JP2822942B2 - Token failure detection method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a token failure detection method, and more particularly to a token failure detection method for estimating a lost position of a token for controlling the right to use each system bus constituting a duplex ring network.

[0002]

2. Description of the Related Art In a ring network, since a plurality of modules (nodes) cannot use a system bus (medium) at the same time, a special frame called a token (TKN) is used as a method for reserving the use of the system bus. Is transferred between the modules, and a token passing method is used in which an arbitrary module that receives this can reserve the use of the system bus.

Therefore, if this token is lost for some reason, the right to use the system bus cannot be controlled. Therefore, the token is always detected by an arbitrary module. (TKN TO), it is determined that a token failure has occurred and the token has been lost, and a new token is generated.

Conventionally, this kind of token failure detection method has
The configuration is as shown in FIG.
No. 2-175044). In FIG. 1, reference numeral 1 denotes a master module for controlling switching of a system bus 0 system (working system) 3 and system bus 1 system (standby system) 4; and 2, a data exchange with an arbitrary module via each system bus. Thus, n slave modules # 1 to #n that execute a predetermined process. In particular, in this network, the system bus is composed of system bus 0 and 1
The system is duplicated, and if a failure occurs in one of them, it is switched to the other and used.

In the master module 1, reference numeral 10 denotes a master processor having a token timeout counter 13 for counting the number of token timeouts generated in the system buses 3 and 4, and reference numeral 11 denotes a system bus 3 or 4 and a master processor 10 based on a token passing method. And a system bus adapter (SB) having a token timeout detection circuit 12 for transmitting and receiving data between the
A-M), and are provided in duplicate with respect to the system buses 3 and 4, respectively.

In the slave module 2, reference numeral 20 denotes a slave processor which is a higher-level device for performing predetermined processing by exchanging various data via the system buses 3 and 4, and 21 denotes a system bus 3 or 4 based on a token passing method. And a slave processor 20 for transmitting and receiving data and detecting a token timeout.
A system bus adapter (SBA-S) having
Each is provided in duplicate with respect to the system buses 3 and 4.

Now, a system bus (0 system) 3 is used as an active system.
When the token timeout is detected by the token timeout detection circuit 12 of the master module 1, the master processor 10 counts the number of detections by the token timeout counter 13 of the corresponding system bus 3. Therefore, when the number of times of detection becomes equal to or more than a specified value set according to the network scale or the like, it is determined that the currently used system bus 3 has a fault, and the other system bus 4 is switched. Had to be controlled.

[0008]

Therefore, in such a conventional token failure detection method, it is determined whether or not a token failure has occurred based on the number of token timeouts detected by the token timeout detection circuit 12 of the master module 1. Therefore, there is a problem that it is not possible to estimate the location where the token failure has occurred and the location where the failure has occurred.

A method of confirming whether or not a token timeout of the active system bus 3 in each slave module 2 has been detected by the master processor 10 via the standby system bus 4 can be considered. Since the token timeout is sequentially detected in all the modules after the module in which the timeout is detected, it is difficult to accurately estimate the location of the failure that occurs not fixedly but intermittently or suddenly. An object of the present invention is to solve such a problem, and an object of the present invention is to provide a token failure detection method capable of accurately estimating a place where a token loss failure has occurred.

[0010]

In order to achieve the above object, a token failure detecting method according to the present invention provides a method for detecting a token failure in each module when a token is not detected from a system bus used as an active system for a predetermined time or more. And a token timeout counter that counts the number of token timeouts detected by the token timeout detection circuit. Each module detects occurrence of a token timeout by the token timeout detection circuit. and detects, counts the detection number of tokens timeout by the token timeout counter, the master module if the token number of timeouts detected by the own device, reaches a predetermined allowable number of times or more, the primary system
From one system bus to the other
After switching to the system bus, it became a new working system.
Transfer the token to the other system bus and
At the same time, each module is sent from the token timeout counter of all modules via the active system bus.
The number of token timeouts detected by the module is collected, and the collected token timeout counts are compared with each other to estimate the failure location where the token is lost.

Therefore, when the number of token timeouts detected in the master module becomes equal to or more than the predetermined allowable number, the master module newly adds the current working time.
Transfer the token to the system bus
To it from the token time-out counter of all modules via the system bus, which became the new active system
The number of token timeouts detected by each module is collected, and the collected token timeouts are compared with each other to estimate the failure location where the token is lost.

[0012] Also, the master module determines the number of token timeouts collected from all modules.
The failure occurrence position where the token is lost is estimated by comparing the token timeout times collected from two adjacent modules. Therefore, the adjacent 2
The number of token timeouts collected from the two modules is compared with each other, and the fault location where the token is lost is estimated.

[0013]

Next, the present invention will be described with reference to the drawings. FIG. 1 is a configuration diagram of a duplex ring network using a token failure detection method according to an embodiment of the present invention. In the drawing, the same or equivalent parts as those described above are denoted by the same reference numerals.

In the slave module 2, reference numeral 25 denotes a token timeout (TKN T.O) for transmitting and receiving data between the system bus 3 or 4 and the slave processor 20 based on a token passing method and detecting a token timeout. Detection circuit 2
6, and a system bus adapter (SBA-S) having a token timeout counter 27 for counting the number of token timeouts detected by the token timeout detection circuit 26.
4 is provided in duplicate.

Next, the operation of the present invention will be described with reference to FIG. FIG. 2 is an explanatory diagram showing a failure point detection method, and particularly shows only a portion related to the system bus (0 system) 3 used as the active system. A master module 1 and n slave modules 2 are connected to the system bus 3 in a ring shape. At normal times, tokens are transferred on the system bus 3 and the right to use the system bus 3 is controlled.

Here, if any failure occurs at the point P, the token transmitted from the slave module # k-1 is lost at the point P. Accordingly, the slave module #k adjacent on the downstream side cannot receive the token, and has transmitted the token in the past in all the modules. The token timeout is detected by the token timeout detection circuit 26.

Next, the token timeout detecting circuit 26 of the slave module # k + 1 detects the token timeout, and similarly, the slave module # k + 2
(Not shown),..., Slave module #N, master module 1, slave module # 1,.

Further, in the token timeout circuit 26 of each module, after the detection of the token timeout, a predetermined time is set again to wait for the detection of the token timeout. Therefore, after the token timeout is detected by the slave module # k-1, the second token timeout is sequentially detected again from the slave module #k.

The number of times of these token timeouts is counted by a token timeout counter 27 in the system bus adapter 25 of each slave module 2. On the other hand, the number of token timeouts in the master module 1 is also counted by the token timeout counter 13 in the master processor 10.

Here, the allowable number m _th of token timeouts is set in advance in the master processor 10 as a criterion for system bus switching, and the master processor 10 determines the allowable number m _th and the count value of the token time counter 13. m, and when m becomes equal to or greater than m _th , switching from the currently used system bus to the standby system bus is performed. In this case, switching from the system bus 3 to the system bus 4 is performed. Do,
A new token is generated and transmitted to the system bus 3.

At the same time, the master processor 10
By controlling the system bus adapter (SBA-M) 11,
By transmitting a token timeout read command as shown in FIG. 3A from the system bus 4 to each slave module 2, the token timeout counter 27 of each slave module 2 is transmitted from all slave modules # 1 to #n. Collect the value of Note that FIG.
Are based on general HDLC (High Level Data Link Control), and detailed description is omitted.

Accordingly, the system bus adapter 25 on the system bus (system 1) 4 side of each slave module 2
Is the system bus (system 0) 3 used as the active system
The count value is read from the token timeout counter 27, and the master module 1 is notified of the number of token timeouts detected in its own station by the token timeout number response status as shown in FIG.

As described above, the master module 1
The number of token timeouts is collected from all the slave modules 2 and these values are compared with each other to estimate the location where a failure has occurred. First, as shown in FIG.
The m-th token timeout is the slave module #
k, a new token is transmitted to the system bus 3 when the m-th token timeout is detected in the master module 1, so that the slave modules # 1 to # k-1 downstream from the master slave The m-th token timeout is no longer detected.

Therefore, when the token timeout counts collected from all the slave modules 2 are compared, the token timeout counts of the slave modules #k and slave # k-1 on both sides of the point P where the failure has occurred are respectively different. m and m−1, and a difference occurs. By examining the difference, it is possible to estimate which module has a fault.

Thereafter, the system bus adapter 11 is controlled by the processor 10 of the master module 1,
A token timeout count clear command as shown in FIG. 3 (c) is transmitted by broadcast to the system bus 3, and in response to this, the system bus (0
(System) The count value of the token timeout counter 27 on the 3 side is cleared, and the counting operation of the new token timeout detection count is started.

When a failure occurs between the master module 1 and a slave module #n adjacent to the master module 1, a token timeout is first detected by the master module 1, so that the master module 1
Is larger than the other slave modules 2. Therefore,
If the number of token timeouts m detected by the master module 1 becomes greater than or equal to the allowable number of times m _th, and is larger than the token timeout of the other slave module 2, the master module 1 and its upstream adjacent slave module It is presumed that there is a failure occurrence point between #n.

When a failure occurs between the master module 1 and the slave module # 1 adjacent to the master module 1, the slave module # 1 first detects a token timeout, so that the master module 1 detects a token timeout. The number of token timeouts detected in 1 becomes equal to that of the other slave modules 2. Therefore, the token timeout count m detected by the master module 1
Is equal to or greater than the token timeout of the other slave module 2 when the number of times becomes equal to or more than the permissible number m _th , there is a failure location between the master module 1 and the slave module # 1 adjacent to the master module 1 on the downstream side. It is estimated to be.

As described above, a token timeout detecting circuit and a token timeout counter for storing the detected number of token timeouts are provided in all the modules constituting the network so that the number of token timeouts detected by the master module 1 is equal to the allowable number of times. In this case, the number of token timeouts is collected from all slave modules 2 and the location of the failure is estimated by comparing these token timeouts. Not only that, even if a token loss failure occurs intermittently or suddenly due to noise or the like, the failure occurrence location is estimated.

In the above description, a case has been described in which the number of token timeouts of adjacent modules is compared, and a failure occurrence location is estimated based on the difference between the number of token timeouts. However, the present invention is not limited to this. Instead, the one in which the number of token timeouts in these adjacent modules suddenly changes, the one in which the difference is the largest, or the one in which the difference is larger than a predetermined value may be used as a criterion for failure occurrence point estimation. Alternatively, all the token timeout times may be compared, and the location of the failure may be estimated based on the largest value.Using the criterion according to the occurrence situation of the token failure in each network, The same operation and effect can be obtained.

[0030]

As described above, according to the present invention, each module is provided with a token timeout detection circuit and a token timeout counter for counting the number of detected token timeouts. The number of token timeouts detected from the system bus is counted, and if the number of token timeouts detected in the master module is equal to or greater than a predetermined allowable number, the master module causes the token timeout counters of all modules to return to the respective devices. The number of token timeouts detected in step (1) is collected, and the collected token timeout counts are compared with each other to estimate the failure occurrence location where the token is lost.

Therefore, it is possible to estimate the location where the failure has occurred not only in the case of a fixed failure due to disconnection of the system bus, but also in the case of a token loss failure that occurs intermittently or suddenly due to noise or the like. It becomes possible.
In addition, the master module estimates the fault occurrence location where the token is lost by comparing the token timeout counts collected from two adjacent modules among the token timeout counts collected from all modules. Thus, it is possible to more accurately estimate the location where a failure has occurred.

[Brief description of the drawings]

FIG. 1 is a configuration diagram illustrating a duplex ring network using a token failure detection method according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing a token failure detection method.

FIG. 3 is an explanatory diagram showing various commands and status.

FIG. 4 is a configuration diagram illustrating a duplex ring network using a conventional token failure detection method.

[Explanation of symbols]

1: Master module, 10: Master processor, 11
... System bus adapter (SBA-M), 12 ... Token timeout detection circuit, 13 ... Token timeout counter, 2 ... Slave module, 20 ... Slave processor, 25 ... System bus adapter (SBA-M)
S), 26: token timeout detection circuit, 27: token timeout counter, 3: system bus 0 system,
4: System bus (system 1)

Continuation of front page (56) References JP-A-4-129350 (JP, A) JP-A-4-35251 (JP, A) JP-A-4-14934 (JP, A) JP-A-3-195146 (JP) JP-A-3-11841 (JP, A) JP-A-2-224545 (JP, A) JP-A-2-78344 (JP, A) JP-A 8-307438 (JP, A) (58) Field surveyed (Int.Cl. ⁶ , DB name) H04L 12/42-12/437

Claims (2)

(57) [Claims] 1. A master module for performing switching control of two independent ring-shaped system buses for transferring various types of information in predetermined directions, and one of these system buses as an active system and the other as a standby system as required. And a plurality of slave modules that perform arbitrary processing by exchanging various kinds of information via each of the system buses. As an access method to the system bus, a token is sequentially transferred between each of the modules. In a duplex ring network that uses a token passing method that controls the right to use the system bus, a token timeout that detects a token timeout when each of the modules is not detected from the system bus used as the active system for a predetermined time or more. Detection circuit A token timeout counter that counts the number of token timeouts detected by the token timeout detection circuit. Each module detects the occurrence of a token timeout with the token timeout detection circuit, and detects the occurrence of the token timeout with the token timeout counter. The master module counts the number of timeouts detected, and if the number of token timeouts detected in its own device is equal to or greater than the predetermined allowable number, the master module switches the active system to one system.
Switch from the bus to the other system bus
After the replacement, the other system bus that has become the active system
Transferred tokens and became active
The number of token timeouts detected by each module is collected from the token timeout counters of all modules via the system bus, and the collected token timeout counts are compared with each other to estimate the failure location where the token is lost. A token failure detection method characterized by doing so. 2. The token failure detection method according to claim 1, wherein the master module compares the number of token timeouts collected from two adjacent modules among the number of token timeouts collected from all modules. A token failure detection method characterized by estimating a failure occurrence location where a token has been lost.