JPH04180150A - Decentralized processing system - Google Patents

Abstract

PURPOSE:To attain a decentralized processing system having high extending performance by forming the software with each object defined as a unit and handling a data base, a knowledge base, etc., in an integrated way with all object made resident in a main storage. CONSTITUTION:The computer systems 11i1 - 11in are connected to each other via a local area network LAN 12 and also communicatable with the computer systems 11j1 - 11jn which are connected to a LAN 14 via a wide area network WAN 13. The data, the knowledge, and the procedure included in the system 11i1 consist of the units of objects which are all resident in a main storage. That is, the system 11i1 consists of the objects 15i1 - 15in in regard of all functions of the system. Thus a decentralized processing system having high extending performance can be obtained.

Description

[Detailed Description of the Invention] [Objective of the Invention (Industrial Application Field) The present invention relates to a distributed processing system constructed by connecting a plurality of computer systems via a network and utilizing functional distribution and communication. Regarding.

(Prior art) Conventionally, multiple computer systems are connected to a LAN (Local A).
rea Network) and WAN (Wide Area Network)
In information processing systems connected via networks such as A Network, there are various databases, knowledge bases, and programs that use them on each computer system, and a large number of processes communicate as necessary. A distributed processing system has been realized while communicating using software.

(Problems to be Solved by the Invention) However, such a distributed processing system has the following disadvantages.

(1) Inefficiency due to coexistence of databases and knowledge bases Databases and knowledge bases originally developed as independent technologies, so they exist in different forms. Facts and frames in a knowledge base are basically data.
There are many duplicates of information in the database. For this reason, when inferring using a knowledge base, a method has been used in which information required at the start of inference is extracted from the database and converted into a knowledge base format. In this method, the time required to convert the database into a knowledge base created overhead, which slowed down the inference speed. Particularly in inference using large-scale data, overhead accounts for a large proportion of the total inference time, which has become a major problem.

On the other hand, there is a method of accessing the database each time data in the database is needed during inference. Although this method eliminates unnecessary database access, switching between the knowledge base system and the database system occurs frequently, and the overhead caused by this reduces inference speed.

(2) Inefficiency due to the mixture of different types of databases There are various types of databases, such as hierarchical, network, and relational types. This is not a problem in application systems that target only a single data system, but in application systems that handle information related to multiple database systems, it is necessary to collect related information and create a new database, or to create a new database by collecting related information. It is necessary to take measures such as developing applied systems with awareness of the existence of With the former method, there was a problem with data consistency due to having duplicate data, and with the latter method, there were problems such as a heavy burden on the application system.

(3) Communication opacity When communicating via a network between processes existing in different computer systems, it is necessary to use communication procedures defined between the computer systems. Therefore, the structure of software is different when a certain process is realized by multiple processes within the same computer system through functional distribution and when it is realized by multiple processes between multiple computer systems. It had become.

(4) Slow speed of auxiliary storage devices Database management systems for managing large amounts of data use auxiliary storage devices such as magnetic disks to store data. Although auxiliary storage devices have the advantage of being low cost, their data access speed is much slower than that of main storage devices, and this slow speed has been a problem in real-time systems that require high-speed response.

Based on the above, the present invention eliminates differences in information storage formats such as databases and knowledge bases, and eliminates differences in communication means between computer systems and between computer systems via a network, thereby achieving higher performance. The purpose of this research is to provide a distributed processing system that allows efficient construction of information systems.

[Configuration of the Invention (Means for Solving the Problem) The distributed processing system of the present invention constructs software for arithmetic processing in each computer system using a plurality of objects resident in the main memory, and each object is configured to: An element object part for specifying other objects to be referenced, a slot part for defining relationships with other objects and attributes of the own object, and a rule part for defining knowledge for making inferences. and a procedure section in which procedures for performing arithmetic processing are defined.

(Function) In the present invention, since software is constructed using objects as units, and all objects are made to reside in main memory, databases, lxI knowledge bases, procedures, etc. can be uniformly accessed from any computer system. can be handled. This makes it possible to realize a highly reliable and highly scalable distributed processing system.

(Example) Hereinafter, one embodiment of the present invention will be described with reference to the drawings.

FIG. 1 shows a functional configuration diagram of a distributed processing system according to an embodiment of the present invention. Each computer system 11i, , computer system lli, . . . computer system 11i,
are connected to each other via a LAN 112, and to another computer system 1lJ1, computer system 11j2, . . . which are connected to another LAN 112 via a WAN 13.
- Communication is also possible with the computer system 1ljn. Computer system 111. The data, knowledge, and procedures within are all constructed using objects that reside in main memory as units. That is, computer system lli has object 15i,1, object 15i,...
All of its functions are realized by objects 15i, n, etc. The same applies to other computer systems.

FIG. 2 is an explanatory diagram of each object of the present invention, and FIG. 2(A) shows the basic configuration of the object. The object 15 basically includes an element object part I6, a slot part 17, a rule part 18, and a procedure part 19.
It consists of The element object section 16 is for declaring the objects that make up the object 15, and is made up of arbitrary object names. In other words, it is possible to create a composite object by combining several objects to create a new object.

Furthermore, the objects to be combined may be those defined in other computer systems.

Next, the slot section 17 is made up of arbitrary number of slots and combinations of slot values, and defines attributes of the object and relationships with other objects. A 1s-a relationship is provided to define a parent-child relationship as a special relationship.
Inheritance between objects is now possible.

Additionally, each slot can have facets for defining metadata for the slot.

Knowledge for making inferences is defined in the rule section 18.

The procedure section 19 is a collection of procedures, etc. for accessing the data defined in the slot section I7, and is composed of a procedure name, an argument, and a procedure unit. All access to slots is performed only through procedures defined in procedure section 19. One object is an element object part 16,
Although it is not necessary to have all of the slot part I7, rule part 18, and procedure part 19, one of them is definitely required. The configurations of other objects are also similar.

FIG. 2(B) shows the object structure of the database system implemented in the distributed processing system of the present invention. A database system is a single complex object 1
It is realized as 5A. A group of objects for storing data is defined in the element object section 16A. Attributes of the database system as a whole are defined in the slot section 17A.

Processing for the entire database system is defined in the procedure section 19A. Another object group is defined in the element object section of each object of the object group defined in the element object section 16A, and a database system is realized by the network configuration of the object group as a whole. Using such a network configuration, multiple database systems that share lower-level objects can be easily realized. The terminal objects located at the lowest level of the network configuration are shown in Figure 2 (
It has only a slot section 17B and a procedure section 19B as shown in C), and serves as the final data storage location. The slots in the slot portion 17B represent data names, and the slot values become data. A demon added to the slot is defined in the procedure section 19B.

Data is added to the database by creating new objects, creating slots, etc., and adding element object declarations. However, if an object state already exists, only the element object declaration is added. Data is deleted from the database by deleting a declaration from the element object 16, deleting the corresponding object, deleting a slot, etc. However, if the object to be deleted is used elsewhere, only the declaration of the element object 16 is deleted.

Updating data in the database is achieved by changing the slot value of the corresponding object. Database search processing is realized by calling the corresponding procedure. Furthermore, all objects reside in the main memory within the computer system, realizing a high-speed yarn database management system.

Next, FIG. 2(D) shows the object structure of the knowledge base system realized in the distributed processing system of the present invention. A group of objects storing facts is declared in the element object section 15c. Rule section 18C
The production rules used for forward and backward inference are defined.

A procedure group called from a rule and a demon procedure group added to a slot are defined in the procedure section 19G.

Addition or deletion of facts by inference or the like is performed by creating or deleting a corresponding object or slot, and adding or deleting declarations of element objects. Furthermore, if the object to be added or deleted already exists or is being used elsewhere, only the addition or deletion of element object declarations is performed, as in the case of database systems. In this way, the database system and the knowledge base system can use objects with the same configuration. Therefore, the overhead of data conversion due to differences in storage formats can be eliminated, making efficient processing possible.

All communication between processes is realized as communication between objects, and interprocess communication such as data transfer and procedure activation is handled exactly the same whether the objects are on the same computer system or on different computer systems. This makes communication transparent and enables efficient software development.

FIG. 3 is a system configuration diagram when the distributed processing system of the present invention is applied to plant monitoring and control. Computer system lli and computer system ltJ are connected to LAN12.
connected via the computer system lli.

is a computer system 11J1 that performs plant monitoring and control functions.
performs a plant diagnostic function. computer system l
In the case of the plant database 15a, the plant monitoring system 15b, the plant control system 15c, etc., the functions of li are realized by a group of objects including objects. The functions of the computer system 11J are realized by a group of objects including complex objects such as the plant diagnosis knowledge base system 15d.

FIG. 4 shows the object structure of the plant database 15a and the plant diagnosis knowledge base system 15d. The plant database 15a includes an element object section 16a that declares a group of plant equipment objects, a slot section 17a that defines plant database attributes, and a procedure section 19a that defines a plant database processing program.
It consists of The plant diagnosis knowledge base system 15d also includes an element object section 16d that declares a group of plant equipment objects, a slot section 17d that defines plant diagnosis knowledge base attributes, a rule section 18d that defines diagnosis rules, and a procedure section that defines processing necessary for diagnosis. Part 19d
It is composed of Element object parts 16a, 16d
The plant equipment object group declared in is composed of equipment object A, equipment object B, .

Update of plant data in accordance with changes in plant status is performed by a procedure defined in an object corresponding to each piece of equipment inputting plant data via the process input/output device PIO and updating slot values.

Additionally, processing such as converting engineering units and saving historical data is performed by procedures defined within the object.

Calculation processing using a plurality of plant data is realized by a procedure defined within a composite object made up of related objects, and by updating slot values within the composite object. Data searches throughout the plant database are performed by procedures defined within the plant database object 11.

The plant diagnosis knowledge base system 15d has a rule section 1.
By making inferences using the diagnostic rules defined in 8d, abnormalities in the plant are diagnosed. Plant data required for executing diagnostic rules is accessed through procedures defined in the corresponding objects. These object groups related to plant data are the same as the object group managed by the plant database +5a, and no special data conversion is required for use as a knowledge base, allowing efficient execution.

In this embodiment, the main memory of the computer system lli includes a plant database 15a and a plant monitoring system 1.
5b, the plant control system 15c is assigned a group of objects constituting the plant database, and the computer system oJ is assigned the plant diagnosis knowledge base 15d, but no matter how these objects are arranged in terms of the software configuration. No need for modification,
Added a new computer system and established LAN due to increased functionality.
It is possible to easily construct a distributed processing system using three computers connected to 12.

[Effects of the Invention] As described above, according to the present invention, databases, knowledge bases, procedures, etc. can be handled in a unified framework. Furthermore, in a distributed processing system using a plurality of computer systems connected via LAN, WAN, etc., it is possible to construct an information processing system with high execution efficiency and high expandability.

[Brief explanation of drawings]

FIG. 1 is a functional configuration diagram of a distributed processing system according to the present invention.
FIG. 2(A) is a configuration diagram of an object according to the present invention, FIG. 2(B) is a configuration diagram of an object of a database system according to the present invention, and FIG. 2(C) is a configuration diagram of an end object constituting the database system. FIG. 2(D) is an object configuration diagram of the knowledge base system according to the present invention, FIG. 3 is a system configuration diagram in which the distributed processing system of the present invention is applied to a plant monitoring and control system, and FIG.
It is a block diagram of the plant database object and knowledge base object in the system of a figure. 11...computer system, 12.14-LAN,
13.WAN, 15...Object, 16...Element object section, I7...Slot section, 18...
Rules Department, 19...Procedure Department. Agent Patent Attorney Crest 1) Makoto 11i+ 11i2 1
1inN1 diagram (A) (B) (D) M 2 Figure 11t1 Figure 3

Claims (1)

[Claims] In a distributed processing system in which a plurality of computer systems are connected via a network system and information processing is distributed among the plurality of computer systems, software for performing arithmetic processing on each of the computer systems is stored in a plurality of computers residing in the main memory. Each of the objects includes an element object part for specifying another object to be referenced, and a slot part for defining the relationship with the other object and the attributes of the own object; A distributed processing system characterized by always having at least one of a rule section for defining knowledge for performing inference and a procedure section for defining procedures for performing arithmetic processing.