JPH0619860A - Control method for distributed system - Google Patents

Abstract

PURPOSE:To permit a processing site to easily learn the state transition time of a controller so as to make the load of control monitor as for the controller to be appropriate in a distributed system consisting of a control site having the controller and the processing site transmitting a control reguest to the control site. CONSTITUTION:The processing site 1A adds request output time (value beta) to the control request and transmits it to the control site 2. The control site 2 executes the request and turns on the control flag 25 of 1A so as to hold the value beta at the access time 24 of 1A. When the state of the controller 4 is transited, the control flag 25 of 1A is turned off. When it does not transit, the control flag is held to be ON. Then, the processing site 1A adds request output time (value alpha) to a monitor request as for the controller 4 and transmits it to the control site 2. The control site 2 returns the present flag 25 of 1A and state transition time alpha-beta to the processing site 1A with a state buffer 26 showing the state of the controller 4.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to one or more sites (referred to as control sites) that have a control device and perform control based on received control requests, and one or more sites that issue control requests to this control site. In a distributed system consisting of (referred to as a processing site), a control method for the processing site to easily know the state transition time of the control device and to appropriately control the frequency (load) of issuing a request such as a control request to the control site. Regarding

In the following figures, the same reference numerals indicate the same or corresponding parts.

[0003]

2. Description of the Related Art FIG. 5 shows an example of the configuration of this type of distributed system. In the figure, 1 (1A, 1B, 1C) is a processing site, 2 is a control site, 3 is a network connecting the processing site 1 and the control site 2, and 4 is a control device of the control site 2. 11 is each processing site 1 (1 in this example)
The processor in A).

Next, in the control site 2, 21 is a processor, 22 is a controller for directly controlling the control device 4, and 23 is a bus connecting the processor 21 and the controller 22. In the distributed system including the plurality of sites 1 and 2 as described above, the control site 2 indicates the request from the other plurality of processing sites 1A to 1C to the control site side to show the state of the control device. A state buffer (26, which will be described later) and the like are provided for sequential processing. However, the processing sites 1A to 1C may control the processing of the control device in a time-divisional manner and control them in parallel.

By the way, in such a distributed system,
It is desirable that the processing site 1 monitor the state of the control device 4 at an appropriate frequency (load) and send a control request to the control site 2 at an appropriate timing. To this end, the processing site 1 changes the state transition time of the control device 4 (that is, how much time has passed since the state of the target control device 4 of the control site 2 changed (how long and how long it remains in that state). )) Need to know. To this end, (a) the same clock is guaranteed in all systems, and the time at which the process of changing the state of the control device 4 is performed on the control site 2 side is monitored. And processing site 1
(B) Save the time when the processing in which the state of the control device 4 is changed at the processing site 1 is saved, and after the confirmation of the status confirmation request, save the current time. There are two methods of calculating the state transition time from the time difference.

[0006]

However, in the above method (a), the clock guarantee is a problem (1 second at the processing site is not necessarily the same as 1 second at the control site). In the method (b), when there are a plurality of processes for the control device in the processing site 1, it is necessary to have a structure in which the processing time can be shared and accessed.

Therefore, an object of the present invention is to solve this problem and provide a control method for a distributed system in which a processing site can easily know the state transition time of a control device based on its own clock. .

[0008]

In order to solve the above-mentioned problems, the control method of the present invention controls a control device (4 or the like) of the control device itself based on a received control request and controls the control device. One or a plurality of control sites (2 or the like) that hold the latest state data (in the state buffer 26 or the like) and one or a plurality of processing sites (1 or the like) that send the control request to the control site are provided. In the distributed system, when transmitting the control request, the processing site attaches to the control request data (312, etc.) the output time of the request (311, etc., hereinafter referred to as the first access time) read from its own clock. To the control site, and the control site starts control of the control device in response to the transmitted control request, and at the same time, the control flag related to the processing site within itself ( No. 5, etc.) is turned on and the first access time is stored (as the access time 24 (value β) etc.), and thereafter, the control site changes the state of the control device to a control request of another processing site. The control flag related to the processing site is kept on until the transition based on the other hand, while the control flag is turned off when the state transits for some reason, and after the control request is transmitted, When the processing site sends a monitoring request to the control site, the monitoring request data (312 or the like) outputs the request output time (311 (value α) or the like read from its own clock, hereinafter referred to as the second access time. ) Is added to the control site, and the control site responds to this monitoring request by the control flag 321 in the received data 32 (as the control flag related to the processing site. Etc.) and date data (received as data 3 of the control device
2) and the time of the difference between the second access time and the first access time (as the state transition time 322 (value α−β) in the received data 32). Send it to the processing site.

[0009]

The processing site sends the request time (β) (access time) read from its own clock at the time of control request together with the control request data, and the control site saves the access time for each processing site. Then, in response to a request from the processing site such as for status monitoring, the control site stores the new access time (α) added to this request data and the saved access time (β) at the time of the control request. Difference (α
-Β), and returns it to the processing site as the state transition time of the control device 4. By doing so, the processing site concerned can easily know the state transition time of the control device 4.

[0010]

Embodiments of the present invention will be described below with reference to FIGS. FIG. 1 shows an embodiment of the configuration of a distributed system according to the present invention. This distributed system comprises a plurality of processing sites 1A to 1C and a control site 2 having a control device 4. The processor 21 of the control site 2 uses the access time 24 and the control flag 2 for each processing site 1A, 1B, 1C.
5 and also has a status buffer 26 for storing the latest status of the controller 4.

FIG. 2 is an explanatory diagram of the operation when the processing site 1A controls the control device 4. That is, the processing site 1A transmits to the control site 2 as the transmission data 31, the control request data 312 indicating the content of the control request, and the access time 3 as the control request output time read from its own clock.
Send with 11 added. The processor 21 of the control site 2 which has received the transmission data 31 sends the control request data 312 to the controller 22 of the control device 4, controls the control device 4 in response to this request, and the controller 22 sends the result to the control device 4. It is input as the status 27 of No. 4 and stored in the status buffer 26 for storing the latest status in the processor 21. Then, the content of the access time 24 held for the processing site 1A is updated to the received access time 311 (value is β), and the control flag 25 corresponding to the processing site 1A is turned on (“1”). Then, the control flags 25 of the other processing sites 1B and 1C are turned off.
Set to (“0”).

FIG. 3 is an explanatory diagram of an operation when the control device 4 itself transits to a state for some reason other than the control request of the processing site 1. That is, the control device 4 notifies the controller 22 of the state transition, and the controller 22 notifies the processor 21 by the control device status 27.
The processor 21 stores the state in the state buffer 26 of the control device in the processor 21 and turns off the control flags 25 of all the processing sites 1A to 1C. However, the access time 24 remains unchanged.

FIG. 4 is an explanatory diagram of the operation when the processing site 1A monitors the control device 4. That is, the processing site 1A attaches the monitoring request data 312 as the transmission data 31 to the control site 2 with the access time 311 (value is α) as the monitoring request output time read from its own clock. Then, the processor 21 of the control site 2 that has received this, along with the contents 323 of the latest state buffer 26 of the control device 4 in the processor 21 in the received data 32 to be returned to the processing site 1A, and the newly received monitoring request. From the value α of the access time 311 of the previous access time 24 of the processing site 1A in the processor 21
The value α-β obtained by subtracting (value β) is set as the state transition time 322, and the control flag 25 for the processing site 1A is set.
As the control flag 321.

When the control flag 321 of the received data 32 is off, the processing site 1A checks the contents 323 of the sent state buffer and newly detects that some state transition has occurred after the control processing of its own site. Determine whether to control.

[0015]

According to the present invention, the processing site sends its request time (β) (access time) read from its own clock at the time of a control request together with its control request data, and the control site sends the access time to the processing site. Each time, a new access time (α) added to this monitoring request data in response to a request from the processing site such as status monitoring, and the saved access time (β) at the time of the control request Since the difference (α-β) is returned to the processing site as the state transition time of the control device 4, the processing site itself does not have a clock synchronization mechanism between distributed systems and a special shared access structure of processing time. It is possible to know the state transition time of the control device relating to a plurality of processes (processes divided into sites and processes in the same site) executed by the same method. The requesting side (processing site side) flexibly adjusts the weight of processing to the control device (for example, loose monitoring for long state transition time) according to the frequency of change of state of the control device (state transition time). You will be able to judge.

[Brief description of drawings]

FIG. 1 is a configuration diagram of a system as an embodiment of the present invention.

FIG. 2 is an explanatory diagram of control operation from a processing site

FIG. 3 is an explanatory diagram of a state transition operation of the control device.

FIG. 4 is an explanatory diagram of a status confirmation operation from the processing site.

FIG. 5 is a block diagram of a distributed system

[Explanation of symbols]

 1 (1A to 1C) Processing site 2 Control site 3 Network 4 Control device 11 Processor 21 Processor 22 Controller 23 Bus 24 Access time 25 Control flag 26 State buffer 31 Send data 32 Receive data 311 Access time 312 Request information 321 Control flag 322 State Transition time 323 Controller status data

Claims (1)

[Claims] 1. One or a plurality of control sites that control their own control device based on a received control request and hold the latest status data of the control device, and send the control request to the control site. In the distributed system including one or a plurality of processing sites, the processing site outputs the control request data to the control request data at the time of outputting the request (hereinafter referred to as the first access time). Is added to the control site, and the control site starts control of the control device in response to the transmitted control request, and at the same time turns on a control flag related to the processing site in itself, The first
The access time of the control site is stored, and thereafter, the control site keeps the control flag related to the processing site on until the state of the control device transits based on a control request of another processing site, and the like. When the state is changed for some reason, the control flag is turned off, and after transmitting the control request, the processing site concerned transmits the monitoring request to the control site, and the monitoring request data includes its own The request output time read from the clock (hereinafter referred to as the second access time) is sent to the control site, and the control site responds to the monitoring request with the control flag and the control flag related to the processing site. A time difference between the second access time and the first access time is transmitted to the relevant processing site together with the latest state data of the control device. Control method of a distributed system that.