JPH01122230A - Data transmission system for network system - Google Patents

Abstract

PURPOSE:To effectively distribute a load in multistage bridges by relaying frames, which meet relay conditions stored in a bridge itself, out of reception frames from one LAN to another. CONSTITUTION:Information indicating relay conditions of frames 11-14 to be transmitted from one LAN to the other are preliminarily stored in bridges 3 and 4, and each bridge relays only frames, which meet the relay condition stored in the bridge itself, out of reception frames from one LAN to the other. Thus, it is unnecessary for each station which transmits frames to pay attention to bridges which should relay frames, and the CPU time for bridge selection and the change of an already set frame format due to bridge designation are unnecessary. Further, it is unnecessary for the station to pay attention to entrance and secession of bridges.

Description

Detailed Description of the Invention [Field of Industrial Application] The present invention relates to a data transmission method in a network system, and more specifically, to load balancing of bridges in a network system in which a plurality of LANs are connected by multiple bridges. Concerning data transmission methods suitable for.

[Conventional technology]

Conventionally, for example, there is a network system in which multiple LANs are connected by multiple bridges.

Globecom 85 I Triple E Global
Telecommunications Conference 1019
Pages to 1023 (GLOBIECOM '85IRE
! E Global Telecommunica
tionConferenespplo19-10
23).

In the conventional system described above, the sending station determines the LAN number and bridge number to pass through to the destination station, and embeds this information in the frame.The bridge refers to this information in the frame and determines whether it is designated as If so, relay that frame.

[Problem that the invention seeks to solve]

However, in conventional systems, each transmitting station
The following problems arise because a specific bridge is specified from among a plurality of bridges that multiplexly connect networks existing on the path to the destination station.

(i) Each station is aware of the existence of multiple bridges,
Selecting one bridge from among them consumes CPU time.

(it) Since each station selects a bridge to pass through without having a means of knowing the number of traffic passing through each bridge in the multiple bridges, it is difficult to effectively distribute the traffic load among the multiple bridges.

SUMMARY OF THE INVENTION An object of the present invention is to provide a data transmission system that can effectively reduce the load in the multi-level bridge described above.

[Means for solving problems]

In order to achieve the above object, the data transmission method of the present invention is a network system in which two LANs are interconnected by a plurality of bridges. Information indicating relay conditions is stored in advance, and only those received frames that satisfy the relay conditions stored in the own bridge are relayed from one LAN to the other LAN. do. The relay condition is given to each bridge as, for example, a specific n-bit value in the source address or destination address included in the received frame, or a value corresponding to the length of the received frame. In this case, each bridge counts its own traffic volume, and if the load becomes uneven between bridges, the above relay conditions can be corrected using the protocol between the bridges, and the load can be automatically distributed. can be achieved.

[Effect]

According to the present invention, when interconnecting LANs using multiple bridges (multiple bridges), each bridge selectively relays only frames that satisfy the relaying conditions assigned to it. Each station transmitting a frame does not need to be aware of the bridge to which the frame should be relayed, and there is no need to use CPU time to select a bridge or change the existing frame format to specify a bridge. Furthermore, there is no need for stations to be aware of joining and leaving bridges.

〔Example〕

Embodiments of the present invention will be described below with reference to the drawings.

1 to 4 are diagrams for explaining the first embodiment of the present invention, and FIG. 1 shows the load between bridges when LANI and LAN2 are interconnected by bridges 3 and 4. This is a diagram illustrating dispersion.

Bridges 3 and 4 only perform this processing when the lower n bits (n=2 in this example) of the source address in the received frame match the corresponding number in the respective tables 5 and 6 stored in memory in advance. The frame is captured and "1" is added to the corresponding passing traffic column in each table.

For example, if terminal 7 of LANI and terminal 10 of LAN2 have frame 13 and frame 12. A case will be described in which the terminal 8 of the LAN and the terminal 9 of the LAN 2 exchange frames 14 and 11. In this case, each of the above terminals is
0”, “5”.

Assume that addresses r31J and r6J are given.

Among the four frames, frame 11. transmitted by terminal 9 is frame 11.
Source address lower 2 of frame 13 sent by terminal 7
These frames are captured by bridge 3 since the bit is rOJ or "2". On the other hand, the lower two bits of the source addresses of frames 12 and 14 are rl and r3, respectively, so these frames are taken into the bridge 4.

In this way, each bridge uses the source address in a received frame to determine whether or not to capture the frame, so that load balancing on the bridge is automatically performed without the need for each station that sent the frame to be aware of it. There is. Generally, in each LAN, station addresses are often assigned consecutively, so if a system is adopted as in this embodiment, in which the bridge that should process the frame is determined based on the lower two bits of the source address, each Frames relayed to the bridge can be allocated evenly. In addition, as shown in Tables 5 and 6 in Figure 1, if traffic is unequally allocated to each bridge, each bridge By changing the source address bit number that is responsible for
The load between bridges can be made nearly uniform.

According to the protocol of FIG.

When the total transit traffic reaches the value (200 in this example) or the failure is recovered (step 101), other bridges (in this example, bridge 3 is connected to bridge 4)
Inquire about the person in charge number traffic and total traffic (102). As a result, if the difference from the total traffic of other normal bridges exceeds the allowable range (80 in this example) or there is a bridge that does not respond (103), the assigned number will be reassigned (104). Notify the normal bridge and reset the passing traffic field (105
).

Otherwise, each normal bridge only resets the transit traffic field. FIG. 3 shows a changed table 5' corresponding to tables 5 and 6.

The contents of 6' are shown below.

Communication between bridges is performed using addresses (6 bytes in this example) shown in FIG. This is made up of the multiplex bridge number (46 bits in this example) and a 2-bit number that differs for each bridge.For example, if the multiplex bridge number is 100, the address of bridge 3 in Figure 1 is r400J. The addresses of r402J and bridge 4 are "401" and "403". After the assigned numbers are changed as shown in FIG. 3 by the algorithm shown in FIG. 2, the addresses for bridge 3 are r402J, and for bridge 4 are r400J, r401J, and r403J.

Using the above address, for example, process 1 in FIG.
02 has both source address and destination address as r40
This can be done using a frame set to 1J.

Using the algorithm of FIG. 2, the bridge failure occurs in step 103. Failure recovery is discovered in step 101. After this, the removal of the failed bridge and the entry of the failed bridge are performed in steps 104 and 104, respectively.
105 and 102° 103, 104, 105. Therefore, with this algorithm, it is possible not only to distribute the load by multiple bridges but also to take measures against bridge failures.

In this embodiment, relaying duties are assigned to each normal bridge according to the specific n bits of the source address, but the destination address may be used instead of the source address. Also,
Source addresses may be used for some types of traffic and destination addresses for others. For example, if the extended LAN topology is hierarchical, the upper LAN
Relay assignments can be assigned using the destination address for traffic from the lower LAN and the source address for traffic from the lower LAN.

Next, a second embodiment of the present invention will be described with reference to FIGS. 5 to 8.

Similar to FIG. 1, FIG. 5 shows a network in which LANI and LAN2 are interconnected by bridges 3 and 4. Here, the maximum frame length of LANI and LAN2 is 2047 bytes. Bridge 3. The bridge 4 captures the frame only when the length of the received frame falls within the corresponding frame length range registered in advance in tables 5 and 6, and adds "1" to the corresponding transit traffic column. Operate. Here, Table 5. The frame length delimiter in Table 6 is 256 bytes.

An example of load distribution will be explained using communication between terminal 8 and terminal 10. In order for terminal 8 to communicate with terminal 10, a link must be established. For this purpose, the terminal 8 sends a link setup request frame 15 to the terminal 10. Terminal 10, which received this, sends link setting confirmation frame 1 to terminal 8.
Send 6.

Since these two frames have short frame lengths, the bridge 3 will be in charge of them. After this, terminal 8 and terminal 1
Since the frame communicated by 0 is relatively long, bridge 4 is often in charge of the frame.

By dividing the relay bridges according to the frame length in this way, load distribution is performed without each station being aware of it. As shown in Tables 5 and 6 in Figure 5, if traffic is unequally allocated to each bridge, the range of frame lengths each bridge is responsible for is determined according to the inter-bridge protocol shown in Figure 6. By changing the , it is possible to make the loads between the bridges more even. In this case, each bridge needs to be able to distinguish frames used for inter-bridge protocols from other frames, and to be able to receive these frames regardless of the frame length in charge of its own bridge. For this purpose, it is convenient to divide the bridge address system into multiple bridge numbers and individual bridge numbers, as shown in FIG. 7, for example. When using this address system.

Each bridge receives only the frame having the multiplex bridge number to which it belongs, regardless of the frame length.

To explain the algorithm in Figure 1, in each bridge,
When the total passing traffic reaches a value (200 in this example) or the failure has been recovered (step 201), the range of frame lengths that other bridges (in this example, bridge 3 is responsible for relative to bridge 4) is Each transaction is queried (202).

As a result, if the difference between the total traffic of other normal bridges exceeds the measurement range or there is a bridge that does not respond (203), the frame length range in charge will be reassigned.
04), Notify the normal bridges of this result and reset the passing traffic column (205). Otherwise, make each normal bridge only reset the passing traffic column (206). The changed contents of Tables 5 and 6 are shown as Tables 5' and 6'.

Using the algorithm shown in FIG. 6, the occurrence of a bridge failure is discovered in step 203, and failure recovery is discovered in step 201. After that, the removal of the failed bridge and the entry of the bridge that has recovered from the failure are performed in step 2, respectively.
04,205 and 202°203.204,205. Therefore, it can be seen that this algorithm enables not only load distribution by multiple bridges but also countermeasures against bridge failures.

〔Effect of the invention〕

According to the present invention, even when LANs are interconnected by multiple bridges, each bridge captures only frames that meet the relay conditions assigned to itself, so each station does not need to be aware of multiple bridges. . Therefore, there is no need for CPU time to select a relay bridge or to change the existing frame format for specifying a relay bridge.

Further, when the relay condition is set as a frame point, if the buffer length is divided into the maximum value of the frame in charge of each bridge, the memory for the buffer can be used effectively.

[Brief explanation of the drawing]

FIG. 1 is a diagram for explaining load distribution between bridges according to the first embodiment of the present invention, FIG. 2 is a diagram for explaining an algorithm for dynamically allocating n-bit values to bridges, and FIG. The figure shows the state of the table modified by applying the algorithm in Figure 2, and Figure 4 shows the address system of bridges when exchanging data between bridges in the algorithm in Figure 2. 5 is a diagram for explaining load distribution between bridges according to the second embodiment of the present invention, FIG. 6 is a diagram for explaining an algorithm for dynamically assigning frame lengths to bridges, and FIG. FIG. 7 is a diagram showing the address system used for communication between bridges in order to perform the algorithm in FIG. 6, and FIG. 8 is a diagram showing the state of the table modified by applying the algorithm in FIG. 6. . 1.2...LAN, 3,4...bridge, 5,6...
...Bridge table, 7~1o...Terminal, 11~
18...Frame, 5', 6'...Table after change.

Claims (1)

[Claims] 1. In a network system in which two LANs are interconnected by a plurality of bridges, each bridge stores in advance information indicating relay conditions for frames to be transmitted from one LAN to the other LAN. A data transmission method in a network system, characterized in that, among received frames, only those that satisfy relay conditions stored in the own bridge are relayed from one LAN to the other LAN. . 2. The data transmission system in the network system according to item 1, wherein the relay condition corresponds to a specific bit in an address included in a received frame. 3. The data transmission system in the network system according to item 1, wherein the relay condition corresponds to the length of a received frame.