JPS61118044A - System for detecting and controlling fault of repeat function - Google Patents

Abstract

PURPOSE:To detect a fault of a repeat function caused in a node by providing a recovery number count means of a free token in each node. CONSTITUTION:When a node N0 is in the active monitor state having a free token recovery function, a TPC5 supervises a reception interval of the free token, and when the next free token is not received for a prescribed time, an interruption is given to a microprocessor 8 to start a free token recovery control section 9. The free token recovery control section 9 executes a free token recovery sequence to cause the TPC5 to transmit a new free token and increments a number of time counter 10 by 1. A limit detection section 11 compares the count of the counter 10 with a predetermined limit value and when the count exceeds a limit value, it is informed to a status transition section 12, from which the monitor is transited from the active monitor state to an active monitor give-up state.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a communication control system in a ring network such as a LAN, and in particular, detects a repeat function failure occurring in a node and This invention relates to a failure detection control method for automatically replacing the sending node of a computer.

[Conventional technology]

FIG. 2 shows the system configuration of an example of a conventional ring network to which the present invention is directed. In the figure, LNo, Ll'L+,...
LN7 is an automatic configuration control device, No, N+,...
N7 represents a node. Insertion into and removal from the ring, that is, network configuration, of each node is controlled by a higher-level automatic configuration control device.

In the example of Fig. 2, nodes No., N+, NZ
, N3 have their configurations controlled by an automatic configuration controller LN, and for example, as shown in FIG.
,,N, are inserted into the network.

Node N, has been bypassed and removed from the network.

In a ring network, it is necessary to prohibit data transmission from two or more nodes at the same time.

One method for this purpose is a token passing method in which free tokens are circulated around the ring and the transmission right is given to the node that first captures the free token among nodes that have a transmission request.

FIG. 4 is a configuration diagram of a communication control adapter used in a node of a ring-shaped network element using a token passing method. In the figure, 41 is a communication control adapter, 42
is the token passing controller TPC (hereinafter TP
43 is a transmitting/receiving buffer, 44 is a microprocessor (μmP), 45 is a ROM, 46 is a RAM,
47 is BUS. The basic operating functions will be described below.

The TPC 42 takes in received data from the ring side and sends it to the host device via the transmission/reception buffer 43.

In other words, it is sent to the CPU, and if it is data to be transferred to another node, it is output as transmission data to the ring side using the repeat function, and if there is data addressed to the own node when in reception mode, it is sent to the CPU. Import it with.

On the other hand, if the Fleet-Kun is captured when there is a data transmission request from the CPU of the host device, ! 81 nodes acquired the transmission right.

A frame is assembled from the data in the transmission (K/reception buffer 43) and outputted to the ring side as transmission data. These controls are performed by the microprocessor μ-244.

FIG. 5 shows the internal configuration of the TPC 42 in FIG. 4. In the figure, 51 is a repeat circuit for transferring data not received by the own node to the lower node in FIG. 4, 52 is a receiving circuit, 53 is a transmitting circuit, and 54 is a TPO.
The whole is controlled by a microprocessor μ-P44 (Figure 4).
This is a control circuit that also serves as an interface with the

FIG. 6 further shows the internal configuration of the repeat circuit 51 in FIG. 5. In FIG. is the transmission shift register TSR.

Since the data transmitted on the ring is in a bit serial format, the received data is transferred to the receive shift register R3R6.
1, 9 bits are converted into parallel data for each frame detected and stored in the buffer BFA62.

The data in the buffer BFA62 is sent to the receiving circuit 52 in FIG.
is sent to the CPU, but at the same time the buffer BFE
63 or buffer BFO64. The buffer BFE 63 and the buffer BFO 64 are capable of performing write and read operations simultaneously and alternately. During repeat operation,
The data alternately written in these buffers BFE63 and BFO64 are continuously and repeatedly transferred by alternately reading them from the opposite side.

The transmission shift register TSR 65 is used to convert 9-bit parallel data into bit serial data, and is used to convert data output from the transmission circuit 53 in FIG. 5 when transmitting data from its own node, or from buffer B during repeat operation.
Parallel data read from FE63 or BF064 is converted into bit serial data and sent to the ring side.

As mentioned above, in a ring-shaped network using the token passing method, fleets circulating on the ring are
The right to transmit is acquired by capturing Kun, but
One node is always given the function of reproducing Fleet Tokens, and if that node cannot receive Fleet Tokens for a certain period of time, it reproduces new Fleet Tokens,
It is designed to circulate around the ring.

The state of a node having such a free token regeneration function is called an active monitor state, and the state of other nodes is called a passive monitor state.

If a failure occurs in a node in the active monitor state, it temporarily shifts to the active monitor abandonment state, and during that time one node is automatically selected from the remaining nodes in the passive monitor state.

Move to active monitor state (i.e. recover)
. In order to remove the light, nodes that have been placed in an active monitor abandonment state are moved to a passive monitor state. This is to prevent the failed node from being elected as the active monitor state node again. In addition, nodes that are not selected remain in the passive monitor state. FIG. 7 is a state transition diagram.

[Problem that the invention seeks to solve]

By the way, if a repeat function failure occurs in a certain node within the ring, normal repeat transfer of fleet tokens will be interrupted. Previously, if a repeat function failure occurred in a node in a passive monitor state, it was possible to detect it and disconnect the node from the ring.

However, if a repeat failure occurs in a node that has a Free Token regeneration function, that node will continue to regenerate Fleet Tokens at regular intervals, and the problem is that the node will never notice that the repeat function has failed. was there.

[Means for solving problems]

In order to solve the above-mentioned problems, the present invention provides each node with means for counting the number of playbacks of free tokens.
A node with a free token playback function uses this counting means to monitor the number of times a free token has been played, making it possible to detect repeat failures.
In a ring communication mechanism using a token passing method, the ring communication mechanism is composed of one or more nodes, at least one automatic configuration control device having a function of bypassing the nodes, and a branch line connecting the nodes and the automatic configuration control device. , a node with the function of reproducing Fleet Tokens counts up the number of times the Fleet Tokens have been played, and when this reaches a certain value, it temporarily stops the Fleet Token regeneration function of the node and The feature is that the nodes that play Kun are alternated.

〔Example〕

The details of the present invention will be explained below based on examples.

FIG. 1 is a schematic diagram of one embodiment of the present invention, and particularly shows the main part configuration of a communication control adapter. In the figure, 1 is the automatic configuration control device LNo.

2 is a node N0, 3 is a CPU of a host device, 4 is a communication control adapter, 5 is a TPC ()-Kun passing controller), 6 is a status control unit, 7 is a transmission/reception buffer, 8 is a microprocessor μ-P, 9 is a free token reproduction control section, 10 is a number counter, 11 is a limit detection section, 12 is a status transition control section, 13 is a ROM, and 14 is a RAM.

15 represents BUS.

l automatic configuration control device 1. Nodes N0 and 2 correspond to LN and No in the ring network shown in FIGS. 2 and 3, respectively. Here node N. Although only the configuration of the communication control adapter 4 is shown, the communication control adapters of all other nodes have similar configurations.

The basic functions of the communication control adapter 4 are shown in Figures 4 to 7.
Although it is almost the same as the conventional one explained in the figure, it especially has a function of detecting and controlling repeat failure based on the present invention. A number counter 10 and a limit detector 11 shown in the microprocessor μ-P of 8 are used as the main means for this purpose.

The status control unit 6 of the TPC 5 controls the control state, that is, the status of the node. The details will be described later, but there are three major states: active monitor state, active monitor abandonment state, and passive monitor state, and within each of these states, there are further detailed states, namely, repeat state and token holding state.

There are a token sending state, a token reproducing state, a monitor recovery state, a beacon sending state, etc.

In the microprocessor μ-P8, the fleet token reproduction control section 9 is functionalized only when the node is in the active monitor state, and controls the reproduction and transmission operation of the fleet token. A number counter 10 counts the number of times a free token is played. The limit detection unit 11 monitors the count value of the number counter IO, and determines that a repeat failure has occurred when the limit value is exceeded. This limit value is, for example, 1 to
Set the number of times equivalent to 2 seconds. The status transition control unit 12 controls state switching of the status control unit 6 of the TPC 5. In particular, when the limit detection reactor 11 determines that a repeat failure has occurred, it causes the status control unit 6 to transition the active monitor state of the own node to the active monitor abandonment state.

When one of the other nodes establishes the active monitor state, it also transitions to the passive monitor state.

Next, we will discuss specific operations.

First, it is assumed that the node N0 is currently in an active monitor state with a free token regeneration function and the 2-times counter 10 has been cleared. In this state, the TPC 5 monitors the reception interval of fleet tokens, and if it cannot receive the first fleet token for a predetermined time (T1), it issues an interrupt to the microprocessor (μmP) 8. , activates the free token reproduction control section 9.

The free token playback control section 9 executes a free token playback sequence and causes the TPC 5 to send out new free tokens. At the same time, the counter 10
Amplify the count by 1.

In this way, each time 5 free tokens are reproduced, the count value of the number counter 10 increases by one. The limit detection section 11 compares the count value of the number counter 10 with a predetermined limit value, and notifies the status transition control section 12 when the count value exceeds the limit value.

Causes a transition from the active monitor state to the active monitor abandonment state.

Next, state transition in this embodiment will be explained. FIG. 8 is a state transition diagram, and FIG. 9 shows the correspondence between the states of each status FF provided in the status control unit 6 that controls each state and its control events.

In FIG. 8, state transitions are roughly divided into the following three monitor states.

Active monitor state (1) Active monitor abandonment state (n) Puffy monitor state (III) Each of the above monitor states is further divided into a normal state and an abnormal state.

First, under normal conditions, N), (n),
A node in any of the monitor states (■) undergoes a transition from repeat state (■) to token holding state (■) to token sending state (■) to repeat state (■). The operation in the normal state will be explained below.

A. In normal state repeat state (2), the received fleet token is transmitted as is. However, when a transmission request is made to the own node, upon receiving the Flinidokun, the node captures it and moves to the token holding state (2).

In the token holding state ■, the node acquires the transmission right and performs a frame sending operation. After sending the frame, the state moves to the token sending state ■.

In the token sending state (2), a free token sending operation is performed, and when the free token is sent out, the state returns to the repeat state (2).

Next, each monitor state (1), (II), (II
The operation in an abnormal state at the node [) will be explained.

B. Abnormal condition? In the repeat state of the node in (1), if neither the fleet token nor the busy token can be received for a certain period of time T,
T+ Timeout causes an interrupt to the microprocessor 8, and the state shifts to the token regeneration state (2).

In the token regeneration state■, the ring barge frame R performs the free token regeneration operation and first clears the inside of the ring.
Sends PF and moves to token sending state ■. At this time, the number counter 10 is incremented by one. If a free token is sent in the token sending state ■, the state returns to the repeat state ■. If no token is received again at time T2, the state shifts to the token regeneration state (2), and the above-described operations are repeated. When the count value of the counter 10 exceeds the sum and g values once in the token regeneration state (2), this node moves to the repeat state (2) of the active monitor abandonment letter u (n).

Active monitor abandonment letter G (II) In this repeat state ■, the time T2 is greater than T.
When a token cannot be received during T2, T2 times out, an interrupt is applied to the microprocessor 8, and the state shifts to the beacon sending state (2).

In the beacon sending state (2), a beacon frame BCF is sent out, and if the BCF can be received as a result, the state returns to the repeat state (2). Note that automatic configuration control device LN. The failure recovery process is performed by monitoring the beacon frame BCF, disconnecting the BCF sending node, and if it does not recover, disconnecting the previous node as an abnormal node.

As in the above case, active monitor abandonment letter Nu(I
If there is no abnormality in the node I), it receives a ring purge frame RPF sent by another node (SA≠MA) in its repeat state ■.

Assuming that the replacement of the active monitor state node with the free token regeneration function has been completed, the puffive monitor state g
Move to repeat state (i) (■).

B21] Sanshichiguchi (I Engineering) Each node in the passive monitor state monitors frame reception at T2 for one hour, and when T2 times out, shifts to the monitor recovery state (2).

In this monitor recovery state ■, an operation is performed to select a new active monitor state node.

Each node in the puffive monitor state sends out a monitor recovery frame MRF in one order after the T2 timeout occurs. At this time, the address MA of the own node is written into the MRF as the source address SA. It is assumed that the node with the larger SA is at the higher level.

Next, each node receives the MRF that has made one round.

The source address SA is compared with the own node address MA, and if SA=MA, the MRF transmission is retried, and the S
If A > MA, it is determined that there is another higher-order node that should become an active monitor state node, and the process returns to repeat state (2).

If SA<MA, the MRF is transferred to the next node. In this way, the lower nodes are weeded out one after another, and the highest node remains at the end. This node is S A = M
A, and under this condition, the active monitor state (I) moves to the repeat state (■). In other words, the selection of a new active monitor state node is completed.

Note that the node that selected the MRF in monitor recovery state ■ monitors the time it takes for its own MRF to complete one cycle and return, and if the MRF does not return for a certain period of time or more and the MRF cannot be retried, the node will issue an MRF retryout. Moves to beacon sending state ■, sends a beacon frame BCF, causes the automatic configuration control device to perform fault recovery processing for disconnecting the abnormal 4-board, and confirms that there is no abnormality in the own node BCF
If received, it returns to the repeat state.

As described above, when an abnormality occurs in the reception of fleet tokens by a node in the active monitor state, the active monitor state node is automatically replaced, and it becomes possible to disconnect the node having a failure in the 5-repeat function.

Next, each state FF in the state control section shown in FIG. 9 will be explained. In the figure, nine status FFs indicated by 91 to 99 are in the transition states (1), (II), (III), ■, shown in FIG. 8, respectively.
In response to ■,■,■,■,■.

Used to display current status.

Status FF9L indicates the active monitor state (mouth). This state is sent when an MRF with SA=MA is received, and is reset when the counter value minus the limit value is detected.

Status FF92 is active monitor abandonment letter v (
II). This state is set by the detection of the counter value=limit value and is sent when the SAeMA RPM is received.

The status FF 93 displays a passive monitor type G (■). This state is sent when an RPF with SA≠MA is received and reset when an MRF with SA=MA is received.

The status FF 94 displays a repeat status ■. This state is entered at initialization or by token sending, and is reset by each token capture T1 timeout Tz timeout.

The status FF95 displays the token holding state ■. This state is cleared by token capture and reset by frame sending.

The status FF 96 displays the token sending status ■. This state is sent by sending a token or sending an RPF, and is reset by sending a token.

The status FF97 displays the token playback status ■. This condition is sent by T1 timeout and R
Reset by sending PF.

Status FF98 is in beacon sending state, and MR
It is sent at the time of F transmission retryout or when a disconnection is detected, the explanation of which is omitted here, and is reset by BCF reception.

Status FF99 displays the monitor recovery state. This condition is set by T2 timeout and S
A>MA's MRF reception, SA=MA's MRF reception, MR
Reset by each F transmission retryout.

〔Effect of the invention〕

As described above, according to the present invention, a node with a failure in the repeat function repeatedly plays back Fleet-Kun, and the playback is stopped after a certain number of times, and the node having the Fleet-Kun playback function is transferred to another node. Therefore, it is possible to prevent a state in which free tokens are repeatedly played back without noticing the failure of the repeat function for a long time.

[Brief explanation of the drawing]

FIG. 1 is a schematic configuration diagram of one embodiment of the present invention, FIG. 2 is a configuration diagram of a ring network, FIG. 3 is an explanatory diagram of connections between an automatic configuration control device and nodes, and FIG. 4 is a diagram of a communication control adapter. The configuration diagram, Figure 5, shows the token passing controller (T
Figure 6 is the configuration diagram of the repeat circuit, and Figure 7 is the configuration diagram of the repeat circuit.
8 is a detailed state transition diagram in the embodiment of the present invention, and FIG. 9 is an explanatory diagram of the status FF of the status control unit in the embodiment of the present invention. In the diagram, is 1 on the automatic configuration side? B device No. 2 is node N
0 and 3 are CPUs of host devices, 4 is communication control adapter, 5
is a token passing controller (TPC), 6 is a status control unit, 7 is a transmitting/receiving buffer, and 8 is a microprocessor (μ-P). 9 is a free token playback control unit; 10 is a number counter;
11 is a limit detection section, 12 is a status transition control section,
13 represents ROM, 14 represents RAM, and 15 represents Bus.

Claims (1)

[Claims] In a ring communication mechanism using a token passing method, which is constituted by at least one or more nodes, at least one automatic configuration control device having a function of bypassing the node, and a branch line connecting the node and the automatic configuration control device. , a node with the function of reproducing free tokens counts up the number of times the free token has been regenerated, and when this reaches a certain value, the node's free token regeneration function is temporarily stopped and the free token is regenerated. A repeat function failure detection control method characterized by replacing nodes.