JP4569024B2 - Operation / standby switching method, operation / standby determination method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a network control system, and more particularly to an operation / standby switching or determination method when nodes constituting a network are duplicated into an active system and a standby system.
[0002]
[Background Art and Problems to be Solved by the Invention]
In recent years, in the field of control systems, a protocol defined by the FA Control Network Technical Committee (“FA Link”) for the purpose of unifying the specifications for control LAN in the industry. Protocol ").
[0003]
Since the specifications of the FA link protocol are open to the public, they will be briefly described here.
First, in the past, a control network such as a plant control system (referred to as a network control system) has adopted a unique method for each company. There was impossible or great difficulty. Conventionally, it has been necessary to construct a special network for control.
[0004]
The FA link protocol enables the network to be realized by a general-purpose LAN such as Ethernet in order to solve the above problem. That is, conventionally, general-purpose LANs such as Ethernet have low reliability due to data collisions and the like, and are unsuitable for use for control. However, with this FA link protocol, the reliability of general-purpose LANs is improved. It can be used for control, and has, for example, the following characteristics.
(1) The masterless token method is adopted, and the network does not stop even if some nodes fail.
(2) If the token is not transmitted for a certain time, the next node transmits the token.
(3) The operation state of another node can be recognized. etc
Two data communication methods are supported: (1) cyclic transmission for performing periodic data transmission and (2) message transmission for performing aperiodic data transmission.
Cyclic transmission is a function that supports periodic data exchange that occurs between nodes.
[0005]
The token flow is basically determined by the node number assigned in advance to each node. For example, token rotation is performed in ascending order (in ascending order) from the smallest node number. In this case, the node with the largest node number passes the token to the node with the smallest node number.
[0006]
Further, each node includes a “token monitoring timer” as a configuration for detecting that a token is not transmitted for a certain period of time (2). This token monitoring timer measures the time since the last token flow on the network. When the token monitoring timer is up (time up), the next node reissues the token.
[0007]
Here, in the information processing field, as in a duplex system or the like, the reliability of the system is improved by duplicating devices (nodes) (active system / standby system).
[0008]
In the above network control system, it is conceivable to improve the reliability of the system by duplicating such nodes. However, when switching from the standby system to the active system, or the standby system / active system has not been decided yet. In an unstable situation, it is desirable to stably and promptly determine one of the duplicated nodes as an active system and the other as a standby system in a more appropriate manner.
[0009]
In addition, a program for performing such processing is complicated.
That is, in the past, variables such as “variable indicating the status of the active system” and “variable indicating the number of times the abnormal status of the active system continues” are prepared for the standby system, and transition to the active system is performed. In some cases, it was a common technique to refer to these variables, branch according to their values, and perform processing appropriate to that state. However, in this method, since processing is performed with reference to a plurality of variables in one state, a complicated structure with many branches is likely to be obtained, and processing such as variable updating is likely to be complicated. Also, since there was only a definition of steady state (active system, standby system) and unsteady state was not clearly defined, for example, even if it recognizes that its own state is a standby system, it is just a standby system. Since it may be in the middle of transition from the active system to the operating system (ie, unsteady state), even in the steady state, it will be judged every time whether it is now unsteady state. The amount of program description has increased. Therefore, it is desirable to simplify the program.
[0010]
An object of the present invention is to provide an operation / standby switching method in which switching of an active system / standby system or determination thereof can be performed stably and quickly in a network control system in which nodes are duplicated, and the program configuration is simplified. It is to provide an operation / standby determination method.
[0011]
[Means for Solving the Problems]
  In the operation / standby switching method according to the present invention, some or all of the nodes connected to the network are duplicated into an active system and a standby system.The system adopts FA link protocolAn operation / standby switching method for each redundant node in a network control system, in which a standby system transitions to an active system in addition to the standby system and active system state definitions.An operation ready token waiting state and an operation preparation token holding state are defined as non-stationary states, each node has a token monitoring timer, and the next node indicates that an arbitrary node does not transmit a token for a certain period of time. When detected by the time-up of the token monitoring timer, the next node reissues the token,AboveThe standby system of any node that has not transmitted a token for a certain time or more transitions to the operation ready token waiting state when observing the token reissued by the next node, and then again receives the token addressed to itself. When received, the state transits to the operation ready token holding state and uses the operation / standby switching timer set to a time shorter than the token monitoring timer to transit from the operation preparation token holding state to the active system or to the standby system. Decide what to do.
[0012]
In this way, by clearly defining the unsteady state (intermediate state), for example, in an unstable unsteady state where it is unknown whether to switch from the standby system to the active system, it is clear whether or not the transition is made. Since it is possible to transition to this intermediate state once until it is done, stable switching can be performed. In addition, since the transition to the intermediate state is clearly made in the unsteady state, it is not necessary to check whether the current state is the unsteady state in the steady state.
[0014]
As described above, the present invention is mainly directed to the network control system adopting the FA link protocol.
In addition, whether or not to switch from the standby system to the active system is determined by using the operation / standby switching timer set to a time shorter than the token monitoring timer, so that the token is reissued again from the next node by mistake. It can be done at high speed without In addition to the definition of the intermediate state, the effect of simplifying the program is increased.
[0021]
DETAILED DESCRIPTION OF THE INVENTION
The present invention provides a network control system having a configuration in which nodes constituting a network are duplicated. In this configuration, for example, when switching from a standby system to an active system, when starting a node, adding a new node, In the case of (unstable) unsteady state, such as in the case of duplexing, the steady state can be shifted to the steady state in an appropriate manner.
[0022]
In the present invention, the unsteady state is clearly defined as the operation ready state (intermediate state), and stable switching can be performed by temporarily transitioning to the intermediate state when switching from the standby system to the active system, for example. Furthermore, since the steady state and the unsteady state (transition process) can be clearly separated, only a description corresponding to each state is sufficient, and the program structure can be simplified.
[0023]
Embodiments of the present invention will be described below with reference to the drawings.
A network control system that employs the FA link protocol is referred to as an FA control network.
[0024]
FIG. 1 is a diagram showing an example of the system configuration of the FA control network according to the present embodiment. In the figure, the production system is shown as an example, but the present invention is not limited to this. The subject of the present invention is the entire control system.
[0025]
This FA control network 10 is a control device such as a robot controller (RC) 11, a programmable controller (PC) 12, a numerical controller (NC) 13 or a control personal computer 14 in a production system (hereinafter, these are not distinguished). This is a network control system in which control is performed by connecting data to each other via Ethernet and exchanging data with each other. The FA control network 10 can be further connected to an information network including an EWS, a personal computer, a server, etc. via a gateway or the like, or can be further connected to an external network.
[0026]
Each “node” constituting the FA control network 10 has a function of performing communication control of a token passing method (adopting the FA link protocol) (basic data using UDP / IP prevailing on Ethernet). Thus, the data exchange and the like are performed.
[0027]
In this example, each “node” is duplicated. That is, there are an active system and a standby system. However, not all nodes are duplicated. In addition, the entire device is duplicated. For example, when the PC is duplicated, two PCs having the same node number are connected to the network.
[0028]
FIG. 2 is a schematic configuration diagram of the node.
Each node 21 includes a buffer (reception buffer, transmission buffer) 22, a memory 23, a processing unit 24, a token monitoring timer 25, an operation / standby switching timer 26, etc. (required components are shown for explanation, and others are omitted). ) The buffer 22, the memory 23, and the token monitoring timer 25 are not particularly described (refer to the FA link protocol specification for details). However, the memory 23 may store a “number of receptions” described later and a basic set value of the timer 26. The processing unit 24 executes various processes described below. The operation / standby switching timer 26 will be described in detail later.
[0029]
FIGS. 3A to 3C are diagrams showing a simplified state of duplexing the nodes, and a state of switching and token transmission when a failure occurs in the active system.
In the figure, each “node” such as PC, RC, NC, personal computer, etc. in FIG. 1 is shown in a simplified manner without any particular distinction. Here, it is assumed that all nodes are duplicated, the active node is indicated by a solid line, and the standby node is indicated by a dotted line.
[0030]
The node numbers assigned to each node are # 1, # 2, and # 3, and for duplication, there are system A and system B, and in the example shown, the node of system A is the operating side. It shall be. Thus, in the following description, for example, the two nodes of node number # 2 are described as 2A for the node of system A and 2B for the node of system B, but the node numbers are # 2A and # 2B. In both cases, the node number is # 2 (that is, the active node and the standby node have the same node number).
[0031]
First, FIG. 3A shows a state before a failure occurs.
In this state, the token circulates in the order of node 1A → node 2A → node 3A... Node 1A →. The standby-system nodes 1B, 2B, and 3B monitor the token delivery, and determine that they can remain the standby system if tokens are normally delivered.
[0032]
Next, in FIG. 3B, an example in which the token is not transmitted from the node 2A after the token is passed from the node 1A to the node 2A due to some cause such as failure of the node 2A, for example. Show. In this case, the node 3A, which is the next node, reissues a token as shown in the figure when the token monitoring timer 25 provided therein is up. At this time, the node 2B on the standby side of the failed node 2 has transitioned to an unsteady state (intermediate state) (this will be described with reference to FIG. 4).
[0033]
Then, the token circulates again. First, the node 1A receives the circulated token and transmits it to the node 2 (as described above, both the nodes 2A and 2B have the node number # 2. Therefore, the node 1A does not distinguish between the two and passes the token), but the node 2B transitions from the intermediate state to the active system when the node 2A does not transmit the token even after a predetermined time elapses, A token is transmitted (details will be described with reference to FIG. 4).
[0034]
FIG. 4 is a state transition diagram of the node.
In the steady state, the active system is between the “active token waiting” 31 state and the “active token holding” 32 state, and the standby system is “standby token waiting” 33 and “standby token holding” 34. State transitions are made between these states.
[0035]
More specifically, when the active system receives a token in the state of “waiting for active system token” 31, it transits to the state of “hold active system token” 32, and when a token is transmitted in this state, “waiting for active system token” It returns to the state of 31.
[0036]
When the standby system detects that the active system has received a token in the state of “waiting for standby system token” 33, the standby system transits to the state of “holding standby system token” 34, and that the active system has transmitted a token in this state Is detected, the state returns to the “waiting for standby token” 33 state.
[0037]
In this example, a “operation ready token waiting” state 35 and an “operation preparation token hold” state 36 are further defined as unsteady states (intermediate states).
The state transition related to this intermediate state will be described later.
[0038]
Using the above state definition, each node 21 (particularly, the processing unit 24) executes the switching process to the active system and the active / standby system determination process described below.
First, switching processing to the active system will be described with reference to FIG.
[0039]
Assume that some kind of failure occurs in the active system (node 2A in the example of FIG. 3; hereinafter, the parentheses correspond to the example of FIG. 3).
As described above, the standby system (node 2B) enters the state of “holding standby system token” 34 when the active system receives a token addressed to itself, but the active system (node 2A) has a failure. As a result of the elapse of time without sending a token, the token monitoring timer 25 is up in the next node (time up). Therefore, the next node (node 3A) reissues the token.
[0040]
When the standby system (node 2B) observes this token (step S1), the standby system (node 2B) transitions to a state of “waiting for operation preparation token” 35 (step S2). Thereafter, when the token addressed to the own node is received again (step S3, YES), the state transits to the “operation ready token hold” state 36 (step S4). At this time, the operation / standby switching timer 26 is started (step S5). The set time of the timer 26 is shorter than the token monitoring time. When the timer 26 is up (step S6, YES), it is considered that the operating system has stopped, and the node (node 2B) transmits a token. (Step S7), transition to “wait for active token” 31 (Step S8). If the active system transmits a token before the active / standby switching timer 26 is up (for example, when the active system is restored) (step S9, YES), the process transits to “waiting for standby system token” 33 ( Step S10).
[0041]
As described above, in this example, when the active system does not transmit the token (that is, when there is a possibility that some abnormality has occurred), the standby system is not immediately switched to the active system, but temporarily When a transition is made to a state (unsteady state) and a token is received again, the operation / standby switching timer 26 is started, and if a token is not transmitted even after waiting for a while, a transition is made to the standby system (the active system is Considering a situation where the token was not sent once for some temporary reason, not a malfunction). The reason why the time of the operation / standby switching timer 26 is set to be shorter than the token monitoring time is, of course, to prevent the next node from transmitting a token again.
[0042]
Next, a standby / active system determination method will be described.
That is, either one of the two systems (system A and system B) constituting the duplex node is an active system and the other is a standby system, but a situation that has not yet been determined occurs. This will be described below.
[0043]
The operation / standby switching timer 26 is also used when determining the standby / active system.
In addition, while the active system and the standby system are operating in the joined state, the active node stores the token reception count in the memory 23 or the like (this count is equal to or greater than the number of rules). Will not be increased).
[0044]
FIG. 6 is a diagram illustrating an operation example of the operation / standby switching timer when determining the standby system / active system.
In the figure, the arrow toward the right side represents the flow of time.
[0045]
The set time of the operation / standby switching timer 26 is set shorter than the token monitoring time.
Further, the basic set value of the time-up time of the operation / standby switching timer 26 is set in advance so that there is a difference between the A system and the B system, and the time-up time of the actual operation / standby switching timer 26 is further increased. Varies depending on the number of times the token is received. However, as described above, the number of receptions does not increase beyond a certain number of times.
[0046]
In FIG. 6, as an example, the time-up timing of the operation / standby switching timer 26 of the A system and B system nodes when the number of token receptions is 0 (“0 times A system”, “0 times B system”). And the timing of time-up of the operation / standby switching timer 26 of the A-system and B-system nodes when the number of token receptions is the maximum (the above-mentioned specified number of times) ("regulated times A-system", " The position indicated as “regulated times B system”).
[0047]
As described above, since a time difference is provided between the A system and the B system, for example, as shown in the first example described below, when the number of token receptions is 0 for both the A system and the B system, In this example, the operation / standby switching timer 26 of the A system node is timed up earlier, so that the A system node becomes the active system, and both systems do not accidentally become the active system. Also, the A system node does not always become the active system. For example, when the number of tokens received by the A system node is 0 and the number of tokens received by the B system node is the specified number of times, As described above, this time, the operation / standby switching timer 26 of the B-system node is timed up first, so that the B-system node becomes the active system.
[0048]
Based on the above description, a first example and a second example will be described below for the standby / active system determination method.
First, the first example is a method for determining the standby system / active system when the A system / B system starts up at the same time.
[0049]
When the A system / B system starts up at the same time, both are once set in an operation ready state (or standby system), and are operated by using the operation / standby switching timer 26 with a time difference between the A system and the B system as described above. Determine the system and standby system. As a result, even if the A system / B system are started up at the same time, it is possible to prevent a situation in which both of them become the operating system by mistake.
[0050]
This will be described in more detail with reference to FIG.
When the A / B systems start up at the same time, both nodes once enter the “wait for operation preparation token” state 35 (step S11). Thereafter, the process shifts to “hold operation preparation token” 36 (step S12), sets the time-up time of the operation / standby switching timer 26, and starts (step S13). Then, it is determined whether it becomes a standby system or an active system depending on whether or not its own operation / standby switching timer 26 times out before the other party (step S14).
[0051]
As the set value of the time-up time of the operation / standby switching timer 26, different basic setting values are registered in advance in the A system / B system, and are changed from the basic setting values according to the number of token receptions ( Setting value = basic setting value−number of token receptions × α; α is an arbitrary coefficient).
[0052]
In this case, since both the A system and the B system are just started up, the number of token receptions is “0” in both cases, and therefore the operation / standby switching timer 26 of the A system node is earlier in the example of FIG. Thus, the basic set value is registered so that the time is up, so that the A-system node transitions to the active system after the token transmission (step S15) (step S16). The B-system node detects the token transmission and transitions to the standby system (step S17).
[0053]
Next, a second example will be described.
The second example is a standby / active system determination method when another node joins again after all nodes other than a certain duplex node have left the network.
[0054]
First, when all nodes other than a certain duplexing node leave the network, the duplexing node also enters a state in which no token is transmitted. Thereafter, when another node joins again, the network is constructed again.
[0055]
The processing contents at this time are basically the same as those in the first example (FIG. 7). That is, this duplex node first makes a transition to the “wait for operation preparation token” 35 state in both systems. Thereafter, the process proceeds to “hold operation preparation token” 36 and sets the time-up time of the operation / standby switching timer 26 to start. As described above, the time-up time of the operation / standby switching timer 26 for each node is shortened in accordance with the “number of receptions” (stored in the memory 23 of the node). The person who is there is very likely to time up first. For example, in the example of FIG. 6, if the “reception count” of the A-system node is 0 and the “reception count” of the B-system node has reached the specified number of times, The standby switching timer 26 times up earlier. In this case, the B-system node transitions to the active system after token transmission. The A-system node transitions to the standby system by detecting this token transmission.
[0056]
In this way, with a very high probability, a node that was active before the network is reconstructed can be made active and a node that was standby can be made standby. In other words, it is possible to maintain the consistency of the state where the active system continues to be the active system and the standby system continues to be the standby system.
[0057]
In the network control system of this example, even if, for example, both the A system and the B system become active due to malfunctions, the network control system can quickly and appropriately shift to a normal state. A malfunction is more likely to occur during an unsteady state than during a steady state. There are various situations in which malfunctions are likely to occur. An example will be described below.
[0058]
As an example of a situation in which a malfunction is likely to occur, for example, there is a time when a node that has not been duplicated is duplicated. For example, in FIG. 3, it is assumed that only the node 1A has initially joined the network and the node 1B has not joined. In this state, the node 1A naturally transmits and receives tokens as an active system. Then, when the node 1B joins later, both have the node number # 1 as described above, so that both may become active due to malfunction or the like.
[0059]
In this case, when it is detected that both nodes 1A and 1B are in the active system, the state temporarily shifts to the “wait for operation preparation token” 35 state. Thereafter, the state transits to the “operation ready token hold” state 36 and the operation / standby switching timer 26 is started. In such a case, as described in the second example above, basically, the one that was first in the active system (node 1A) timed up first, so the node 1A transitioned to the active system after transmitting the token. To do. The node 1B joined later detects this token transmission and transitions to the standby system.
[0060]
Next, referring to FIG. 8, a description will be given of a case where A system and B system are joined to different networks to transmit and receive tokens, and then the two networks are combined to form one network. To do.
[0061]
Conventionally, the bus network type is mainly the bus type, but the star topology is mainly used at present, and the form using the switching HUB is generally used. Therefore, the form using the switching HUB is shown in FIG. As an example.
[0062]
In the example of FIG. 8, the first network in which the nodes 1A, 2 and 3 are connected to the HUB 41, and the second network in which the nodes 1B, 4 and 5 are connected to the HUB 42 are separately provided. It shall exist. In any network, the nodes are not duplicated.
[0063]
In such a configuration, in the first network, the token circulates as node 1A → node 2 → node 3 → node 1A →..., And in the second network, node 1B → node 4 → node 5 → node 1B. → ... and so on. However, since both the node 1A and the node 1B have the node number # 1, each node recognizes that “node 1” exists.
[0064]
From such a state, as shown by the dotted line in the figure, when the HUB 41 and the HUB 42 are connected and the two networks are formed as one network, for example, the node 5 → node 1B (the same applies to the node 3 → node 1A). When the token is sent out, as described above, it is not distinguished as node 1A / node 1B, but is recognized as “node 1”. Therefore, both node 1A and node 1B have tokens addressed to themselves. As received, tokens such as node 1A → node 2 and node 1B → node 4 are transmitted. Then, both the node 1A and the node 1B recognize the token issued by the partner station, and confirm that “a node other than the own station is assigned the same node number # 1 and this is currently in operation”. Recognize (In the FA link protocol targeted in this example, all tokens are transmitted by broadcast. That is, each node recognizes the existence of all tokens flowing on the network. When the “destination node number” in the data matches the node number of the local station, it is determined that the token addressed to the local station is received (transmission right is acquired).
[0065]
As described above, when it is detected that there are two nodes having the same node number in the active system, each node executes processing described below. This process is basically the same as the process described in the first example.
[0066]
First, both the node 1A and the node 1B make a transition to the “wait for operation preparation token” state 35 once. Thereafter, the state transitions to the “hold operation preparation token” state 36 and the operation / standby switching timer 26 is set and started.
[0067]
Before the connection between the HUB 41 and the HUB 42, since both the node 1A and the node 1B were operating as active systems, when it is considered that the “number of receptions” has reached the specified number of times, the example of FIG. When used, the node 1A times up first, and after token transmission, transitions to the active system. The node 1B detects this token transmission and transitions to the standby system.
[0068]
In this way, in this example, even if two nodes having the same node number become both active due to some cause, the active / standby switching timer 26 is used to determine the active / standby system. Therefore, it is possible to avoid an unstable state where both the A system and the B system are operating systems for a long time.
[0069]
【The invention's effect】
As described above in detail, according to the operation / standby switching method and the operation / standby determination method in the network control system of the present invention, in addition to the definition of the conventional steady state (active system, standby system), unsteady state Is clearly defined (operation ready state), so that the steady state and the unsteady state can be separated, stable switching can be performed, and the program configuration can be simplified.
[0070]
In addition, an operation / standby switching timer is provided, and by using this operation / standby switching timer for operation / standby switching or operation / standby determination, the operating system / standby system is determined quickly, and there is no need for both systems to operate. It is avoided that a stable state continues. Also, by changing the timer up time of the operation / standby switching timer according to various conditions, what was formerly the active system will continue to be the active system, and what was the standby system will continue to be the standby system State consistency is maintained.
[Brief description of the drawings]
FIG. 1 is a diagram showing an example of a system configuration of an FA control network according to the present embodiment.
FIG. 2 is a schematic configuration diagram of a node.
FIGS. 3A to 3C are diagrams schematically showing a state in which the nodes are duplicated, and a state of switching and token transmission when a failure occurs in an active system. FIGS.
FIG. 4 is a state transition diagram of a node.
FIG. 5 is a flowchart for explaining an example of switching processing from a standby system to an active system;
FIG. 6 is a diagram illustrating an operation example of an operation / standby switching timer when determining a standby system / active system.
FIG. 7 is a flowchart for explaining standby / active system determination processing when an A system / B system is started up simultaneously;
FIG. 8 is a diagram for explaining a case where two networks are combined to form one network.
[Explanation of symbols]
10 FA control network
11 Robot controller (RC)
12 Programmable controller (PC)
13 Numerical controller (NC)
14 PC for control
21 nodes
22 buffers (reception buffer, transmission buffer)
23 memory
24 processor
25 Token monitoring timer
26 Operation / standby switching timer
31 “Waiting for active token” (status)
32 “Holding active token” (status)
33 “Waiting for standby token” (status)
34 “Standby token holding” (status)
35 “Waiting for operation preparation token” (status)
36 “Operation ready token hold” (status)
41 HUB
42 HUB

Claims (1)

A part or all of the nodes connected to the network is a system in which the active system and the standby system are duplexed, and the operation / standby switching method of each duplex node in the network control system adopting the FA link protocol ,
In addition to the standby system and active system state definitions, the operation ready token waiting state and the operation ready token holding state are defined as non-stationary states in the process of transitioning the standby system to the active system .
When each node has a token monitoring timer and the next node detects that any node does not transmit a token for a certain time or more due to the timeout of the token monitoring timer, the next node reissues the token. ,
The standby system of an arbitrary node that has not transmitted a token for a certain time or more transitions to the operation ready token waiting state when observing a token reissued by the next node, and then again a token addressed to itself Is transferred to the operation ready token holding state, and using the operation / standby switching timer set to a time shorter than the token monitoring timer, the operation preparation token holding state is changed to the active system or to the standby system. An operation / standby switching method characterized by determining whether to make a transition .

Images