JPS61253946A - Network control system - Google Patents

Abstract

PURPOSE:To reconstitute a network precisely even if a communication right transfer instruction disappears by starting an address searching means when the communication right transfer instruction is initially received after recovery processing. CONSTITUTION:Whether a received token is the initial token obtained after the connection of a power source in a self-node or not is checked, and when the token is the initial one, the token is processed at the time of token transfer operation after the reception of the token. If the token concerned is not the initial token, whether the token transfer operation is the initial token transfer operation after the succeeding recovery processing or not is checked, and in case of the initial transfer operation, the token is generated by any one of nodes to execute the recovery processing for restarting the circulation of the token. When a fixed period has not yet passed, the token is transmitted by using the node address value used at the preceding token transfer operation and stored in a memory circuit 6 as a destination address.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field] The present invention relates to a network control method for a network system in which a plurality of transmission control devices are connected to a communication medium.

"Prior Art" In recent years, local area network systems, so-called "LAN", have become popular, in which a large number of communication control devices (hereinafter referred to as nodes) are connected by sharing a single transmission path, and data communication is performed between these nodes. It's thriving. Among the various communication methods in a LAN, the token passing method is particularly superior in that it can provide communication services evenly to all nodes even when the network is under high traffic conditions, and therefore does not cause a decrease in transmission efficiency. It is.

An example of the system configuration of this LAN is shown in FIG.

In the figure, 1 is a transmission line 100, and 160 is each node A to G forming the LAN. Each node A-G (100-160
) respectively have hosts A to G (
200'' to 260) are connected.

In this way, bus-type LANs generally share a single transmission path, so if each node is allowed to send calls without management, a situation will occur where two or more nodes will be transmitting at the same time, causing problems on the transmission path. A so-called "disaster" phenomenon occurs in which data is mixed.

One method for preventing this is a token passing method. In contrast, in the token passing method, “
Transmission right delegation commands called "tokens" are exchanged between each node via the transmission path, and only the node that receives the token acquires the right to transmit (transmission right). If necessary (if there is a transmission request), performs a series of communication processing such as transmission processing at that point, and when this processing is completed or when there is no need for transmission processing at all, the next transmission right is granted. Node to be delegated (
This method prevents the subtext phenomenon by sending tokens to downstream nodes).

-Token'' is thus passed to the next node every time the transmission process is completed at the node with the transmission request, and is circulated to each node in the network in turn.In this way, each node in the network Nodes have equal opportunities to acquire transmission rights, that is, opportunities for communication services.

There are two methods for token circulation control, depending on the shape of the network: the single token ring method and the "human bus" method. In the token bus method, the transmission path is usually formed with a single linear conductor. Transmission data sent from one node is simultaneously received by all other nodes connected to the transmission path of the network (transmission path 1 in Figure 1).Therefore, the data containing the aforementioned token command is A destination node address is always attached to the destination node, and this is compared with the address value uniquely assigned to each node, and if it matches the node address assigned to the second node, the transmitted data is transferred to the own node. After taking in data as destination data or indiscriminately taking in transmitted data, the addresses are compared, and only when they match is treated as received data officially addressed to the eye node.

In this token bus system, the node to which a certain node should next pass a token (downstream node) is generally defined as the node with the lowest address value that is closest to its own node address. For a node, it can be defined as "the node with the largest address value".

Now, naturally, in a network system, there will be parts that become inoperable due to failures. This cannot be avoided no matter how expensive and highly reliable parts are used. When a failure of one part affects the operation of the entire system, it becomes a big problem.

If one node that makes up the network system goes down, the tokens disappear at this node and cannot be sent to the next node. Therefore, if the situation continues as it is, the token will no longer be circulated, and a serious situation will occur in which communication will be impossible at any node in the network.

In the past, it was sometimes acceptable for systems to go down due to failures, but in recent years, fault tolerance has become more important as a measure of system reliability. In order to strengthen this resistance, it is necessary to take measures such as, for example, disconnecting a failed part from the system and continuing operation with the remaining parts, or reintroducing parts that have recovered from a failure during operation. be done.

Furthermore, in a network, there are circumstances in which it is impossible to deny the existence of inoperable nodes even though they are not out of order. In other words, since the devices and offices connected to a network are spread out over a wide area, it is more convenient for users to start up nodes in the network partially rather than all at once. It would be advantageous to be able to power down the node of the device that has been terminated in a timely manner. In these cases, the network will be operated while including some inoperable nodes.

Considering the above, it should be assumed that changes in the configuration of operable and inoperable nodes in the network can always occur.
On the other hand, each node in the network constantly monitors the status of the downstream node to which it should pass the token next, and if a downstream node becomes inoperable, it immediately changes this node to the network configuration. It is necessary to select another suitable node as the node to which the token should be transferred from now on, and switch to this node.

FIG. 83 shows an example of token circulation in the system shown in FIG. 1 when, for example, node B 110 and node F 150 are inoperable nodes.

As shown in the figure, it is necessary to circulate the token while avoiding inoperable nodes.

Furthermore, a node with an address value intermediate between the node that has been passing tokens and the eye node becomes operational, and is unable to enter the network (receive tokens).
When ready, it is necessary to switch the downstream node to which the token should be passed to this new node.

The above operation can be rephrased as the definition of a node that should use the token described above as ``A node that is closest to the node address and is smaller among the nodes that are in an operational state.''

If there is no corresponding node, the node with the highest value address among the "operable" nodes.

Now, in the past, there have been few methods that have been able to sufficiently follow the changes in the configuration of nodes in a network as described above, and therefore, for example, there have been many cases where the failure of some nodes leads to the failure of the entire network.

Furthermore, even if a device has a means for dealing with the above-mentioned configuration variations, its processing procedure is extremely complicated, and in particular, the overhead of communication control firmware is large, so a simple algorithm is required.

The present invention was developed in view of the above-mentioned drawbacks of the prior art, and is capable of reliably reconfiguring a network even if a node constituting the network becomes abnormal and the communication rights delegation instruction disappears. The purpose of this study is to provide a network control method that can do this using a simple algorithm.

[Example] Hereinafter, an example of the present invention will be described in detail with reference to the drawings.

FIG. 4 is a block diagram of a node according to an embodiment of the present invention.
1 in the figure is the LAN network transmission line shown in Figure 1;
2 is a node which is a transmission control device, and 3 is various computer equipment and office equipment connected to the node 2 (corresponding to the hosts 230 to 260 shown in FIG. 1).

The node 2 includes a transmitting/receiving circuit 4 that performs data communication with the transmission path 1, and a control unit (hereinafter referred to as CPU) that performs overall control of the node, processing of transmitted data, decoding and decomposition of received data, and timing control of communication operations. 5. It is comprised of a memory circuit 6 for storing transmitted and received data, an interface circuit 7 for communicating with the host 3, and an address setting section 8 comprising switches and the like for setting a unique address number for each code. The address value set in the address setting unit 8 is read by the CPU 5 and used as the destination address and source address during transmission and reception.

The data to be transmitted from the host 3 is once stored in the memory circuit 6 in the node 2 and appropriately formatted (packetized) as one piece of communication data. After adding the setting value of the setting circuit 8 as a source address, a token addressed to the eye node is received, and when the transmission right is acquired, it is sent to the transmission path l.

On the other hand, the other node receives all the communication data on the transmission path l, checks the destination address value in the received data, and determines that the setting value of the address setting circuit 8 in the second node is the data addressed to the second node. Once determined, this data is taken into the eye node, subjected to some decomposition and editing processing by the CPU 5, and then delivered to the connected host 3.

In addition, in the token passing method, all the data flowing on the transmission path l is not only the communication data between the hosts 3, but also the communication data between the CPU in each node such as a token and the CPU of the destination node. It also includes so-called communication control data.

Based on the above hardware configuration and the principle of token passing mentioned above, each node in the network receives and passes tokens to downstream nodes one after another, thereby performing communication using a single transmission path. be.

Now, in each node, generally the downstream node to which the token should be passed (that is, the node specified by the destination address added when transmitting the token) is fixed once the token has circulated on the network. There are many cases where

On the other hand, in this embodiment, the following RECO
A method is adopted in which each node is required to perform an NFIGURATION (reconfiguration) operation, and downstream nodes are flexibly changed in response to changes in the network configuration.

The data transmission control of this embodiment will be explained below with reference to the flowchart of FIG.

First, in step SlO, the transmission data on the transmission path l is monitored to check whether a token addressed to the eye node has been received.

If the token is not received, the process proceeds to step Sll, where it is checked whether the destination address is the set address value of the eye node and the transmission data is addressed to the eye node. If the transmission data is addressed to the own node, this data is transmitted at step S12. The received data is received and sent to the host 3 as necessary.

When a token addressed to the eye node is received, step SIO
The process then proceeds to step S13, where it is checked whether there is data to be transmitted from the host 3 and whether there is a transmission request. here,
If there is a transmission request, data transmission processing is executed in step S14, and the process proceeds to step S15. If there is no transmission request in step 313, the process directly proceeds to step 315.

Step 515 checks whether the received token is the first token after power-up in the eye node, and if it is the first token, the process proceeds to step 521, which indicates that the node has started a transmission operation (e.g., powered on). This represents the time of token delegation operation after receiving a token for the first time, and it means that the receiver has just started up and the receiver does not yet know the downstream node to which it should pass the token that it took, and this means that This is to do so.

If it is determined in step S15 that it is not the first token, the process advances to step S16, and due to the occurrence of an unexpected situation such as a failure power down, the token that has been circulating until then disappears, and the first token after the subsequent recovery process is It is checked whether the token transfer operation is being performed or not, and if it is the first token transfer, the process proceeds to step 521. This often occurs in systems where communication rights are transferred by tokens1. This is because if a node is powered down, the token disappears and the token cannot be passed on to a subsequent node.

In such a case, generate a token at one of the nodes,
This is because it is necessary to perform recovery processing to restart the circulation of the token.

If it is determined in step S16 that this is not the first token received after the token disappears, the process proceeds to step S17, and the process proceeds to step S17, where the token is transferred to the node every one foot period, or when the token is transferred every - foot time period. If a certain period has elapsed, the process proceeds to step S21. This is to deal with a node newly joining the network. In other words, if a node with an address value closer to the eye node than the previous downstream node starts up and becomes operational, and the nodes before and after this newly joined node do not know about this, the token will remain unchanged forever. It is not passed on to the participating node, so 6 RECONFIGURAT, which will be described later, is sent at a certain period.
The purpose of the ION operation is to recognize these newly joining nodes and allow them to join the network.

If the fixed period has not elapsed in step 317, the process proceeds to step S1g, and the token is transmitted using the node address value stored in the memory circuit 6 and used in the previous token transfer as the destination address. Then, in the following step S19, it is checked whether the token transfer was performed normally. The means for identifying this differs depending on the transmission control procedure of each network, but for example, the receiving node returns an ACK response as an affirmative response when receiving the token, or the receiving node starts a new communication operation and This is done by determining that the source address of the communication data sent to transmission path 1 is this node address.

If the token has been successfully transferred, the process returns to step SIO to prepare for the next data transmission.

If the token has not been successfully transferred in step 519, it is checked in step 520 whether or not the token has been sent for the second time, and if it is the first time, the process proceeds to step 518;
Execute the token retransmission process. If the token has been sent twice in step 520, the process proceeds to step S21. This is to deal with the case where the other node to which the token has been passed suddenly goes down (failure or power is turned off). This is to find a new node to which the token should be transferred.

In step S21, RECONFIG-11R is a network reconstruction process that newly detects a token delegation light.
Perform the ATION operation, successfully delegate the token, and return to step SlO.

Details of the RECONFIGURATION operation in step S21 will be explained below with reference to the flowchart of FIG.

First, in step Sl, the own node address is set in the search address memory area of the memory circuit 6, regardless of the note address stored as the destination address in the memory circuit 6 as described above, and in the subsequent step S2, the set search The node address is subtracted by 1, and in step S3, the token is transmitted using the search node address obtained by this subtraction as the destination address. Then, in the following step S4, it is checked whether the token transfer was successful or not.

The means for identifying this is the same as in step 319.

If the token is not successfully sent to the downstream node (if the transmission fails), the process proceeds to step S5, assuming that the other node is an inoperable (communication) node, and the search node address value is the minimum specified by the network. Check whether it is the address value or not, and if it is not the minimum address value, step S
Returning to step 2, the search node address value is subtracted by one again, and token transmission is attempted using this new search node address value as the destination address.

In this way, the address value is gradually subtracted and a token is attempted to be sent to the 7th address of the address value, and this operation is repeated until a token is received by any node.

If the token transfer is not successful even if the subtracted address value reaches the minimum address value defined in the network, the process proceeds from step S5 to step S6, and the search node address value is determined in the network. The process returns to step S3 as the maximum node address value.

From then on, the process of receiving and receiving tokens repeats passing and subtracting address values.If the minimum value to be compared is the minimum existing node address, if the value is smaller than that value, proceed to step S6. .

If the token transfer is successful, the process proceeds from step S4 to step S7, where this operation is stopped.
The search node address value at that time is stored in the memory circuit 6, and thereafter the address value at that time is used as the downstream node address.

Since the processing of R1IC0NFIGtlRATION operation takes time, steps S15, l6, 2
In specific cases such as the 4th t6 YES, and in the case of YES in step S17, control is given so that it is performed only every cycle, and in other normal token transfers, the previous RECONFIGURATION operation is requested, memo! j
By using the next node address value stored in the MS as it is as a fixed value, consideration is given to not reducing the data transmission efficiency of the network.

As explained above, according to this embodiment, token delegation can be performed reliably in any case.

In addition, even if token delegation cannot be performed normally, the network can be rebuilt with almost no loss in data transmission efficiency.

In this embodiment, a network using a bus-type communication medium has been described, but another configuration 1, for example, a ring-type communication medium may be used and token delegation is performed using a token bus method.

"Effects" As explained above, according to the present invention, a network control method can be provided that can reliably reconstruct a network system even if a communication transfer command is lost.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the configuration of a bus type network system. FIGS. 2 and 3 are conceptual diagrams showing how tokens circulate in the token bus system. 4th V! J is a block diagram of an embodiment according to the present invention. FIG. 5 is a flowchart showing data transmission control in this embodiment, and FIG. 6 is a flowchart of network reconfiguration operation in this embodiment. here,! ...Network transmission path, 2...Node, 3...Host, 4...Transmission/reception circuit, 5...C
PU, 6... memory circuit, 7... interface circuit, 8... address setting circuit. Figure 1 Figure 2 Figure 5 Sudar←

Claims (1)

[Claims] A network control method for a network system that performs data communication by connecting a plurality of transmission control devices to each other via a communication medium and delegating communication rights using a communication right delegation command having a destination address, wherein the communication rights are delegated to the transmission control devices. The transmission control device includes address search means for searching for a destination address to which a command should be sent, and sending means for sending a communication right delegation command based on the address searched by the address search means, and the transmission control device A network control system characterized in that the address search means is activated when a communication right transfer command is received.