55 (Step S2)] FIG. 10, the master management process 1
2 is a flowchart of a process of the _third process placement portion 55. Step S11: When receiving the process replacement start request 70 shown in FIG. 9, the process placement unit 55 checks whether the process placement node “nodeB” included in the request has been registered in the node management table 59 shown in FIGS. Determine whether or not.

Step S12: The process placement unit 55
If it is determined in step S11 that the request has not been registered, based on the process replacement start request 70, the process management node name “n” included in the request is stored in the node management table 59.
odeB ”entry. In the present embodiment, the process placement node “nodeB” is already registered in the node management table 59 shown in FIG.
Is not performed.

Step S13: The process placement unit 55
It is determined whether or not the process name included in the process replacement start request 70 is registered in the process management table 58. Step S14: The process placement unit 55 determines in step S1
If it is determined in step 3 that the process has not been registered, an entry of the process name included in the process replacement start request 70 is added to the process management table 58. At this time, the process ID of the entry of the added process is set to 0. In the present embodiment, the process name “YonB” included in the process replacement start request 70 shown in FIG.
No entry is added because it is already registered in the process management table 58.

Step S15: The process placement section 55
Node name to place the process on the computer node 10 ₂ of the slave management process 12 _second process generating function 53 of "nodeB", and outputs the process generation request 71 containing the information shown in FIG. 11 (step S shown in FIG. 8
3).

[0081] [process of the slave management process 12 _second process generating function 53 (step S4)] FIG. 12 is a flowchart of a process of process generation function 53 of the slave management process 12 _2. Step S21: When the process generation function 53 receives the process generation request 71 shown in FIG. 11, the process module 14 whose location is specified by the process generation request 71
₂ Download “Y-1”. As the download protocol, “ftp” or “http” is used.

Step S22: Process generation function 53
Determines the process ID to be assigned to the new process 16 ₄ to be generated. As the process ID, a numerical value other than 0 is assigned so as to be unique within the computer node. In the present embodiment, the new process 16
₄ is assigned “21” as a process ID.

Step S23: Process generation function 53
Generates a new process 16 ₄ executes a program module 14 ₂ downloaded in step S21 (step S5 shown in FIG. 8). The new process 16 _4, a process is provided information shown in FIG. 13.

[0084] [initial processing of new process 16 ₄ (step S6)] new process 16 ₄ Once generated by the slave management process 12 ₂ performs the initial processing described below. Figure 14 is a flow chart of initial processing of the new process 16 _4. Step S31: new process 16 _4, in the initialization process, calls the export processing unit 31 of the reference Risoruba 254 shown in FIG. 4, for concurrent object of its own process, performs the export processing, FIG. 1
The local reference table 37 shown in FIG. This allows
The concurrent object New Process 16 ₄ becomes visible to other processes. In the export processing, the parallel objects existing in a new process 16 _4, allocation so that a unique remote ID and the local ID, creates a local reference table 37. In the export process, the message transfer condition in the local reference table 37 and the reference to the entry of the transfer destination parallel object in the remote reference table 38 are not set and are left blank. In addition, by performing the export processing, the remote ID notifying unit 33 illustrated in FIG. 4 can operate.

Step S32: The new process 16 ₄
Process information setting request 73 which includes process information of the new process 16 ₄ shown in FIG. 16, the master management process 12
The process output is output to the process information setting unit 56 of the process arrangement function 52 of No. ₃ . Then, in the process information setting unit 56,
The process shown in FIG. 17 is performed, the process information setting request response 74 including process information for an existing process 16 ₃ is produced. Further, by the process shown in FIG. 17 is performed, the process management table 58 shown in FIG. 5 of the master management process 12 ₃
The node management table 59 is rewritten from FIG. 6 to FIG.

Step S33: The new process 16 ₄
From the process information setting unit 56, enter process information setting request response 74 including process information for an existing process 16 ₃ shown in Figure 18.

Step S34: The new process 16 ₄
Process information setting request response 7 input in step S33
4 is temporarily saved.

Step S35: The new process 16 ₄
The import processing unit 32 of the reference resolver 254 shown in FIG. 4 is called, and the import processing shown in FIG. 20 is performed on the parallel object existing in another process. The import processing shown in FIG. 20 will be described later in detail. Import process shown in FIG. 20, in step S52, the process information notification request 75 containing a list of process name is output to the process information notification unit 57 of the process arrangement features 52 of the master management process 12 _3. Then, the process shown in FIG. 23 is performed in the process information notification unit 57, and process information including a process ID, a node name, a network address, and a communication medium corresponding to the process name included in the process information notification request 75 is obtained. A process information notification request response 76 including the process information is input.

The processing shown in FIG. 23 will be described later in detail. By import processing, new process 16 ₄
The remote reference table 38 and the process reference table 39 are updated as shown in FIG. 21 to FIG. When the processing of the step S35 is completed, the reference table of the new process 16 ₄ is as shown in FIG. 24, the existing new process 1
6 ₃ lookup table is as shown in FIG. 25.

Step S36: The new process 16 ₄
After the end of the import process, it is determined whether or not the process ID of the saved existing process information is 0. New Process 16 _4, if the process ID shunting the existing process information is 0, it means that existing processes do not exist, the initial process is completed.

Step S37: The new process 16 ₄
If the process ID of the saved process information is other than 0, the existing process exists, and the reference resolver 25 shown in FIG.
A message transfer information setting request 79 is output to the transfer information setting unit 34 of FIG. In the case of FIG. 1, as shown in FIG. 18, the process ID of the process of the existing processes 16 ₃ retracted is because "20", with respect to existing process 16 _3, outputs a message transfer information setting request 79 shown in FIG. 26 I do. Existing processes 16 ₃ inputs the message transfer information setting request 79 shown in FIG. 26 performs the processing shown in FIG. 27 in the transfer information setting section 34 shown in FIG. 4, the remote reference table 38, process reference table 39 and a node reference An entry is added to the table 40 and the local reference table 37 is updated. In addition, the transfer information setting unit 34
₃ local reference table 37 after the updating of the outputs to the new process 16 ₄ included in the message transfer information setting request response 80.

Step S38: The new process 16 ₄
A message transfer information setting request response 80 is input. When the step S38 is completed, the lookup table is an existing process 16 _3, are as shown in FIG. 28.

Step S39: The new process 16 ₄
Reference resolver 25 of own process shown in FIG.
A message transfer information setting request 79a including the information shown in FIG. 29 is output to the transfer information setting unit 34 of FIG. As a result, the transfer information setting unit 34 of the new process 16 ₄
Performs the processing shown in 7, lookup table New Process 16 ₄ is updated as shown in FIG. 30.

[0094] processing [master management process of the process 12 ₃ Process information setting section 56] the process information setting unit 56, for example, in the initialization process of the new process 16 ₄ shown in FIG. 14, a new process 16 _4, FIG. This is performed by inputting the process information setting request 73 shown in FIG. When the process of the process information setting unit 56 is completed, the process management table 58 and the node management table 59 of the master management process 12 ₃ is updated as shown in FIG. 19 from FIG. Figure 17 is a flowchart of a process of process information setting unit 56 of the master management process 12 _3. Step S41: process information setting unit 56, the process information setting indicating a new process 16 ₄ 16 request 73
Is entered, and the process name “Yon
An entry corresponding to "B" is retrieved from the process management table 58 and the node management table 59 shown in FIG.

Step S42: Process information setting section 56
Saves the process information set in the entry searched in step S41.

Step S43: Process information setting section 56
Is the new process 1 included in the process information setting request 73 shown in FIG.
6. Set the process information of ₄ . As a result, the process management table 58 and the node management table 59 are
Updated as shown in

Step S44: Process information setting section 56
, In step S63 shown in FIG. 23 to be described later, process information notification unit 57 of the master management process 12 _3,
It is determined whether a process ID is waiting. Then, the process information setting unit 56 sets the process ID
If there is a wait, the process of step S45 is executed, and if not, the process of step S46 is executed.

Step S45: Process information setting section 56
Is in the process information notification unit 57 of the master management process 12 ₃ that the waiting process ID, process I
Notifies that D has been set.

Step S46: Process information setting section 56
May process information setting request response 74 including process information for an existing process 16 ₃ shown in FIG. 18 that has been retracted in step S42, and outputs the new process 16 _4. Thereafter, in the new process 16 ₄ , step S shown in FIG.
33 is performed.

[0100] [Process of import processing unit 32 of the new process 16 ₄ Reference Risoruba 254 (FIG. 14
FIG. 20 is a flowchart of the processing of the import processing unit 32.

[0102] Step S51: import processing unit 32 of the new process 16 ₄ with respect concurrent object other processes required in its own process, to create a remote reference table 38 and the process reference table 39. At this time, the reference to the remote ID and the process reference table 39 in the remote reference table 38 and the reference to the process ID and the node reference table 40 in the process reference table 39 are left blank. For example, when the process of step S51 ends, the new process 16
₄ of the remote reference table 38 and the process reference table 39,
It is as shown in FIG.

Step S52: The new process 16 ₄
Process information in the master management process 12 ₃ notification section 57
, A process information notification request 75 including a list of necessary process names is output. Specifically, the new process 16
₄ are the process names “XonA” and “X
A process information notification request 75 including “onC” is output to the process information notification unit 57.

Step S53: The new process 16 ₄
The master management process 12 ₃ Process information notification unit 57, and inputs the process information notification request response 76 containing process information about the process name included in the process information notification request 75. Specifically, the new process 16 ₄
A process information notification request response 76 including process information including the process ID, the node name, the network address, and the communication medium for the processes 16 ₁ and 16 ₂ having the process names “XonA” and “XonC” is input.

Step S54: The new process 16 ₄
Process information notification request response 7 input in step S53
6, the node reference table 40 is used.
Create

Step S55: The new process 16 ₄
Process information notification request response 7 input in step S53
Process reference table 3 using the process information contained in
The process ID and the item of reference to the node reference table 40 which are blank in 9 are set.

Step S56: The new process 16 ₄
A remote ID notification request 7 including a list of concurrent object names requiring a remote ID to the remote ID notifying unit 33 of the reference resolver 254 in another process.
7 is output. Specifically, the remote ID notification request process 16 ₁ Remote ID notification unit 33 including the "objA" Concurrent Object names shown in FIG. 21, also,
Remote ID notification request including concurrent object names shown in FIG. 21 to "objB" outputs to the process 16 ₂ Remote ID notification unit 33. Process 16 ₁ and 16
_When the remote ID notifying unit 33 receives the remote ID notifying request 77, the remote ID notifying unit 33 obtains the remote ID of the concurrent object name included in the request with reference to the local reference table 37, and transmits the remote ID to the new process 16 Notify ₄ . Step S57: new process 16 _4, the remote ID notification unit 33 of the reference Risoruba 254 in other processes, inputs the remote ID. Specifically, the remote IDs of “objA” and “objB” are input from the remote ID notifying units 33 of the processes 16 ₁ and 16 ₂ .

Step S58: The new process 16 ₄
When a remote ID is input from another process, the remote ID is set in the remote reference table 38. Thereby, the reference resolver 2 of the new process 16 ₄
When the process of step S58 is completed, the remote reference table 38 and the process reference table 39 are stored in FIG.
22 is updated as shown in FIG. 22, and the node reference table 40 shown in FIG. 22 is generated.

[0108] process of process information notification unit 57 of [the master management process 12 processes the _third process information notification unit 57] The master management process 12 _3, steps S shown in FIG. 20
In 52, the new process 16 _4, the process information notification unit 57 of the master management process 12 ₃ is executed when the process information notification request 75 is output. Figure 23 is a flowchart of a process of process information notification unit 57 of the master management process 12 _3.

Step S61: Master management process 12
Process information notification unit 57 of the _3, in step S52 shown in FIG. 20, the import processing unit 32 of the new process 16 ₄
From process information notification request 7 including a list of process names
When "5" is input, an entry of the process management table 58 shown in FIG. 19 corresponding to the process name included in the request is searched. Specifically, as described in step S52 shown in FIG. 20 described above, the process information notifying unit 57 stores the process names “XonA” and “XonC” shown in FIG.
Is input.

Step S62: process information notifying section 57
Determines whether the process ID of the entry in the process management table 58 searched in step S61 is 0, and if 0, performs the process in step S63;
64 is performed.

Step S63: Executed when the process ID is determined to be 0 in step S62,
Is not set, the process information notification unit 57 is in a waiting state until the process ID is set.

Step S64: process information notifying section 57
Obtains the process information of the entry searched in step S61. The process information includes a process ID, a node name, a network address, and a communication medium.

Step S65: process information notifying section 57
Determines whether or not the process information has been generated for all the process names included in the process information notification request 75. If it has, the process of step S66 is executed, and it is determined that the process information has not been generated. If so, the process of step S61 is repeated. Specifically, the process information notifying unit 57 determines that the process names “XonA” and “XonA”
Generate process information for processes 16 ₁ and 16 ₂ having "C".

Step S66: process information notifying section 57
Generates process information notification request response 76 containing the process information generated by step S65, and outputs it to the new process 16 _4. Thereafter, in the new process 16 ₄ , the processing after step S53 shown in FIG. 20 is performed.

[Transfer information setting unit 3 of existing process 16 ₃ ]
4 of Process] process of the transfer information setting unit 34 of the existing process 16 _3, for example, in step S37 shown in FIG. 14, a new process 16 _4, 26 to the transfer information setting unit 34 of the existing processes 16 ₃ It is executed when the message transfer information setting request 79 shown is output.

FIG. 27 is a flowchart of the processing of the transfer information setting unit 34. Hereinafter, with reference to FIG. 27, illustrating the process of transferring information setting unit 34 of the existing process 16 _3.

[0117] Step S71: transfer information setting unit 34 of the existing process 16 _3, in step S37 shown in FIG. 14, to enter the message transfer information setting request 79 shown in FIG. 26 the new process 16 ₄ is output. As shown in FIG. 26, the message transfer information setting request 79 includes a local reference table 37 shown in FIG. 15 New Process 16 _4. The transfer information setting unit 34 sets the new process 16 ₄
A local reference table 37 shown in FIG. 15, the local reference table 3 shown in existing processes 16 ₃ in FIG. 25 is a self-process
7 to determine whether the same concurrent object name exists. Comparing FIG. 15 with FIG. 25, since the parallel object name “Ftax” is the same, the transfer information setting unit 34 performs the process of step S72.

Step S72: The transfer information setting unit 34
Figure New Process 16 ₄ is a transfer destination process information contained in the message transfer information setting request 79 shown in FIG. 26 1
Based on the process information shown in 6, a computer node that the new process 16 ₄ are disposed 10 ₂ (address "s
It is determined whether an entry of “2.dvl.co.jp”, communication medium “Ethernet”, and node name “nodeB”) is registered in the node reference table 40 shown in FIG. And
Since it has not been registered, the process of step S73 is executed.

Step S73: The transfer information setting unit 34
New Process 16 ₄ computer node is located 10 ₂ of node name entry "nodeB", 2
5 is registered in the node reference table 40 shown in FIG. As a result, the node reference table 40 becomes as shown in FIG.

Step S74: The transfer information setting section 34
Process process name "YonB" shown in the process information of the new process 16 ₄ shown in FIG. 16, it is determined whether or not it is registered in the process reference table 39 shown in FIG. 25. Then, since it is not registered, the process of step S75 is executed.

Step S75: The transfer information setting section 34
The entry process name "YonB" shown in the process information of the new process 16 _4, registered in the process reference table 39 shown in FIG. 25. Thereby, the process reference table 39
Is as shown in FIG.

Step S76: The transfer information setting section 34
Entries in step S71 corresponds to the parallel object name of the new process 16 ₄ local reference table 37 shown in FIG. 15 it is determined that the same name "Ftax" is the remote reference table 38 of the existing processes 16 ₃ shown in FIG. 25 It is determined whether or not it has been registered. Then, since it has not been registered, the process of step S77 is executed.

Step S77: The transfer information setting unit 34
The local reference table 37 of the new process 16 ₄ shown in FIG.
, The remote ID “211” registered in the entry of the concurrent object name “Ftax” is extracted.

Step S78: The transfer information setting section 34
By using the remote ID "211" extracted in step S77, the in remote reference table 38 of the existing processes 16 ₃ shown in FIG. 25, as shown in FIG. 28, the entry of concurrent object name "Ftax" and remote ID "211" Add.

Step S79: The transfer information setting section 34
It is determined whether or not the request source process information included in the message transfer information setting request 79 is the process information of the own process. Then, it is determined that the information is process information of another process, and the process of step S81 is executed.

Step S81: The transfer information setting section 34
Step S in the local reference table 37 shown in FIG.
71, the parallel object name "F
The message transfer condition of the message transfer information setting request 79 shown in FIG. 26 is set in the message transfer condition of the entry of “tax”, and the parallel object name “F” added in step S78 is added to the reference of the entry of the remote reference table 38.
Set a reference to the "tax" entry. This allows
The local reference table 37 shown in FIG. 25 is updated as shown in FIG.

Step S82: The transfer information setting section 34
The local reference table 37 shown in FIG.
New process including message transfer information setting request response 80
16 _FourOutput to

[Transfer information setting section 3 of new process 16 ₄ ]
For 4 treatment] new process 16 _fourth transfer information setting unit 34, for example, in step S39 shown in FIG. 14, message transfer a new process 16 ₄ shown in FIG. 29 with respect to transfer information setting unit 34 of its own process This is executed when the information setting request 79a is output.

[0129] Figure 27 is a flowchart of a process of transferring information setting unit 34 of the new process 16 _4. Hereinafter, FIG.
Referring to 7, illustrating a process of transferring information setting unit 34 of the new process 16 _4.

[0130] Step S71: transfer information setting unit 34 of the new process 16 _4, in step S39 shown in FIG. 14, to enter the message transfer information setting request 79a shown in FIG. 29 by the own process is output. As shown in FIG. 29, the message transfer information setting request 79 includes a local reference table 37 shown in existing processes 16 ₃ in FIG. 28. Transfer information setting unit 34, as compared to the local reference table 37 shown in existing processes 16 ₃ of FIG. 28, the local reference table 37 shown in FIG. 24 New Process 16 ₄ is its own process, the same concurrent object name It is determined whether or not exists. Comparing FIG. 28 with FIG. 24, since the parallel object name “Ftax” is the same, the transfer information setting unit 34 performs the process of step S72.

Step S72: The transfer information setting section 34
Based on the process information shown in FIG. 18 of the existing processes 16 ₃ message is a transfer destination process information included in the transfer information setting request 79a shown in FIG. 29, the existing process 16 ₃ computer node is located 10 ₂ (address " It is determined whether an entry of “s2.dvl.co.jp”, communication medium “Ethernet”, node name “nodeB”) is registered in the node reference table 40 shown in FIG. Then, since it is not registered, the process of step S73 is executed.

Step S73: The transfer information setting section 34
Computer node 10 ₂ of node names on the basis of the process information of the existing processes 16 ₃ shown in FIG. 18 the entry "nodeB", and registers the node reference table 40 shown in FIG. 24. As a result, the node reference table 40 becomes as shown in FIG.

Step S74: The transfer information setting section 34
Process process name "YonB" shown in the process information of the existing processes 16 ₃ shown in FIG. 18, it is determined whether or not it is registered in the process reference table 39 shown in FIG. 24. Then, since it is not registered, the process of step S75 is executed.

Step S75: The transfer information setting section 34
Process name indicated in the process information of the existing processes 16 ₃ shown in FIG. 18 the entry "YonB", registered in the process reference table 39 shown in FIG. 24. As a result, the process reference table 39 becomes as shown in FIG.

Step S76: The transfer information setting section 34
Concurrent object name of the local reference table 37 of the existing processes 16 ₃ shown in Figure 28 it is determined that the same name in step S71 the entry corresponding to "Ftax" is the remote reference table 38 of the new process 16 ₄ shown in FIG. 24 It is determined whether or not it has been registered. Then, since it has not been registered, the process of step S77 is executed.

Step S77: The transfer information setting unit 34
Local Reference Table existing processes 16 ₃ shown in FIG. 28 37
, The remote ID “201” registered in the entry of the parallel object name “Ftax” is extracted.

Step S78: The transfer information setting section 34
By using the remote ID "201" extracted in step S77, the in remote reference table 38 of the new process 16 ₄ shown in FIG. 24, as shown in FIG. 30, the entry of concurrent object name "Ftax" and remote ID "210" Add.

Step S79: The transfer information setting section 34
It is determined whether or not the request source process information included in the message transfer information setting request 79 is the process information of the own process. Then, it is determined that the information is the process information of its own process, and the process of step S80 is executed.

Step S80: The transfer information setting section 34
The message transfer condition included in the message transfer information setting request 79a shown in FIG. 29 is inverted.

Step S81: The transfer information setting section 34
Step S in the local reference table 37 shown in FIG.
71, the parallel object name "F
The message transfer condition of the entry of “tax” is set to the message transfer condition inverted in step S80, and the reference to the entry of the remote reference table 38 is referred to the entry of the parallel object name “Ftax” added in step S78. Set. Thus, the local reference table 37 shown in FIG. 24 is updated as shown in FIG.

Effect of New Process Generation Through the above processing, the computer node 1 shown in FIG.
0 _2, a new process 16 ₄ of the process name "YonB"
Is generated. Further, the computer node 10 _2,
Existing processes 16 ₃ process name "YonB" exists to continue. Further, the processing of the new process 16 ₄ shown in FIG. 14, reference tables for reference Risoruba 254 New Process 16 ₄ is as shown in FIG. 30, reference table of existing processes 16 ₃ Reference Risoruba 254 28 It becomes as shown in. Incidentally, 16 _1, 16 ₂ of the reference table process shown in Figure 1 is the same as the previous generation of new process 16 _4.

[0142] New processes after generation of the operation (Part 1) below, on April 1, 0:00 for example March 31 23:00 50 minutes 00 seconds earlier than 0 seconds, the computer node 10 ₁ on shown in FIG. 1 Process 16 ₁ concurrent object "obj
The operation performed when “A” calls the concurrent object “Ftax” will be described. FIG. 31 shows that at 23:55:00 on March 31, the computer node 10 ₁ shown in FIG.
Concurrent object "objA" of process 16 ₁ above
Is a diagram for describing an operation when a concurrent object “Ftax” is called. FIG.

As shown in FIG. 31, in the process 16 ₁ , the name of the concurrent object to be called “Ftax”
The message transmission request including the message and the argument is output from the concurrent object “objA” to the application programming function 252. And the application
In the programming function 252, the argument included in the message transmission request is converted from the internal expression to the common expression.

Next, in the message sender 256 of the process 16 ₁ , the transmission reference resolution unit 35 of the reference resolver 254 is called, and the parallel object name “Ftax” included in the message transmission request is used as a key in FIG. The remote reference table 38 shown is referred to, and the destination information (remote ID, process ID, and network address) of the message is obtained. next,
Communication media interface 255 for process 16 ₁
, A message which is a byte sequence such as a packet in a format suitable for output via the network 8 is generated from the destination information input from the message sender 256 and the argument data converted into the common expression, Is output to the network 8. The message has a transmission time “23:50 on March 31” as attribute information.
Minutes 00 seconds "is added.

[0145] The message over the network 8, are received in the communication media interface 255 existing processes 16 _third computer node 10 _2, 3 of the message dispatcher 258 will be described later
Step 2 is performed. In the processing shown in FIG. 32, the local reference table 37 shown in FIG. 28 is referred to determine whether or not the attribute information (transmission time) included in the received message satisfies the message transfer condition of the entry of the concurrent object “Ftax”. Since the message transfer condition is not satisfied, the application programming function 252 is instructed to execute the concurrent object “Ftax”. At this time, the reference resolver 254 refers to the local reference table 37, and obtains the local ID from the remote ID included in the message.

The existing process 16_ThreeApplication
Message programming function 252
Argument data included in a page is converted from a common expression to an internal expression
And the converted argument and the reference resolver
-The execution account obtained from the local ID obtained in 254
A parallel object that processes 3% of the consumption tax using a dress
The project “Ftax” is executed. Then, the parallel
The execution result of the object "Ftax" is the existing process 1.
6 _ThreeApplication programming function 252,
Message Sender 256 and Communication Media Interface
After processing by face 255,
Roses 16₁Sent to.

[0147] Hereinafter, a message dispatcher 258 existing processes 16 ₃ in the above-described operation (Part 1)
Will be described in detail. FIG. 32 is a flowchart of a process when the message dispatcher 258 receives a message.

[0148] Step S91: Message Dispatcher 258 existing processes 16 ₃ o'clock communication media remote ID "201" as the interface 255 is included in the message received from the process 16 ₁ and attribute information (transmission time "March 31, 23 50:00 "), and outputs the reference resolution request to the reference resolution unit 36 at the reception of the reference resolver 254 to perform reference resolution. Reference resolver at the time of reception of reference resolver 254
Performs a process shown in FIG. 33 upon input of the reference resolution request, refers to the local reference table 37 shown in FIG. 28, and does not satisfy the message transfer condition.
The local reference to the concurrent object in the process including the local ID “20201” corresponding to “1” is output to the message dispatcher 258 as a reference result.

Step S92: The message dispatcher 258 inputs a reference result from the reception reference resolving unit 36 of the reference resolver 254.

Step S93: The message dispatcher 258 determines that the reference result input in step S92 is
It is determined that the reference is a local reference to a concurrent object in its own process, and the process of step S94 is executed.

Step S94: The message dispatcher 258 outputs the local ID included in the local reference input in step S92 and the argument included in the message to the application programming function 252.

In the following, the processing of the reception time reference resolving unit 36 called by the reference resolver 254 in step S91 shown in FIG. 32 in the above-described operation (part 1) will be described. FIG.
6 is a flowchart of a process of No. 6; Step S101: When receiving a reference resolution request including a remote ID and a transmission time from the message dispatcher 258, the reception-time reference resolution unit 36 searches the local reference table 37 shown in FIG. 28 using the remote ID "201" as a key. .

Step S102: Reception reference resolution section 36
Determines that the remote reference of the transfer destination of the entry of the remote ID “201” of the local reference table 37 shown in FIG. 28 retrieved in step S101 is set, and executes the processing of step S103.

Step S103: Reception-time reference solving section 36
Is the remote ID in the local reference table 37 shown in FIG.
The message transfer condition “after 00:00:04 on April 1” of the entry of “201” is compared with the transmission time “23:55:00 on March 31” included in the reference resolution request. It is determined that the message transfer condition is not satisfied, and the process of step S105 is executed.

Step S105: Reception Reference Resolution Unit 36
Responds to the message dispatcher 258 in FIG.
Remote ID "201" in the local reference table 37 shown in FIG.
Is output to the message dispatcher 258. As a result, FIG.
Step S92 shown in FIG. 2 is performed.

[0156]Operation after creating a new process (part 2) Hereinafter, for example, 0:00 on April 1 after 0:00:00 on April 1
At 10:00, the computer node 10 shown in FIG.₁
Process 16 above₁Concurrent object "objA"
When the concurrent object "Ftax" is called
The work will be described. Fig. 34 shows April 1 at 0: 10: 0
At 0 seconds, the computer node 10 shown in FIG. ₁Top professional
Seth 16₁Of the concurrent object “objA”
Operation when calling the object "Ftax"
It is a figure for explaining.

As shown in FIG. 34, in process 16 ₁ , the name of the concurrent object to be called “Ftax”
The message transmission request including the message and the argument is output from the concurrent object “objA” to the application programming function 252. And the application
In the programming function 252, the argument included in the message transmission request is converted from the internal expression to the common expression.

Next, in the message sender 256 of the process 16 ₁ , the transmission reference resolving unit 35 of the reference resolver 254 is called, and the parallel object name “Ftax” included in the message transmission request is used as a key in FIG. The remote reference table 38 shown is referred to, and the destination information (remote ID, process ID, and network address) of the message is obtained. next,
Communication media interface 255 for process 16 ₁
, A message which is a byte sequence such as a packet in a format suitable for output via the network 8 is generated from the destination information input from the message sender 256 and the argument data converted into the common expression, Is output to the network 8. In the message, the transmission time “April 1, 0: 10: 0” as attribute information
0 second "is added.

[0159] The message over the network 8, are received in the communication media interface 255 existing processes 16 _third computer node 10 _2, 3 of the message dispatcher 258 will be described later
Step 2 is performed. In the processing shown in FIG. 32, the local reference table 37 shown in FIG. 28 is referred to determine whether or not the attribute information (transmission time) included in the received message satisfies the message transfer condition of the entry of the concurrent object “Ftax”. Then, a message transmission request is output to the message sender 256 to satisfy the message transfer condition. At this time, the reference resolver 25
4, the local reference table 37, the remote reference table 38, the process reference table 39, and the node reference table 40 in FIG.
Is referenced to obtain a remote reference to the destination of the message. The remote reference includes, for example, a remote ID of a parallel object “Ftax” that performs processing of a consumption tax of 5%.
"211", process ID "21", network address "s2.dvl.co.jp" and communication medium "Ethernet"
Is included.

Then, the message dispatcher 25
8 to the message sender 256, the obtained remote reference, and the argument and attribute information (transmission time “April 1, 00:00:00”) included in the received message are included in the message transmission request. Included.

Then, from the message sender 256,
Via a communication medium interface 255 and the network 8, based on the remote reference contained in the message transmission request, the arguments, and the attribute information included in the message transmission request is sent to the new process 16 _4. That is, from the process 16 ₁ to the existing process 16 ₃
Message output is transferred from the existing processes 16 ₃ new process 16 _4.

New process 16_FourThen, the message
Received, the message dispatcher 258 and
And the reference resolver 254
Using the remote ID “211” included in the message as a key
And refer to the local reference table 37 to process the 5% consumption tax.
Local ID “2” of the concurrent object “Ftax” to be performed
1211 ". And the application program
The consumption tax 5% is processed by the ramming function 252.
The concurrent object "Ftax" is executed. And
The execution result of the concurrent object “Ftax” is new
Process 16 _FourApplication programming machines
No. 252, message sender 256 and communication media
A) After processing by the interface 255, the message
Process 16₁Sent to.

[0163] Hereinafter, a message dispatcher 258 existing processes 16 ₃ in the above-described operation (Part 2)
Will be described in detail. FIG. 32 is a flowchart of a process when the message dispatcher 258 receives a message.

[0164] Step S91: Message Dispatcher 258 existing processes 16 ₃ o'clock remote ID "201" and the attribute information (transmission time "April 1 0 contained in the message to the communication media interface 255 has received from the process 16 ₁ A reference resolution request including “10:00” is output to the reception-time reference resolution unit 36 of the reference resolver 254 to perform reference resolution. Upon receiving the reference resolution request, the reception-time reference resolution unit 36 of the reference resolver 254 performs the processing shown in FIG. 33, refers to the local reference table 37 shown in FIG. 28, and satisfies the message transfer condition. The remote reference to the transfer destination parallel object corresponding to “201” is output to the message dispatcher 258 as a reference result.

Step S92: The message dispatcher 258 inputs a reference result from the reference resolver at the time of reception of the reference resolver 254.

Step S93: The message dispatcher 258 determines that the reference result input in step S92 is
It is determined that it is a remote reference, and the process of step S95 is executed.

Step S95: The message dispatcher 258 outputs arguments and attribute information included in the message to the message sender 256 based on the remote reference input in step S92.

Step S94: The message dispatcher 258 outputs an execution request to the destination parallel object.

Hereinafter, the processing of the reception time reference resolving unit 36 called by the reference resolver 254 in step S91 shown in FIG. 32 in the above-described operation (No. 2) will be described. FIG.
6 is a flowchart of a process of No. 6; Step S101: When receiving a reference resolution request including a remote ID and a transmission time from the message dispatcher 258, the reception-time reference resolution unit 36 searches the local reference table 37 shown in FIG. 28 using the remote ID "201" as a key. .

Step S102: Receive time reference solution section 36
Determines that the remote reference of the transfer destination of the entry of the remote ID “201” of the local reference table 37 shown in FIG. 28 retrieved in step S101 is set, and executes the processing of step S103.

Step S103: Reference resolution unit 36 at the time of reception
Is the remote ID in the local reference table 37 shown in FIG.
By comparing the message transfer condition “after April 1, 00:00:00” of the entry “201” with the transmission time “April 1, 00:00:00” included in the reference resolution request, It is determined that the message transfer condition is satisfied, and step S10
4 is executed.

Step S104: Reception time reference solution section 36
Responds to the message dispatcher 258 in FIG.
The remote reference of the local reference table 37 shown in FIG. Thus, the process of step S92 shown in FIG. 32 is performed.

As described above, the distributed processing system 1
According to 00, for example, before 3:00:00 on April 1, 3
Month 31 days 23 hours 50 minutes 00 seconds, the message output from the process 16 ₁ to the existing process 16 _3,
In the existing process 16 ₃ , the parallel object “Ftax” for processing the consumption tax of 3% is executed, and the result is returned to the process 16 ₁ . Also, at 00:00:00 on April 1 after 00:00:04 on April 1, the process 16
The message output from the ₁ to the existing process 16 _3, as shown in FIG. 35, and transferred from the existing processes 16 ₃ new process 16 _4, performed in the new process 16 _4, the consumption tax 5% of the treated concurrently Object "F
Run the tax ", and returns the result to the process 16 _1.

[0174] Therefore, according to the distributed processing system 100, can start a new process 16 ₄ while operating the existing process 16 _3, the processing of the consumption tax in accordance with the output message from the process 16 _1, the message depending on the transmission time, existing processes 16 ₃ and new processes 16 ₄
And can be switched appropriately. Further, according to the distributed processing system 100, when switching from the existing process 16 ₃ to the new process 16 ₄ , the existing process 16 ₃
It is not necessary to stop the _3, it is possible to accept messages from the process 16 ₁ always.

In the distributed processing system 100,
The messages existing processes 16 ₃ receives, at the time when the transmission time runs out April 1st 00:00 0 previous message than seconds, including process 16 _1, existing processes 16
The remote ID of the entry of the concurrent object "Ftax" in the remote reference table 38 of the process having a reference to ₃
To the concurrent object “Fta” of the new process 16 _4.
x ”is rewritten to the remote ID“ 211 ”. And
To stop the existing process 16 _3. Note that existing process 1
6 ₃ may be as it is operated, to stop at the point in order to reduce the load of the computer node 10 _2, the transmission time runs out even messages older than 0 minutes 0 seconds at 1 April 0 Is desirable.

[0176] Further, according to the distributed processing system 100 described above, in the initialization process of the new process 16 ₄ shown in FIG. 14, while communicating with the new process 16 ₄ with the existing process 16 _3, new processes 16 ₄ and local reference table 37 of the existing process 16 _3, the remote reference table 38, to update and generate process reference table 39 and a node reference table 40, can reduce the burden of the master management process 12 ₃ and slave management process 12 _3.

Further, according to the distributed processing system 100,
In the local reference table 37, the remote reference table 38, the process reference table 39, and the node reference table 40, even if there are unset items, the items of the already set items can be used, so the items are set. Each time, the processing is performed by sequentially using the data of the set items, thereby enabling high-speed processing.

The present invention is not limited to the above embodiment. For example, in the distributed processing system 100 described above, the message existing processes 16 ₃ receives, while the transmission time is present before the message than 0 minutes 0 seconds at April 1 0, including the process 16 _1, remote ID entry of the reference process concurrent object "Ftax" remote reference table 38 with the existing process 16 _3, it is also possible to rewrite the remote ID "211" of concurrent object new process 16 ₄ "Ftax" is there. In this case, a message whose transmission time is earlier than April 1, 00:00:00 is a new process 1 as shown in FIG.
After being received by 6 _4, is transferred to an existing process 16 _3, consumption tax 3% processing according to the message is performed in the existing process 16 _3. Also in this case, the transmission time is 4
It is desirable to stop processing the existing processes 16 ₃ when the moon day 00:00:00 previous message than is exhausted.

The distributed processing system 100 does not transfer the message input from the process 16 ₁ , for example, between the existing process 16 ₃ and the new process 16 ₄ , as shown in FIG. the message transfer process 16 ₅ to collectively input a message from ₁ may be provided. In this case, the message transfer process 16 _5, when the transmission time included in the message input from the process 16 ₁ is earlier than 0 minutes 0 seconds at 1 April 0, existing the message process 1
Transferred to 6 _3, if it is 0 minutes 0 seconds after 4:00 January 0 outputs the message to the new process 16 _4.

In the above-described embodiment, the transmission time of the message is used as the attribute information included in the message. However, the attribute information may include other information such as the address of the process that transmits the message. It may be.

[0181] In the embodiment described above, a new process 16 ₄ to switch an existing process 16 _3, but existing processes 16 ₃ is generated in the computer node 10 ₂ operating, the other computer nodes, such as the computer node 10 ₁ May be generated. Further, the process of switching the existing processes 16 ₃ other but newly generated, may be one that already exists.

[0182] In addition, the new process 16 _4, a message transferred from an existing process 16 _3, based on the predetermined message transfer condition may be further transferred to other processes.

[0183]

As described above, according to the distributed processing system and the method thereof of the present invention, when a running process is switched to another process, an appropriate process according to the switching condition before and after the switching is performed. Can handle the message. Further, according to the distributed processing system and method of the present invention, a running process can be switched to another process without stopping the operation of the system.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a distributed processing system according to an embodiment of the present invention.

FIG. 2 is a diagram illustrating an outline of an internal configuration of a computer node of the distributed processing system illustrated in FIG. 1;

FIG. 3 is a diagram for explaining a message passing function of the process shown in FIG. 1;

FIG. 4 is a configuration diagram of a reference resolver shown in FIG. 3;

FIG. 5 is a configuration diagram of a master management process shown in FIG. 1;

FIG. 6 is a diagram for explaining the process management table and the node management table shown in FIG. 5 before performing the processing shown in FIG. 17;

FIG. 7 is a diagram for explaining a flow of a request and information in the operation of the distributed processing system shown in FIG. 1;

FIG. 8 is a flowchart for explaining the operation of the distributed processing system shown in FIG. 1;

FIG. 9 is a diagram for explaining the process replacement start request shown in FIG. 7;

FIG. 10 is a flowchart of a process of a process arrangement unit of the management process shown in FIG. 5;

FIG. 11 is a diagram for explaining the process generation request shown in FIG. 7;

FIG. 12 is a flowchart of a process of a process generation function of the management process shown in FIG. 5;

FIG. 13 shows process information of a new (generation) process.

FIG. 14 is a flowchart of an initial process of a new process.

FIG. 15 is a diagram for explaining a local reference table of a new process after performing the export process shown in FIG. 14;

FIG. 16 is a diagram for explaining process information of a new process included in the process information setting request shown in FIG. 7;

FIG. 17 is a flowchart of a process performed by a process information setting unit of the management process shown in FIG. 5;

FIG. 18 is a diagram for explaining process information of an existing process included in the process information setting request response shown in FIG. 7;

FIG. 19 is a diagram for explaining the process management table and the node management table shown in FIG. 5 after performing the processing shown in FIG. 17;

FIG. 20 is a flowchart of a process of an import processing unit of the reference resolver shown in FIG. 4;

FIG. 21 is a diagram for explaining a remote reference table and a process reference table of a new process after performing the processing in step S51 shown in FIG. 20;

FIG. 22 is a diagram for explaining a remote reference table and a process reference table of a new process after performing the import processing shown in FIG. 14;

FIG. 23 is a flowchart of a process of a process information notification unit of the management process shown in FIG. 5;

FIG. 24 is a diagram for explaining a local reference table, a remote reference table, a process reference table, and a node reference table of a new process after performing the import processing shown in FIG. 14;

FIG. 25 is a diagram for explaining a local reference table, a remote reference table, a process reference table, and a node reference table of an existing process after performing the import processing shown in FIG. 14;

FIG. 26 is a diagram for explaining the message transfer information setting request shown in FIG. 7;

FIG. 27 is a flowchart of a process of a transfer information setting unit shown in FIG. 4;

FIG. 28 is a diagram for explaining a local reference table, a remote reference table, a process reference table, and a node reference table of an existing process after performing the process of step S38 shown in FIG. 14;

FIG. 29 is a diagram for explaining a message transfer information setting request output to the own process by the new process in step S39 shown in FIG. 14;

FIG. 30 is a diagram for explaining a local reference table, a remote reference table, a process reference table, and a node reference table of a new process after performing the process of step S39 shown in FIG. 14;

FIG. 31 is a diagram for describing an operation when the concurrent object “objA” illustrated in FIG. 1 calls the concurrent object “Ftax” at 23:55:00 on March 31.

FIG. 32 is a flowchart of a process when the message dispatcher shown in FIG. 3 receives a message.

FIG. 33 is a flowchart of a process performed by the reference resolver at the time of reception of the reference resolver shown in FIG. 4;

FIG. 34 is a diagram for explaining an operation when the concurrent object “objA” illustrated in FIG. 1 calls the concurrent object “Ftax” at 00:00:00 on April 1;

FIG. 35 is a diagram for explaining an outline of an operation of the distributed processing system shown in FIG. 1;

FIG. 36 is a diagram illustrating a modification of the operation of the distributed processing system illustrated in FIG. 1;

FIG. 37 is a diagram for explaining another embodiment of the distributed processing system of the present invention.

[Explanation of symbols]

100: distributed processing system, 8: network, 1
0 ₁ , 10 ₂ , 10 ₃ ... computer nodes, 16 ₁ ,
16 ₂ … General process, 16 ₃ … Existing process (Yo
n), 16 ₄ ... new process (Yon), 12 ₁ , 12
₂ , 12 ₃ ... management process, 31 ... export processing unit, 32 ... import processing unit, 33 ... remote ID notification unit, 34 ... transfer information setting unit, 35 ... transmission reference resolution unit,
36: Reference resolution unit at reception, 37: Local reference table, 38
... Remote reference table, 39 ... Process reference table, 40 ... Node reference table, 51 ... Message passing function, 52 ... Process placement function, 53 ... Process creation function, 54 ... Process placement table, 55 ... Process placement unit, 56 ... Process Information setting unit, 57 Process information notifying unit, 252 Application programming function, 253 Message passing operator, 254 Reference resolver, 255 Communication media interface, 2
56 message sender, 257 message evaluation scheduler, 258 message dispatcher

Continuation of the front page (56) References JP-A-7-141204 (JP, A) JP-A-1-232432 (JP, A) Tokuyo Sho 58-502025 (JP, A) (58) Fields investigated (Int) .Cl. ⁷ , DB name) G06F 15/177 676 G06F 9/06 540 G06F 11/20 310

Claims (10)

(57) [Claims] 1. A distributed processing system in which a plurality of computer nodes are connected via a network and perform processing while performing message communication between a plurality of processes operating on the plurality of computer nodes. A first process for outputting a message including predetermined attribute information to invoke a process, a second process for performing a process in accordance with the input message, and a second process when a predetermined condition is satisfied On the basis of the process replacement start request instructing to replace the third process with the new process.
A first process outputs the message to the second process, and the third process outputs message transfer information when the third process is generated by the management process. A setting request is output to the second process, and processing according to the input message is performed. The second process includes a message transfer condition and a message transfer destination process in response to the message transfer information setting request. The reference information of a certain third process is set, and when the attribute information of the input message does not satisfy the message transfer condition, a process according to the input message is performed, and the message transfer condition is satisfied. Distributed processing for transferring the input message to the third process using the reference information Stem. 2. The management process inputs a process replacement start request including location information of a program module, information specifying the second process, and location information of a computer node on which a new process is to be arranged. The distributed processing system according to claim 1, wherein the third process is generated by reading the program module into the computer node and executing the program module. 3. When the third process is generated by the management process, a remote identifier of an object existing in the third process and a local identifier which is an identifier of the object in the process are generated. Generates a local reference table indicating the correspondence between the identifier, the message transfer condition, and the reference to the message transfer destination object, and does not set the message transfer condition of the local reference table and the reference to the message transfer destination object 2. The distributed processing system according to claim 1, wherein an export process is performed. 4. The process according to claim 3, wherein the third process is a process for an object of another process that makes a call, a remote identifier that is an identifier of the object on the network, a process identifier of a process to which the object belongs, and 4. The distributed processing system according to claim 3, wherein an import process for generating a remote reference table indicating a correspondence with the location information of the existing computer node is performed. 5. The management process has a process arrangement table indicating a correspondence between a process identifier of the process and location information of the computer node on which the process operates. 5. The distributed processing system according to claim 4, wherein the query is performed to obtain the process identifier and the location information of the computer node for each object of the remote reference table. 6. The third process, when the second process exists, outputs a message transfer information setting request including the local lookup table and a message transfer condition to the second process. The second process includes the third message included in the input message transfer information setting request such that the message is transferred to the third process when the attribute information of the input message satisfies the predetermined condition. The distributed processing system according to claim 5, wherein the local reference table and the remote reference table of the second process are updated using the local reference table and the message transfer condition of the process. 7. The third process inputs the updated local reference table from the second process, and the input local reference table of the second process;
7. The distributed processing system according to claim 6, wherein the local reference table and the remote reference table of the third process are updated using the message transfer condition. 8. When the message input by the second process no longer has the message that does not satisfy the message transfer condition, the message has a reference to the second process in the first process. 2. The distributed processing system according to claim 1, wherein the remote identifier of the entry of the concurrent object in the remote reference table of the process is changed to the remote identifier of the concurrent object of the third process. 9. A distributed processing method in which a plurality of computer nodes are connected via a network and perform processing while performing message communication between a plurality of processes operating on the plurality of computer nodes. And outputs a message to the second process, and when the predetermined condition is satisfied, the second process, which is the new process, is executed based on a process replacement start request instructing replacement of the second process with a new process. 3
The third process outputs a message transfer information setting request to the second process when the third process is generated, and performs processing according to the input message. The second process includes: In response to the message transfer information setting request, a message transfer condition and reference information of the third process which is a transfer destination process of the message are set, and the attribute information of the input message does not satisfy the message transfer condition. A distributed processing method of performing a process according to the input message, and transferring the input message to the third process using the reference information when the message transfer condition is satisfied. 10. The process replacement start request includes location information of a program module, information specifying the second process, and location information of a computer node where a new process is arranged. The distributed processing method according to claim 9, wherein the third process is generated by being read into a computer node and executed.