JPH06149756A - Method and device for decentralized information processing - Google Patents

Abstract

PURPOSE:To perform an individual process meeting a client's request by moving the process procedure that the client has to a server and performing the process on the server so as to couple the characteristic process procedure that the client has with the function of the server which has data and its data operation processing procedure. CONSTITUTION:An access data analyzing mechanism 201 has position information 202 on data decentralized and arranged in respective computer nodes and analyzes the position of decentralized data that a process procedure 200 being executed by an information process mechanism 11 of a computer node should refer to; if it is evident that data is present in another node 20 when the process procedure 200 tries to refer to the data, the execution of the process procedure is interrupted and the process procedure is transferred to the node 20 and restarted to perform the process cooperatively with a process procedure 2001 originally present in the node 20.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a distributed computer processing system, and more particularly, when the distributed computer systems each have a processing procedure and data, the processing procedures of a plurality of computer systems work together. The present invention relates to a distributed information processing method and an apparatus for realizing unique processing in a computer system that requests processing by performing the above.

[0002]

2. Description of the Related Art A distributed information processing system in which one computer system is used as one computer node and a plurality of the computer nodes are connected by a computer network is divided into those for the purpose of distributing the load of the computer and those for dividing the respective functions. There is a purpose of function distribution that performs processing in parallel.

The present invention relates to the latter function-distributed system. Generally, in such a function-distributed information processing system, as shown in FIG. 21, a plurality of computer nodes 1
0 and 20 are connected via the network 1.
Each of the computer nodes 10 and 20 has an information processing mechanism 1
1 and 21, mechanisms 12 and 22 for storing processing procedures, and data 13 and 23 referred to by the processing procedures.

In the distributed system having such a configuration, each of the computer nodes 10 and 20 has its own purpose and function, and the processing procedures and data required therefor are stored in that node. However, even if such division is performed, it is necessary to refer to the data stored in another node depending on the contents of the procedure processing. FIG. 22 shows this concept, and shows an example in which the processing procedure 14 being executed in the ten computer nodes a refers to the data 23 and 33 of the computer node b and the computer node c.

As a mechanism for accessing the data stored in such another node, a remote data access mechanism has been proposed so that the processing procedure can access the data of its own node and that of the other node without apparently distinguishing them. . As a mechanism for accessing the file system of another node, for example, NFS (Network File System) and RFS (Remote File) which are practically used in the Unix operating system.
e System) is a typical example. Further, a method of providing a resource management mechanism for data on a network and accessing other node data via this management mechanism in order to uniformly access distributed data is disclosed in, for example, Japanese Patent Laid-Open No. Japanese Patent No. 166158 and Japanese Patent Laid-Open No. 2-230439
It is known from the official gazette.

FIG. 23 shows the concept of such a conventional distributed information retrieval system. In FIG. 23, the server node, which is a database server, is requested to search for data that meets a certain condition 1, and the result is returned to the client node. Next, after performing a process in consideration of the search result under the condition 1, the database server is requested to search for data that matches the condition 2, and after obtaining the result, the process proceeds to the next process. The search result in these cases is the condition 1
It is all data corresponding to. When such a procedure is followed, there is a problem that the amount of data to be reciprocated between the client node and the server node becomes enormous when the searched data is multimedia including image data and the like as a target system. .

Further, when the processing procedure for accessing the data distributed to another node is clearly separated, the processing procedure is arranged in advance in the other node, a message is sent from the own node to the other node, and the processing is performed. A method of requesting is also proposed. This is based on a method called a client / server model, that is, the client requests the server to perform processing. From a different point of view, this method is an object-oriented processing method, which is a method of sending a message to an object in which data and a procedure for operating the data are integrated and requesting the processing execution. Known examples of this method include, for example, JP-A-62-126457.

[0008]

As described above, the conventional remote data access mechanism functions extremely efficiently when the number of distributed computer nodes and the frequency of accessing distributed data are low. However, as the number of computer nodes that make up the distributed environment and the amount of data to be referenced increase, the load on each computer node and the network significantly increases, and the processing performance deteriorates.

On the other hand, in the method based on the client / server model, it is necessary to analyze all the processing procedures to be executed in advance and arrange them in the server in advance. This analysis work is not problematic when the number of distributed computer nodes and the number of data are small, but a huge amount of work is required as the number of nodes and the types of procedures to be processed increase. Also, in an actual large-scale and wide-area distributed environment, the required specifications are changed or added, and the type of data used is also updated. Therefore, the processing procedure allocated to the server is changed or added each time. There is a need. In particular, the processing contents requested by the client vary widely, and it is not easy for the server side to prepare processing procedures capable of responding to all these requests.

The present invention has been made in view of the above problems, and an object thereof is to perform a process in a distributed information processing environment, which process corresponds to an individual process request of a computer node requesting the process. Provided are a distributed information processing method and an apparatus thereof, which can be executed on a computer node provided with target data, can reduce data transfer between computer nodes, and can perform information processing efficiently. That is.

[0011]

In order to solve such a problem, the present invention basically transfers the processing procedure peculiar to the client side out of the processing content requested by the client side to the server side. , The processing procedure and data stored in advance on the server side and the transferred processing procedure, for example, embed a subroutine, or the transferred programs read or write a file while exchanging data with each other. It is characterized in that processing is performed in cooperation with each other.

Further, as a more specific example, there is a distributed information processing method and its apparatus obtained from the following configuration. (1) When there is a processing procedure to be executed by the information processing mechanism of the own computer node, which has position information of data distributedly arranged in a plurality of computer nodes, which part of the processing procedure is stored in which computer node It is determined whether or not it is necessary to refer to the distributed data using the built-in position information, the computer node in which the data to be referred is stored as a unit, a unit that divides the process into processing units, and is divided.
It is determined whether the processing unit grouped for each computer node is stored in an executable form in the computer node to be executed, and if not stored, the processing unit is stored in the computer to be executed. Transfer to the node,
Means to be executed by the information processing mechanism of the computer node,
A configuration characterized by having. (2) When there is a processing procedure to be executed by the information processing mechanism of its own computer node, which has positional information of data distributed and arranged in a plurality of computer nodes, which part of the processing procedure is stored in which computer node. It is determined whether it is necessary to refer to the distributed data that has been
A unit for dividing the processing into processing units, with the computer node storing the data to be referred as a unit,
A processing unit grouped for each computer node is a computer load when the processing unit is executed in another computer node to be executed, and a computer and a network when transferring the processing unit to the computer node to be executed. Means for evaluating the load, the load amount for transferring the processing unit through the network through the network, and the load on the computer and the network when the calculation result of the processing unit is returned from another node to the own node; Depending on each load amount estimated by, whether the processing procedure to be executed should be transferred to another node for processing, or whether the data in another node should be transferred to the own node and the processing unit is executed in the own node. A configuration comprising: a means for judging. (3) means for analyzing the position of the distributed data to be referred to by the processing procedure being executed by the information processing mechanism of the computer node, which has position information of the data distributed to each computer node, and the processing procedure If you try to reference
If the data is in its own node, the processing is continued as it is, and if it is found that the data is in another node, the execution of the processing procedure is interrupted, transferred to the other node, and the other node And a means for resuming the execution of the processing procedure. (4) means for analyzing the position of distributed data to be referred to by a processing procedure being executed by the information processing mechanism of the computer node, which has position information of the data distributed in each computer node, and the processing being executed. When the data referred to by the procedure exists in another node, the load amount of the computer and the network when accessing the data via the remote data access mechanism between the nodes and the processing procedure being executed are interrupted to the other node. A means for evaluating the load on the computer and the network when transferring, the load on the computer when the processing procedure is executed by another node, and the load on the computer and the network when returning the calculation result to the own node And a function of interrupting the execution of the processing procedure, transferring to the other node, restarting the execution of the processing procedure at the other node, and returning the calculation result to the own node. Structure characterized by comprising: means was example, the.

[0013]

In the distributed information processing method and the apparatus thereof having the above-mentioned characteristics, the peculiar processing procedure possessed by the side requesting the processing is moved to the processing procedure and the data stored in advance in the destination of the processing procedure. By performing these processes in cooperation with each other, it is possible to realize a unique process that meets the request of the computer system that requests the process.

[0014]

Embodiments of the present invention will be described in detail below with reference to the accompanying drawings. In the following description, Examples 1 to 4 show examples describing the basic concept of the present invention, and Example 5
9 to 9 show examples when the present invention is specifically applied to individual systems. In the drawings for explaining the embodiments, those having the same function are designated by the same reference numeral, and a duplicate description will be omitted. (1) First Embodiment FIG. 1 is an example of a case where processing is performed by moving to another node.

In FIG. 1, there are processing procedures 14 to be executed by 10 computer nodes a. Processing procedure analysis mechanism 1
00 refers to the built-in distributed data arrangement information 101,
The processing procedure 14 is divided into processing units by the computer node in which the distributed data to be referred to is arranged, and then grouped. Reference numerals 141, 142 and 143 in FIG. 1 are examples of divided and grouped processing units.
1 is a processing unit a which refers to the data 13 arranged in the computer node a, and 142 is a processing unit b and a processing unit b ′ which refer to the data 23 arranged in the computer node b.
And 143 is the data 3 arranged in the computer node c.
3 is a processing unit c that refers to 3.

Reference numeral 150 denotes a processing unit execution control mechanism, which controls the execution of the divided processing units 141, 142 and 143 in a predetermined computer node. Figure 2
FIG. 3 is a diagram showing a basic operation flow of 150. 15
0 identifies the corresponding computer node that executes each processing unit, searches for whether the processing unit is registered in the computer node in a form executable by the information processing mechanism of the computer node, and is registered. If not, the processing unit is transferred to the computer node via the network. When the transfer of the processing unit is completed, 150 issues an execution request of the processing unit to the computer node.

In FIG. 1, the processing unit 142 is transferred to 20 computer nodes b, and then 20 computer nodes b are referred to while referring to the data 23 arranged in 20.
Run on. (2) Second Embodiment FIG. 3 shows another example in which processing is performed by moving to another node.

In FIG. 3, there are processing procedures 14 to be executed by 10 computer nodes a. Processing procedure analysis mechanism 1
00 refers to the built-in distributed data arrangement information 101,
The processing procedure 14 is divided into processing units by the computer node in which the distributed data to be referred to is arranged, and then grouped. Reference numerals 141, 142 and 143 in FIG. 1 are examples of divided and grouped processing units.
1 is a processing unit a which refers to the data 13 arranged in the computer node a, and 142 is a processing unit b and a processing unit b ′ which refer to the data 23 arranged in the computer node b.
And 143 is the data 3 arranged in the computer node c.
3 is a processing unit c that refers to 3.

Reference numeral 160 is a means for evaluating the load during execution of the divided processing units. 160 evaluates the load amount described below for each processing unit. There are various factors in the load amount here, but for example, the CPU time of the computer, the time required for input / output, the transfer time on the network, the storage capacity that the processing of the present invention occupies in the storage device of the computer, etc. is there.

A) A computer load amount when the processing unit is executed on a designated computer node. b) The load on the computer and the network when the processing unit is transferred to the designated computer node. c) The amount of load on the computer and the network required to return the calculation result when it is necessary to return the calculation result to the own node after executing the process on the computer node for which the processing unit is designated.

D) A computer in the case where the data to be referred to by the processing unit is transferred from the computer node to the own node and is processed by the own node without transferring the processing unit to the designated computer node. Network load. e) without transferring the processing unit to the designated computer node,
The amount of load on the computer and the network when processing is performed by the own node by accessing the data to be referenced by the processing unit via the information processing mechanism of the computer node and the own node.

The processing unit execution load evaluation mechanism 160 compares and evaluates the above five load amounts, judges the combination that gives the least load to each computer node and the network, and controls the execution of the processing unit based on the judgment. 150 in FIG.
Is a processing unit execution control mechanism, and is used when it is judged that the execution of the processing unit on the designated computer node is the lightest. The remote data transfer control mechanism 151 is used when it is determined that it is best to transfer the data in another node to the own node and execute the processing unit in the own node. In FIG. 5, computer node b
The data b of 23 is transferred to the computer node a by the instruction of 151, and the processing unit b referring to 23 is executed by the computer node a.

The processing unit execution control mechanism 150 controls the execution of the divided processing units 141, 142 and 143 in a predetermined computer node. FIG. 2 is a diagram showing a basic operation flow of the processing unit execution control mechanism 150. Further, the processing unit execution control mechanism 150 identifies the corresponding computer node that executes each processing unit, and whether the processing unit is registered in the computer node in a form executable by the information processing mechanism of the computer node. Is searched, and if it is not registered, the processing unit is transferred to the computer node via the network. When the transfer of the processing unit is completed, 150 issues an execution request of the processing unit to the computer node.

Reference numeral 152 denotes a remote data access control mechanism, which realizes remote data access via the information control mechanism of its own node and the control mechanism of another node. When the above method is judged to be the best, Used. (3) Third Embodiment FIG. 4 shows an embodiment in which a processing procedure being executed moves a computer node.

In FIG. 4, there is a processing procedure 200 being executed by 10 computer nodes a. In the above-described two configurations, the analysis of the processing procedure is not yet executed, but in this configuration, the information processing mechanism 11 is already executing it. The data referred to by the processing procedure 200 is provided to the 200 via the access data analysis mechanism 201. 201 is 20
The location information 202 indicating which computer node the data requested by 0 is located is stored. If 200 accesses the data 13 stored in the computer node a, it is referred to as it is. Further, when the 200 needs to refer to the data in another computer node, the 200 is transferred to the node and the processing is continued.
In FIG. 4, data 23 stored in 20 computer nodes b
Therefore, 200 is transferred to the information processing mechanism 21 of 20 and the process of 200 is continued at 21. If there is a processing procedure 2001 already stored in another node 20, for example, a subroutine is inserted,
The transferred programs execute processing in cooperation with the reading and writing of files while exchanging data with each other.

FIG. 5 shows how the processing procedure being executed is moved. That is, the processing procedure 200 is 10
When it is necessary to refer to the data 13 in the computer node a of the other node 20, it is executed in the computer node a, and when it is necessary to refer to the data 23 in the computer node b of the other node 20 due to the continuation of the process, Move to computer node b. If you need to access 13 more, 10 more
If it is necessary to go back and refer to the data in another computer node c, move to the computer node c.

As described above, the processing procedure 200 circulates a plurality of computer nodes and executes the processing, if necessary. (4) Fourth Embodiment FIG. 6 shows another embodiment in which a processing procedure being executed moves a computer node.

In FIG. 6, as in FIG. 4, there are processing procedures 200 being executed by ten computer nodes a. The data referred to by the processing procedure 200 is provided to the 200 via the access data analysis mechanism 201. The access data analysis mechanism 201 stores the position information 202 indicating in which computer node the data requested by the 200 is arranged. If the 200 accesses the data 13 stored in the computer node a, , Refer as it is. If it is necessary for 200 to refer to data in another computer node, the following load amount evaluation is executed via the processing load evaluation mechanism 204.

A) A computer load when the processing procedure is executed on another computer node. b) Amount of load on the computer and the network when the processing procedure is transferred to another computer node. c) The amount of load on the computer and the network required to return the calculation result when the calculation procedure needs to return the calculation result to its own node after executing the process on another computer node.

D) A computer and a network in which data to be referred to by the processing procedure is transferred from the computer node to the own node and is processed by the own node without transferring the processing procedure to another computer node. Load amount. e) The data to be referred to by the processing procedure is accessed via the information processing mechanism of the computer node and the own node without transferring the processing procedure to the designated computer node, so that the own node performs the processing. The amount of load on the computer and network when it is performed.

The processing load evaluation mechanism 204 compares and evaluates the above load amounts, determines a combination that gives the least load to each computer node and the network, and controls the execution of the processing procedure based on the result. When it is determined that the processing procedure 200 should be executed in the own node without being transferred to the other node 20, the processing is continued in the own node via the remote data access mechanism 205. If it is determined that the execution should be performed on another node, the processing is interrupted, and the processing procedure execution control mechanism 203 causes the other node 20 to execute 200.
To continue processing. Since it is necessary to refer to the data 23 stored in the computer node b of the other node 20 in FIG. 6, 200 is transferred to the information processing mechanism 21 of the other node 20, and the process of 200 is continued at 21.

Also in this configuration, as shown in FIG. 5, if necessary, a plurality of computer nodes are circulated to execute the processing. Next, an example in which the present invention is applied to an actual system will be described. (5) Fifth Embodiment FIG. 7 shows an example of a distributed database system in which a plurality of database systems are used as database servers and are connected to a network.

In FIG. 7, reference numeral 300 denotes a computer which provides a user with a data search service, which functions as a client server. A search condition 301 is given by the user via the man-machine interface, and is composed of a plurality of search conditions x, search conditions y and the like. 302
Is a search condition analysis mechanism, and corresponds to the processing unit execution control mechanism 150 of the configuration shown in FIG. A reference numeral 302 determines which search condition should be executed in which database server by referring to the distributed data arrangement information 303 which stores information regarding the type of data stored in each database server connected to the network. .

The transfer program generation mechanism 304 is a mechanism for generating a search condition execution procedure for actually executing the search condition determined by the search condition analysis mechanism 302.
A required module is taken out from the retrieval condition execution module library 5 and a program to be transferred is generated.
The generated search program 308 is transferred to and stored in the data search processing mechanism 307 of the server node 306. The data search processing mechanism 307 executes a database search via the link mechanism of the database management mechanism 310.

FIG. 8 shows the concept of the operation of the search program of FIG. When the search condition x is given, the client node uses 305 to generate a processing procedure for executing the search condition x. A processing procedure is similarly generated when the search condition y is given. Search condition x
When it is necessary to obtain data that matches the combination condition of the search condition y and the search condition y, a processing procedure for verifying the combination condition is generated.

The server node receives the above-mentioned search execution module, executes the search execution module, and saves the search results in the server node. After that, when the combination condition search program is transferred, the combination condition process is executed using the search results of the condition x and the condition y that were previously executed and saved. The result is returned to the client node as the final search result.

A specific example of data retrieval in this embodiment will be described below. The data stored in 311 is, for example, data such as personnel data, which includes attribute data that can be represented by a personal symbol, image data of the face, and a radiograph showing the health condition of the body. As the search condition, for example, a condition such as a man or a height of 170 cm or more can be represented by a symbol.

In addition, as a search condition, when it is desired to search for a person having a specific facial feature, the facial feature is
It is assumed that a specific image data processing program needs to be searched. This search changes depending on the client and cannot be prepared in advance on the server side. In the present embodiment, the program for determining the facial features is extracted from 305 in FIG. 7 and becomes the processing program 308 in 307. A function 308 provided by the database management mechanism 310 via the link mechanism 309 performs face feature extraction and retrieval.

According to this embodiment, there are the following effects. a) In a normal database search, when trying to obtain data that matches a certain condition from a server, the amount of transfer becomes enormous when the amount of corresponding data becomes large, which causes an excessive load on the computer and the network. However, according to the present embodiment, in the case where the transfer amount of the program is smaller than the transfer amount of the data, the load on the computer and the network is reduced, so that the throughput of the data search is improved.

B) When there are a plurality of search conditions, unnecessary data transfer can be suppressed by holding the intermediate data obtained by each search condition on the server side. c) Since it is possible to generate a search condition execution procedure on the client side and combine the procedure with a data search mechanism that the server side normally supports,
It is possible to execute a search process that is complicated and flexible and that suits the client's individuality. (6) Sixth Embodiment FIG. 9 is an example in which the present invention is applied to an object-oriented system.

In FIG. 9, two computer nodes 32 are shown.
0 and 321 are connected by a network, and each performs information processing by a processing unit called an object. An object is an encapsulation of data and a procedure program for manipulating the data, and the procedure is called a method. The data in the object cannot be directly operated from the outside, and the operation of the data in the object and the result thereof can be obtained by sending a message requesting method activation to the object.

The present embodiment is characterized in that a message transmission analysis control mechanism 322 shown in FIG. 9 is provided. The message transfer analysis control mechanism 322 corresponds to the access data analysis mechanism 201 shown in FIG. The message transfer analysis control mechanism 322 analyzes using the distributed data arrangement information including the destination of the message that each object in the computer node 320 is about to send. When the message transfer analysis control mechanism 322 determines that the message of each object should be sent as it is, it causes the message to be sent to another node as it is. In FIG. 9, the message of the object 323 is the object 32 in the computer node 321 as it is.
It is sent to 5.

On the other hand, when it is determined that the object should be moved, the object is moved to the computer node 321 as indicated by 324 in FIG. 12 and the processing is continued as the object 324 '. The object 324 'is the computer node 32
Processing is continued while exchanging a message with the object 325 in 1. According to this embodiment, there are the following effects.

A) In a normal object-oriented system, data manipulation is requested by a message. The message that conveys the operation instruction has a relatively small amount of data,
In the case where the operation result needs to be returned to the originator of the message, a large amount of data transfer may occur in replying the message. If messages are exchanged frequently and the amount of data is large, the transfer amount will be huge,
Although there is a problem that an excessive load is applied to the computer and the network, according to this embodiment, in the case where the transfer amount of the object is smaller than the transfer amount of the data, the load on the computer and the network is reduced. This has the effect of improving the processing efficiency of the method executed by the object.

B) The object sending the message needs to know what kind of message the object receiving the message can accept. However, according to the present embodiment, since the object that sends the message goes to the partner, the partner node can directly perform the processing, and it is possible to realize the unique processing that meets the client's request. If the object moved to the other party stays there as it is, it can be used as a means for realizing the version upgrade of the function provided by the other party node. (7) Example 7 FIG. 10 shows an example of a portable personal health management system. For health management, it is necessary to carry out tests such as a blood test and a urine test on the health condition of each person, and the wristwatch type portable personal health management system shown in this embodiment is used.

In FIG. 10, reference numeral 350 is a mobile unit in the health management system of this embodiment. The mobile unit 350 includes a CPU 351, a program library 352, and a database 353. Further, the program library 352 is provided with various diagnostic programs 354 individually tuned, and the database 353 is provided with diagnostic item inspection results 355, respectively.

The moving unit 350 is a wristwatch type portable type and is mounted on a human arm and moves with the human. On the other hand, various data inspection mechanisms are installed in various public places where the mobile unit 350 visits. The example of the data inspection mechanism shown in FIG. 10 is a toilet diagnostic system 35.
It is 6. The toilet diagnostic system 356 includes a urine analyzer 357 for analyzing urine components obtained from a urinal, a data storage device 358 for storing test data, and a CPU 359 for data processing.

The urine health examination includes personal analysis items as well as various items such as protein and blood (hematuria) in the urine. It is not possible to store all of these special inspection programs in the toilet diagnostic system 356. In this embodiment, the program unique to the individual is the library 3
52 to the diagnostic program 354 stored in the CPU 359.
The analysis data 361 is stored in the diagnostic item inspection result database 353 of the mobile unit 350 by uploading it to the mobile unit 350. Wireless is used for the download and upload described here.

These processes are carried out by the time an individual enters the toilet, has finished using it, and has left the toilet, and
It is automatically executed in daily activities without the individual being aware of it. Although FIG. 10 shows an example of the urine test in the toilet, the following example of the test may be considered in a public place where humans approach. a) Inspect the complexion while facing the mirror.

B) Check pulse rate while sitting in the chair. c) Check body temperature while taking a bath. It is possible to inspect. As described above, various health checkup systems installed in public places, and the health management subsystem put in human hands downloads to each health checkup program, which is tuned for the individual who owns wherever he or she goes. And store the results in your database. When this individual returns home, as shown in FIG.
Data collection device 36 of health management system 363 automatically
2 sucks up the inspection item data accumulated in the database 353 and performs a predetermined diagnosis, and then the diagnosis result 364
Is output. Of course, the location of the health management system need not be limited to the home, and may be installed in a public place.

According to this embodiment, there are the following effects. a) Inspection mechanisms for health management are distributed and arranged in public places, and various inspections are automatically performed every time when visiting the public places, and the inspection data is naturally accumulated in the mobile unit. Therefore, abundant health management is possible and the burden on each person's health check can be reduced.

B) The contents of processing required for the physical examination differ for each individual, and if necessary, the mobile unit mounted on one's arm is downloaded to various inspection systems, and the results are uploaded. Individual inspection and diagnosis are possible. In addition, the structure of each inspection system is simplified and flexible inspection is possible.

C) The physical movement of the diagnostic program is based on the spatial movement of a person, and when the above-mentioned program is moved to the inspection system, it is wireless by only a short distance, and the means for realizing the movement of the program is extremely simple. become. Further, the transfer amount of the data or the program is performed at a short distance, so that the moving means becomes simple. (8) Example 8 Next, an example in which the present invention is applied to a distributed control system will be described.

A control system for one plant such as a plant monitoring control system, or a wide area control system such as a power system or a traffic control system observes the state quantity of the control target, performs appropriate information processing, and then controls the control target. To generate a control signal. When the controlled object can be further divided into sub-systems, there is a hierarchical distributed control system in which the supervisory control system is also divided and a mechanism for controlling the divided supervisory control system is provided.

FIG. 12 shows a typical configuration example of the distributed supervisory control system. In FIG. 12, reference numeral 404 is a communication network, to which is connected an upper node 400 that controls the entire system as a whole and lower nodes 401, 402, and 403 that control partial systems. Each of the lower nodes 401, 402, and 403 is a control target 40.
In charge of 4, 405 and 406. When an example of a distributed control system is an electric power system or a traffic control system, the controlled object covers a wide range, the controlled object is divided according to the region or the control purpose, and these divided objects are partially divided. A distributed control method of controlling is adopted.

FIG. 13 shows a case where the power system is taken as an example. In FIG. 13, there are lower nodes 401, 402, and 403 below the upper node 400, which issues commands or reports via the communication network. The power system to be controlled is divided into sub-systems 404, 405, and 406, respectively, and 401, 402, and 403 are in charge of control, respectively. In FIG. 13, 407 is a power station, 408 is a power transmission line, and 409 is a substation bus. Although the upper node and the lower node are referred to here, in an actual example, for example, when the upper node is the central power feeding command station, the lower node corresponds to the local power feeding command station or the power station.

Next, as a specific control example, an example of supply and demand control of a thermal power generator performed at a central power supply command station will be shown. In a power system, the total consumption of electricity (called total demand) and the sum of electricity generated must always be equal. It takes a certain amount of time to increase or decrease the amount of power generated by a power generator such as a thermal power generator and cannot be changed rapidly. Therefore, the total demand is predicted for a certain period (several minutes to several tens of minutes), and the generator is instructed to increase or decrease the output in accordance with this prediction. This is called supply and demand control. FIG. 14 shows this concept,
If the total demand is predicted as shown by the curve in FIG. 14 at a time point n minutes after the current time, the difference between the current power generation amount and the predicted value is insufficient. In this case, it is necessary to increase the output of each generator, but each generator has characteristics, one generator can increase the output at high speed, but the other generator is slow Thus, there are variations in characteristics. Figure 1
Fig. 5 is a diagram showing this. At the central power supply command station, it is necessary to give an instruction to increase the output in consideration of the characteristics of these generators. It cannot be uniquely determined without considering various parameters including the dynamic characteristics of. At present, this dynamic characteristic is such that only fixed data such as 100 kW / min exists in the database of the computer of the central power feeding command station, and the output command using this is calculated.

In this embodiment, the monitoring control computer of the power plant is given a precondition to determine the output change characteristic at the present time, and the characteristic is returned to the central power feeding command station.
The purpose is to collect accurate dynamic characteristics at the present time and realize more accurate supply and demand control. FIG. 16 shows a more specific block configuration diagram of this embodiment. In FIG. 16, 4
Reference numeral 00 is a central power supply command station, and an information processing mechanism 410 inside the central power supply command station operates a supply and demand control program 411. Within the supply and demand control program 411, there is a subroutine 412 that determines which generator output will be increased or decreased with respect to the predicted total demand. Reference numeral 413 is a library in which various dynamic characteristic determination programs to be sent to each power station are stored, and 414 is a dynamic characteristic determination program that is generated according to the preconditions given by the operator of the central power supply command station and is generated in each power station. It is a mechanism to send.

On the other hand, the thermal power plant 401 has a characteristic determination program 416 and a plant internal state estimation program 417 sent from the central power feed command station to the internal information processing mechanism 415. Plant internal state estimation program 41
7 has been designed in advance for this power plant. Further, the thermal power plant 401 is provided with observation data 418 obtained from power generation plant components such as a power generator 419 and a boiler 420. Characteristic determination program 4
16 and the plant internal state estimation program 417 cooperate with each other to estimate the state inside the power plant at the present time and calculate the output change dynamic characteristic of the power generator at the present time.

FIG. 17 shows an example of output change dynamic characteristics of this generator. The power plant has a complicated control system inside, and various operation modes are possible according to the given constraint conditions. For example, in FIG. 17, the generator output change dynamic characteristics in consideration of economic efficiency such as fuel ratio, and the dynamic characteristics in the case where maximum output change is obtained even if some damage to the plant components is sacrificed. Is shown as an example.

In a normal operating state, it is usual to place importance on economic efficiency, but in an emergency where the amount of power generation is insufficient and a major blackout may occur, the amount of power generation is increased even if some damage to equipment is ignored. There is a need. In this way, by giving the dynamic characteristic determination program including various prerequisites to the computer of the power plant, it becomes possible to obtain the dynamic characteristic required by the operator of the central power feeding command station.

According to this embodiment, there are the following effects. a) Since the output change dynamic characteristic of the generator can be obtained under various constraint conditions, it is possible to determine the increase or decrease in the output of the thermal power generator according to the operator's request at the central power feeding command station. (9) Ninth Embodiment Next, an example of a hierarchically distributed power system monitoring control system will be described, as in the above embodiments. The present embodiment is an example applied to a system for planning an operating procedure of a power system.

FIG. 18 is a power system diagram as an example.
In FIG. 18, 430, 431 and 432 are power plants, and 433, 434, 435, 436, 437 and 43.
8, 439 are substations. Currently power transmission line 441
And 444 are cut off from the grid and no power is flowing. Here, the case where it is necessary to disconnect the power transmission line 440 from the system for construction is an example of this embodiment. Power transmission line 4
If 40 is disconnected from the grid as it is, the substation 438 located downstream cuts off the power supply and a power failure occurs. It is necessary to perform various operations to prevent this power failure. This is an example of planning an operating procedure for a power system.

In this case, various operating procedures are possible. The simplest procedure is to connect the stopped transmission line 441 to the grid and then disconnect the transmission line 440.
In addition, if the 441 is connected, the power plant 431 supplies power to both the substations 438 and 439, and there is a risk of overload. Therefore, the power transmission line 444 is activated, and the substation 437 generates electricity. Received power from the office 432,
The power station 431 may supply only 438.
FIG. 19 shows the connection state of the system after executing the latter procedure.

FIG. 20 is a block diagram of this embodiment. A local power supply command center 500 is an upper node, and a control station A 501 and a control station B 502 are lower nodes. The control center A controls the substations 433, 434 and 436 and the power transmission lines 440 and 441 in FIG. On the other hand, the control center B controls the substations 435 and 437 and the power transmission lines 443 and 441. The local power supply command station 500 has an operating procedure planning system 503 including an inference engine and a knowledge base as a system for planning operating procedures. Local Power Supply Command Office 500
Controls and manages the entire system shown in FIG. 18, including the subordinate control stations. However, the observed state quantities of individual substations and transmission lines are detailed data in the lower node control stations. is there.

In the illustrated example of FIG. 20, 503 has generated two operation procedures 504 and 505 from the data relating to the rules and the system configuration accumulated in the knowledge base. An overload check is necessary to determine which is better. Also, even if there is no overload, load rates close to overload should be avoided. 504 is a sequence of calculation programs and operating procedures that realize these items to be checked.
505 and are sent to the control stations 501 and 502.
At each control station, based on the received procedure table, the power flow flowing through the transmission line is estimated based on the system configuration (connection status) that changes for each procedure, overload and load factor values are calculated, and the results are fed to the local power supply. Return it to the command center 500. In addition, if there is specific information that makes it impossible to implement the procedure given to each control station (for example, a breaker failure that was not reported to the upper node), include this, and send it to the upper node. , Report the evaluation results of the procedure.

In this way, the operation procedure planning system 503 issues various processing instructions to the subordinate nodes and receives the results, thereby creating a better procedure.
According to this embodiment, there are the following effects. a) Since the upper node opens and evaluates the planned procedure in the lower node, an excellent procedure can be planned by using the data accumulated in the lower node.

[0068]

As can be understood from the above description, according to the present invention, in a distributed information processing environment, a process suitable for a unique process request of a computer node that requests the process can be performed.
It becomes possible to execute it on a computer node that has data to be processed. Further, since data transfer between computer nodes is reduced, efficient information processing becomes possible.

[Brief description of drawings]

FIG. 1 is a block diagram of means for performing processing by moving to another node.

FIG. 2 is a processing procedure movement determination algorithm to another node.

FIG. 3 is a block diagram of a means for performing processing by moving to another node.

FIG. 4 is a block diagram of a means for a processing procedure being executed to move a computer node.

FIG. 5 is a diagram showing the circulation of processing procedures.

FIG. 6 is a block diagram of means for a processing procedure being executed to move a computer node.

FIG. 7 is a block diagram of an embodiment for performing individual data search.

FIG. 8 is a processing flow of unique data search.

FIG. 9 is an operation diagram of a mobile object.

FIG. 10 is a block diagram of a portable health management system.

FIG. 11 is an operation diagram of a portable health management system.

FIG. 12 is a configuration diagram of a distributed monitoring control system.

FIG. 13 is a configuration diagram of a distributed supervisory control system using a power system as an example.

FIG. 14: Total demand forecast curve.

FIG. 15 is a generator output change dynamic characteristic diagram.

FIG. 16 is a block diagram of an application example to a supply and demand control system.

FIG. 17 is a dynamic characteristic diagram of a generator output change under various conditions.

FIG. 18 is a model system diagram.

FIG. 19 is a model system diagram.

FIG. 20 is a block diagram showing an application example to a system operation procedure planning system.

FIG. 21 is a diagram showing a general configuration of distributed information processing.

FIG. 22 is a block diagram showing access to data in another node.

FIG. 23 is a flowchart of condition matching in a database.

[Explanation of symbols]

 1 ... Network 10, 20, 30, 300, 306 ... Computer node 11, 21, 31 ... Information processing device 12, 22, 32 ... Data in each computer node 14, 24 ... Processing procedure in each computer node 100 ... Processing Procedure analysis mechanism 101 ... Analysis data arrangement information 150 ... Processing unit execution control mechanism 160 ... Processing unit execution load evaluation mechanism 201 ... Access data analysis mechanism 204 ... Processing load evaluation mechanism 302 ... Search condition analysis mechanism 350 ... Portable health management system Mobile unit 363 ... Fixed unit of portable health management system 400 ... Central power supply command station 401 ... Thermal power station 500 ... Local power supply command station 501, 502 ... Control station

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Junzo Kawakami 7-1-1, Omika-cho, Hitachi-shi, Ibaraki Hitachi Ltd. Hitachi Research Laboratory

Claims (18)

[Claims] 1. In a distributed information processing method in which a plurality of computer nodes are connected via a network, after a processing procedure sent from the own node moves between computer nodes, a processing procedure stored in another node in advance. And a distributed information processing method, wherein information processing unique to a sender of a processing procedure is realized by performing processing in association with data. 2. When each of the computer nodes is equipped with an inference engine, what moves between the computer nodes is all or a part of a knowledge base referred to by the inference engine. The distributed information processing method according to claim 1. 3. A device that moves between the computer nodes,
A processing procedure and data are included.
The distributed information processing method described. 4. The distributed information processing method according to claim 1, wherein the moving processing procedure circulates through a plurality of computer nodes. 5. The distributed information processing method according to claim 1, wherein the processing procedure moved between computer nodes is resident in the destination computer node. 6. When there is a processing procedure to be executed by the information processing mechanism of the own computer node, which has position information of a plurality of distributed data distributedly distributed to a plurality of computer nodes, which part of the processing procedure is Whether the distributed data stored in the computer node needs to be referenced is determined using the built-in position information, and the processing is divided into processing units with the computer node storing the data to be referenced as a unit. , It is determined whether the processing unit divided and grouped for each computer node is stored in an executable form in the computer node to be executed. If not stored, the processing unit is executed as described above. 2. The distributed information according to claim 1, wherein the processing unit is transferred to a computer node to be executed, and the processing unit is executed by an information processing mechanism of another computer node. Processing method. 7. When there is a processing procedure to be executed by the information processing mechanism of the own computer node, which has the position information of the distributed data distributed to a plurality of computer nodes, which part of the processing procedure is which computer node It is determined whether it is necessary to refer to the distributed data stored in, using the built-in location information, and the processing is divided into processing units with the computer node storing the data to be referred as the unit. A processing unit grouped for each computer node, a computer load amount when the processing unit is executed in another computer node to be executed, and a computer when the processing unit is transferred to the computer node to be executed. And the load on the network, the load to transfer the processing unit through the network, and the calculation result of the processing unit from another node to the own node. If the load amount of the computer and the network is evaluated, the processing load to be executed should be transferred to another node for processing, or the data in the other node should be processed by the estimated load amount. It is determined whether or not the processing unit should be transferred to execute the processing unit on its own node, and whether the processing unit divided and grouped for each computer node is stored in an executable form in the computer node to be executed. It is determined, and if not stored, the processing unit is transferred to the computer node to be executed, and the information processing mechanism of the computer node executes the processing unit. 1. The distributed information processing method described in 1. 8. The position information of the data distributed in each computer node is analyzed, and the position of the distributed data to be referred to by the processing procedure being executed by the information processing mechanism of the own computer node is analyzed, and When the processing procedure being executed tries to refer to the data, if the data is in its own node, the processing is continued as it is, and if it is found that the data is in another node, the processing procedure is executed. The distributed information processing method according to claim 1, wherein the processing is interrupted, transferred to the other node, and the execution of the processing procedure is restarted in the other node. 9. The position information of data distributed to each computer node is analyzed, and the position of the distributed data to be referred to by the processing procedure being executed by the information processing mechanism of the own computer node is analyzed, When the data referred to by the processing procedure exists in another node, the load amount of the computer and the network when accessing the data through the remote data access mechanism between the nodes, and the other node by interrupting the processing procedure being executed. The load on the computer and the network when transferring the data to the node, the load on the computer when the processing procedure is executed on another node, and the load on the computer and the network when returning the calculation result to the own node. When it is determined that it is preferable to execute the process on another node, the execution of the process procedure is interrupted, transferred to the other node, and the process procedure is executed on the other node. Is restarted, the calculation result, characterized in that back to the local node according to claim 1
The distributed information processing method described. 10. In a distributed information processing apparatus in which a plurality of computer nodes are connected via a network, another node in which processing procedures and data are stored in advance and processing unique to a computer system requesting processing. Distributed information comprising: a self-node for sending a procedure to the other node, wherein the processing procedure sent from the self-node is processed in cooperation with the processing procedure and data of the other node. Processing equipment. 11. When each of the computer nodes is equipped with an inference engine, what moves between the computer nodes is all or part of a knowledge base referenced by the inference engine. The distributed information processing device according to claim 10. 12. The distributed information processing apparatus according to claim 10, wherein the one that moves between the computer nodes includes a processing procedure and data. 13. The distributed information processing apparatus according to claim 10, wherein the moving processing procedure circulates through a plurality of computer nodes. 14. The distributed information processing apparatus according to claim 10, wherein the processing procedure moved between computer nodes resides at the destination computer node. 15. When there is a processing procedure to be executed by the information processing mechanism of the own computer node, which has position information of a plurality of distributed data distributedly arranged on a plurality of computer nodes, which part of the processing procedure is which? Means for determining whether or not it is necessary to refer to the distributed data stored in the computer node using the built-in position information, and dividing the processing into processing units with the computer node storing the data to be referenced as a unit Then, it is determined whether the processing unit divided and grouped for each computer node is stored in an executable form in the computer node to be executed. Means for transferring to a computer node to be executed and having the information processing mechanism of another computer node execute the processing unit. 0 dispersed data processing apparatus according. 16. When there is a processing procedure to be executed by the information processing mechanism of its own computer node, which has the position information of the distributed data distributedly distributed to a plurality of computer nodes, which part of the processing procedure is which computer node Means for determining whether it is necessary to refer to the distributed data stored in, using the built-in position information, and dividing the processing into processing units with the computer node storing the data to be referenced as a unit. , When a processing unit divided and grouped for each computer node is executed in another computer node to be executed, and when the processing unit is transferred to the computer node to be executed Computer and network load, the load of transferring the processing unit through the network, and the calculation result of the processing unit from other nodes. Means for evaluating the load of the computer and network in the case of back to de by the load estimated by said means, the processing to be performed procedure, it should be processed transferred to another node,
A means for transferring data in another node to the own node and determining whether the processing unit should be executed in the own node, and a processing unit divided and grouped for each computer node is executed in the computer node to be executed. It is determined whether the processing unit is stored in an executable form. If not stored, the processing unit is transferred to the computer node to be executed, and the information processing mechanism of the computer node executes the processing unit. 11. The distributed information processing apparatus according to claim 10, further comprising: 17. A unit for analyzing the position of the distributed data to be referred to by a processing procedure being executed by an information processing mechanism of the own computer node, the unit having position information of data distributedly arranged in each computer node, When the processing procedure being executed tries to refer to the data, if the data is in its own node, the processing is continued, and if it is found that the data is in another node, the processing procedure is executed. 11. The distributed information processing apparatus according to claim 10, further comprising: a means for interrupting the processing, transferring the processing to the other node, and restarting the execution of the processing procedure in the other node. 18. Means for analyzing the position of distributed data to be referred to by a processing procedure being executed by an information processing mechanism of its own computer node, which has position information of data distributedly arranged in each computer node, and said execution When the data referred to by the processing procedure in the middle exists in another node, the load amount of the computer and the network at the time of accessing the data through the remote data access mechanism between the nodes and the processing procedure being executed are interrupted. The load on the computer and the network when transferring to another node, the load on the computer when the processing procedure is executed on the other node, and the load on the computer and the network when returning the calculation result to the own node. A means for evaluating and interrupting the execution of the processing procedure, transferring to the other node, restarting the execution of the processing procedure at the other node, and returning the calculation result to the own node. Distributed data processing apparatus according to claim 10, wherein the means having a function of, comprising the to.