JPS63187939A - Network connection system - Google Patents

Abstract

PURPOSE:To combine a complicated network with giving the number of degrees of freedom of a composite network constitution by adopting a routing method with less regulation on an address giving limit. CONSTITUTION:The subnetworks of bus-type networks 2 and star-type networks 3 are connected to the backbone networks of ring-type networks 1, and the network of a logically tree structure is constituted. Plural ring nodes 5 are connected to the ring-type networks 1 in a ring, and the network is generated. The stations 4 which have unique station addresses in the whole composite network are connected to transmission lines to be shared in the bus-type networks 2. In the star-type networks 3 the stations 4 which have unique addresses in the whole composite network are connected in a star-type with the controller of the star-type networks such as a PBX in a center. The address of the station on the subtransmitting lines 2, 3 of one node is registered on the table so that the transmitting address contained in the packet transmitted from the station 4 corresponds to the address peculiar to the subtransmitting lines 2, 3. And on the basis of this table, the lines are set up.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to means for interconnecting a composite network in which a ring network and a bus or star network are combined.

〔overview〕

The present invention provides a network in which one or more ring networks and a plurality of bus-type or star-type subnetworks are logically combined into a tree structure, by having station address tables in nodes on the ring network. , it is possible to reduce the routing load.

[Conventional technology]

Traditionally, this type of complex network architecture was created with the assumption that it would be a small network, and the routing load was reduced by placing restrictions on the address assignment method.

[Problem that the invention seeks to solve]

These conventional routing methods place restrictions on address allocation and are based on small-scale networks, making it impossible to build large-scale, complex networks. Since an address is assigned, there is a drawback that the station address changes as the device moves.

The present invention aims to eliminate such drawbacks, and aims to provide a network connection system that employs a routing method with fewer restrictions on address assignment.

[Means for solving problems]

In the present invention, one node on one ring-shaped main transmission path is connected to one node on another ring-shaped main transmission path, and a bus-shaped or star-shaped secondary transmission is connected to each node on each main transmission path. In a network connection method, the ring-shaped The main transmission path of is configured to broadcast packets, and each of the above nodes registers the address of the station on its own node's secondary transmission path based on the source address included in the packet sent from this station, and Register the source address included in a packet sent to a station on the secondary transmission path of the own node and the address unique to the node on the secondary transmission path to which this transmission source station is connected, distinguishing it from the above registration. The apparatus is characterized in that the setting means is configured to set the route based on the table.

[Effect]

S in the SA field of a packet frame sent from a station under a node with a station address
SA child table based on A information, D from the packet frame on the ring addressed to the station under the ring node
DA indicating the correspondence between A information and destination ring node number
A routing table consisting of child tables is automatically created during operation. Search the corresponding node number on the ring in the routing table from the DA field in the packet frame sent from the station under the ring node, and if there is a corresponding node number, send one-to-one transmission to the node on the corresponding ring. A packet is sent via communication, and if there is no corresponding node number, it is sent via broadcast. The receiving node bridges the packet sent by one-to-one communication or broadcast based on the SA child table on the routing table if the packet is destined for a subordinate station, and does not bridge the packet if it is not a packet.

〔Example〕

Embodiments of the present invention will be described below based on the drawings.

FIG. 1 is a block configuration diagram showing the configuration of a composite network that is an embodiment of the present invention. Bus type network 2 is added to the bank bone network of ring type network 1.
and the subnetworks of the star network 3 are connected to form a logically complex network with a tree structure. That is, in this network, there are no redundant buses in the communication route (the communication route is uniquely determined. A plurality of ring nodes 5 are connected in a ring shape to the ring network 1 to form a network. A station 4 having a unique station address in the entire complex network is connected to a shared transmission path.The star network 3 is connected to a station 4 having a unique station address in the entire complex network in a star shape centered on the control device of the star network such as PBX. Station 4 with the address is connected.

FIG. 2 shows an example of a packet format flowing within a ring and a packet format flowing through a bus network 2 and a star network 3. In FIG. A subnetwork packet has a field DA indicating the destination address of the station and a field SA indicating the source address.
and the control field C of the subnetwork,
It consists of a data field DATA on the subnetwork and a frame check field FC3 of the subnetwork. In addition, a ring type network packet includes a field R-DA indicating the destination node number on the ring, a field R-3A indicating the source ring number on the ring, a control field R-CONT on the ring, and a field R-DA indicating the destination node number on the ring. It consists of a network frame check sequence field R-Fe2 and a subnetwork packet field in which a subnetwork packet is loaded as is. However, it is also possible that the frame check sequence FC3 of the subnetwork is not rotated around the ring, but is checked at the ring node, and then generated and added again at the final stage of the ring.

FIGS. 3 and 4 are diagrams explaining the routing method used in the present invention. Here, one ring-type network and a path-type network around it will be explained as an example, but the principle is the same even if the network becomes complex. A node 5A with a node number A, a node 5B with a node number B, a node 5C with a node number C, and a node 5D with a node number are connected to the ring network. Further, a bus type network is connected to the node 5A, and on this bus type network there is a station 4a with a station address of a, and a station 4b with a station address of b.

Station 4c with station address C is connected. Node 5B has a station address h
Station 4h with station address d is connected to node 5C, station 4e with station address e and station 4f with station address f are connected to node 5C, and station 4g with station address g is connected to node 5D. ing.

Figure 3 shows the flow of packets and the generation of the routing table for each node when a packet is sent from station 4a to station 4e in a state where the routing table consisting of the SA child table and DA child table of each node has not been created at all. The results are shown. When a packet is sent from station 4a, node 5A looks at address a in the SA frame of the packet and registers that a is under its own node in the SΔ child table. The SA child table has no address e of the DA frame of the packet and D
Since the A child table does not have a correspondence table corresponding to the address e of the DA frame of the packet, the node 5A broadcasts the packet to all nodes. As a result, other nodes 5B, 5C, and 5D receive R-
Since the packet can be recognized from the 3A frame and SA frame as being sent from station address a under node A, each node 5B, 5C, and 5D
A correspondence table between child table node A and station address e is created. Then, packets broadcasted on the ring network are unconditionally bused to the underlying path network and bridged.

FIG. 4 shows the flow of packets when a packet is sent from station 4e to station 4a after the state shown in FIG. 3 and the results of generating the routing table of each node. The packet sent from station 4e is
Looking at the SA frame of the packet, address e is already SA
You can see that the child table is registered. Also, since a of the DA frame does not exist in the SA child table, the DA child table is referred to. Since there is a correspondence between node A and address a in the correspondence table, one-to-one communication is performed on the ring to node A. When the node 5A receives a packet from the ring, it passes the packet to the bus-type network for bridging.

The table shows the process by which a station address table is created through the use of packets as shown in FIGS. 3 and 4.

(Hereinafter, the margin of this page) FIG. 5 is a flowchart showing the procedure from receiving a packet from a bus network to sending the packet to a ring network or discarding the packet. If the received packet is determined to be communication within a bus-type network, the received packet is discarded within the node, and other packets are sent onto the ring for the time being.

Since some of them are not registered in the SA child table, there may actually be communication within the bus network. In such a case, even if a packet is sent out on the ring and flows to another bus network, no other station will receive this packet because the station address is unique throughout the network. If a correspondence table has been created in the DA child table, the destination node number is known, so one-to-one communication is performed. If the correspondence table has not been created, send the broadcast address (Brc) to R-DA.
By setting , the packet will be sent as a broadcast.

FIG. 6 is a flowchart showing the operational procedure from receiving a packet from a loop network to bridging it to a bus network.

〔Effect of the invention〕

As explained above, by adopting a routing method with fewer restrictions on address assignment restrictions, the present invention provides flexibility in the configuration of a complex network, making it possible to construct a complex network, and station addresses can be set to Since it is not restricted by the location on the network, there is an advantage that there is no need to change the station address even when the device is moved.

[Brief explanation of the drawing]

FIG. 1 is a block configuration diagram showing the configuration of an embodiment of the present invention. FIG. 2 is a format diagram of a packet used in an embodiment of the present invention. FIG. 3 and FIG. 4 are explanatory diagrams of the routing method used in the embodiment of the present invention. 5 and 6 are flowcharts showing the operating procedure of the embodiment of the present invention. ■...Ring type network, 2...Bus type network, 7
network, 3... star network, 4... station, 5... ring node.

Claims (1)

[Claims] (1) One node (5) on one ring-shaped main transmission path (1) and one node (5) on another ring-shaped main transmission path (1) are connected, and each main A bus-shaped or star-shaped secondary transmission line (2) is connected to a node (5) on the transmission line.
, 3) is included in the connected network, and has a setting means for setting a route through which packets transmitted between the stations (4) connected to the slave transmission path and having their own unique addresses. In this case, the ring-shaped main transmission path has a configuration in which packets are broadcast, and each of the nodes (5) transmits the address of a station on its own secondary transmission path in a packet sent from this station. It is registered based on the source address, and also corresponds to the source address included in the packet sent to the station on the slave transmission path of its own node and the unique address of the node on the slave transmission route to which this source station is connected. A network connection system comprising: a table means for registering the route separately from the above registration; and the setting means sets the route based on the table.