JPH03253148A - Selective system for communication channel - Google Patents

Abstract

PURPOSE:To make it possible to allocate the same communication channel to the end systems of both local area network LANs in both transmission and reception by providing a gateway device with a corresponding table for a hush computing element, its computed result and its corresponding communication channel. CONSTITUTION:A processing part PU#1 in a gateway device GW#1 and a processing part PU#2 of a gateway device GW#2 judge the gateway device to be communication at every communication frame received from the end system ES of the local area network(LAN), and at th time of determining a communication channel LN to be used, extract a sending network address and a destination network address from the communication frame and execute hush operation based upon a loop back method. Then the communication channel is selected from the corresponding table corresponding to the communication channels to be allocated in accordance with the range of the hush operation result. Consequently, the same communication channel can be allocated by the same corresponding table corresponding to the communication channels in any side of both gateway devices.

Description

Detailed Description of the Invention [Field of Industrial Application] The present invention relates to a communication line selection method in a network, and more specifically, a first local area network that accommodates a plurality of end systems; a second local area network accommodating an end system, a public network including a plurality of communication lines, a first gateway device connecting between the first local area network and the public network, and a first gateway device connecting the first local area network and the public network; A network system comprising: a second gateway device connecting between a local area network No. 2 and the public network; to a second end system comprised of a local area network of
When transmitting data with the frame bottom assigned with the address of the first end system as the source address and the address of the second end system as the destination address, the frame is received from the first end system. The present invention relates to a communication line selection method for selecting which line to use from among a plurality of lines in the public network in the first gateway device.

[Conventional technology]

FIG. 1 is a block diagram showing one embodiment of the present invention, and since it is also useful for explaining a network to which the present invention is implemented, let us first refer to FIG. Describe the network.

In FIG. 1, ES is an end system, AU is an access unit, LAN is a local area network, GW is a gateway device, and WAN is a wide area WM (public network).The gateway device GW is connected to a processing unit PU and multiple lines. The control unit CU is admonishing this. The public network WAN is protected from multiple communication lines LN.

Local area network LAN and wide area network (public network) WAN have different communication methods and communication speeds, so
A gateway device GW is connected between the two for the purpose of achieving consistency between the two by performing conversion or the like.

Now Local Area II! End system ES accommodated in MLAN#1 via access unit AU#11
When #11 transmits data to end system ES#21 accommodated in another local area network LAN#2 via access unit AU#21, the address of end system ES#11 is changed to the source address ( The write field in its frame is denoted as SA), and the address of end system ES#21 is the destination address (its write field is denoted as DA).
It constructs a frame with both addresses assigned, and performs data communication using this frame.

Please refer to the format of this frame as shown in FIG. 1A.

The frame issued by the end system ES#11 in this manner reaches the gateway device GW#1 via the local area network LAN#1 and the access unit AU#12, where it is connected to multiple communication lines LN#1 to LN#1. 1 from inside
The book is selected, and the frame is sent to the destination end system ES#21 via the following route.

However, when the number of communication frames between local area networks LAN increases, it is necessary to use multiple public XWAN communication lines between gateway devices GW to cope with the increase in the number of communication frames. As can be seen in Figure 1. In the gateway device CW, the line control unit CU that accommodates the communication line is different for each communication line, and the protocol processing of communication frames relayed between the gateway devices GW is also performed by the line control unit CU that accommodates the communication line.
Each will be carried out separately. That is, protocol processing is also performed for each communication line.

Therefore, in communication frame relay processing between specific end systems ES accommodated in both local areas 41iiLANs, the communication line selected by the gateway device GW is the same communication line for both upstream and downstream communication frames. If not, the same end system E
Even in communication between S, the protocol processing within the gateway equipment fGW is different for uplink and downlink, so the order of communication frames may be reversed, or the response may be incorrect due to protocol error processing in the relay gateway equipment GW. Normal communication cannot be guaranteed as the return may be sent via a different communication line.

Furthermore, conventionally, each gateway device GW has multiple communication lines L.
When selecting one of N, a method of preparing in advance a list of the correspondence between the combination of the source network address and destination network address in the communication frame and the communication line used, and selecting the communication line for each communication frame. is known, but
In this method, as the number of end systems ESOs communicating with each other increases, the content of the correspondence table becomes enormous, which not only increases the amount of memory required to process it, but also increases the amount of communication frame Each time, the processing time required to search for the communication line to be used increased, resulting in poor processing efficiency.

Furthermore, in the method of selecting the communication line to be used by the gateway device GW based on the source network address of the end system ES accommodated in one of the local area network LANs, the above-mentioned correspondence list is not necessary. 2
It can only be applied to connections between two gateway devices CW, and in the case of an exchange between three or more gateway devices GW or a tandem connection of gateway devices GW, it cannot be managed by the destination network address, and if the communication direction is different. , it was not possible to select and determine the same communication line. Exchange and tandem connections between gateway devices GW are performed in three ways to expand the interconnection range and improve reliability.
It is commonly used when connecting two or more gateway devices WGW.

Furthermore, in the method of selecting a communication line by performing a logical operation on both the source and destination network addresses in a communication frame, the distribution of selected communication lines depends on the method of assigning the network address, and each digit making up the address If the value of is uneven, the distribution of selected communication lines becomes uneven, and it is difficult to distribute the load of communication frames relayed by the gateway device GW to each communication line.

Regarding the conventional communication line selection method, for example,
“Study of Relay Routing Processing Method for LAN/Wide Area Network Connections,” Hidemi Murakami, Sanyoshi Hanaki, and Takanori Miyazawa (NTTii Telecommunications Research Institute), Journal of the 37th National Conference of the Information Processing Society of Japan (late 1986), Vol. Discussed on page 493.

[Problem to be solved by the invention]

The present invention aims to overcome the above-mentioned problems in the prior art in the communication line selection method. In other words, between the same specific end systems ES accommodated in both local area networks LAN, either uplink or downlink communication frames are Even if the same communication line is selected and the number of end systems ES that communicate with each other increases, the contents of the necessary response list will not become huge, and therefore the processing efficiency of communication line selection will decrease. To be able to support exchange between three or more gateway devices GW and tandem connection of gateway devices GW without inviting, and to make the distribution uniform when selecting a communication line among a plurality of communication lines. is the issue to be solved.

Therefore, an object of the present invention is to provide an excellent communication line selection method that solves the above problems.

[Means to solve the problem]

To achieve the above object, the present invention provides a first local area network accommodating a plurality of end systems, a second local area network accommodating a plurality of end systems, a public network including a plurality of communication lines, comprising: a first gateway device that connects the first local area network and the public network; and a second gateway device that connects the second local area network and the public network. In the network system, the gateway device includes a hash calculator and a correspondence table between the calculation results and the corresponding communication lines.

[Effect]

In the present invention, communication frames from the end system ES in the local area w4LAN are transferred to the local area vj
When a gateway device GW accommodating 4LANs selects a communication line for sending to another gateway device GW, it extracts both the sending source and destination network addresses from the communication frame, and uses both network address values to calculate a hash calculation unit. Based on the result, the communication line to be used is selected from the correspondence table between the obtained hash value (calculation result) and the communication line.

In hash calculations, the processing time is longer due to reasons such as dividing a long digit network address into shorter digits suitable for calculation processing and repeating the calculation process, and processing all digits of the network address uniformly. This has the advantage that the hash-calculated values are uniformized quickly.

As a result of this invention, if both gateway devices GW perform a hash calculation using the same network address (the same address where the source and destination are simply swapped), the same hash calculation result will be obtained and the same communication line will be selected. be done. That is, it is possible to select a communication line using only the arithmetic processing of ``no\'' and ``sh'' operations, and there is no need for a list of correspondences between network addresses and communication lines, which was required in the past.

In addition, since arithmetic processing is performed using both the source and destination network addresses in the communication frame, it does not depend on the connection form of the communication line between the gateway devices GW, but also the connection form between the local area network LAN. Even when the gateway device GW is duplicated and the four gateway devices WGW are cross-connected on both sides, it becomes easy to select one specific communication line from a plurality of communication lines in the same direction.

Furthermore, if a folding method (exclusive OR of each component digit of the network address) is used as a hash operation, the value of each digit of the network address will have an equal influence on the selection of the communication line. The distribution of choices becomes uniform.

Regarding hashing operations, for example, Hashing Techniques and Applications, Kiyoshi Nishihara (University of Tsukuba, Department of Electronics and Information Engineering), Information Processing, Vol. 21, Nα9.1980, No. 9
80-991''.

〔Example〕

FIG. 1 is a block diagram illustrating one embodiment of the present invention, as described above.

In the same figure, when an end system ES#11 accommodated in a local area 1i1 LAN#1 communicates with an end system ES#21 in another local area network LAN#2, a communication frame as shown in FIG. 1A is used. The network address NA#11 of the own end system ES#11 is set in the source network address field SA and destination network address field DA of the FR, respectively.
and set the network address NA#21 of the other end system ES#21, and set the local area network LAN#1.
transmission path.

The end system ES also recognizes the arrived communication frame F
It is determined that the value of the destination network address field DA of R matches the own network address NA, and the communication frame FR is received. The access unit AU connects the end system ES to the local area W! 4L
A unit accommodated in the AN fl (AU#1 l, etc.), a plurality of access units AU (AU#11-AU#1 n,
processing for configuring a local area network LAN with AU#21 to AU#2m), accommodating the gateway device GW in the local area network LAN, and processing for LAN support of the communication frame FR for communication between the inside and outside of the local area LAN (AU
#14, AU #22).

In gateway device GW#1, line control unit CU#
11 is public network WAN (leased line, switched communication line wAw4
, etc.) and the gateway device iGW#1, and protocol processing for communication with the gateway device GW#2 connected beyond that. Processing unit PU#1
Is it the access unit AU#12 and the gateway device? &
Processing to connect GW#1 and local area network LA
End system ES in N (ES#11 to ES#1
It is responsible for calculation processing by a hash calculation unit for selecting a line control unit CU accommodating a communication line LN used to relay the communication frame PR received from n) to another gateway device GW#2.

The gateway device GW entrusts communication-related processing such as protocol processing within the local area network LAN of the communication layer I to the access unit AU, and processes information of two or more communication frames FR of the communication layer. Gateway device G
W#2 also has the same configuration as gateway device GW#1 and performs the same functions.

In the system configuration shown in Figure 1, there are multiple communication lines LN (including physical lines in the case of a leased line and logical channels in the case of a packet line) of the public network WAN (LN#1 to LN#k).
), the same communication line LN#1 is used for both local area mLAN end systems (ES1#11 and ES#21).

The reason for this is, as already mentioned, that the end systems ES (ES #11 and ES #11 and ES
#21) If the communication line LN#1 of the gateway device GW used for communication between the same end systems ES differs for each communication frame FR or communication direction, the line control unit CU and the communication line LN This is because the communication delay time differs for each communication frame FR depending on the usage situation, which may reduce communication performance. In addition, line control unit C
In U protocol processing, the received communication frame FR
This is because if a protocol error is to be returned to , it will be done via another line control unit CU or another communication line LN, which will make it impossible to manage the order of communication frames FR and prevent correct communication. .

In other words, in FIG. 1, the processing unit PU#1 of the gateway device GW#1 includes a hash calculation unit (not shown), a hash calculation result obtained by the calculation, and a communication line (LN
#1-LN#k), and a frame F having a format as shown in FIG. 1A.
When R is received from the end system ES#11 via access unit AU#11, local area network LAN#1, and access unit AU#12, the frame FR
The address of the destination network and the address of the sending network are extracted from the DA and SA fields in the DA and SA fields, a hash calculation is performed using the hash calculation unit, and the communication line is determined by referring to the correspondence table from the obtained hash calculation result. Select and determine from among the lines (LN#1 to LN#k).

Similarly, the processing unit PU#2 of the gateway device GW#2 has a hash computing unit and a predetermined correspondence table between the hash computing results obtained by the computing and the communication lines (LN#1 to LN#k). In advance, frame FR is transferred from end system ES#21 to access unit AU#2.
1. When received via local area 1qLAN#2, access unit AU#22, D in the frame PR
A.

The address of the destination network and the address of the sending network are taken out from the SA field, and a hashish calculation unit performs a hashish calculation based on these. From the obtained hashish calculation result, refer to the correspondence table to determine the line to be used, and the communication line (LN#L). ~LN#k).

Therefore, no matter which gateway device GW performs the calculation,
If the frame PR is transferred between the same end systems ES, the same communication line will be selected.

FIG. 2(a) is an explanatory diagram of network addresses and hash calculation results.

In the same figure, the end system ES#11 is assigned a 3-byte address lined with zeros, and this address is the sending source network address of the frame PR when the end system ES#11 becomes the sending source. Destination network address field DA of frame FR when written to field SA and destination
This is the address written to. Other end systems E
A 3-byte address is also assigned to each S, and these are written in the address fields DA and SA of the frame PR in exactly the same way.

Refer to the column in which the address is written at the top of the diagram in FIG. 2 (a). Then, here is the end system ES
The address of #11 (3 bytes, binary), the address of end system ES #21 (3 bytes, binary), and the operation result obtained when a hash operation is performed using both addresses (1 byte, 2 (advanced) is shown.

Similarly, if you refer to the column where the address is written at the bottom, the address of the end system ES#1n (
3 bytes, binary), the address of end system ES#2m (3 bytes, binary), and the operation result obtained when performing a hash operation using both addresses (1 byte, 2
(advanced) is shown.

FIG. 2(b) is an explanatory diagram showing an example of a correspondence table between hash calculation results and communication lines.

When converting the hash operation result from binary to decimal, if the value is 0 to 60, communication line LN#1 is assigned, if it is 61 to 100, communication line LN#2 is assigned, and so on. The table in FIG. 2 (b) shows that if the number is 220 to 255, the communication line LN#k is allocated.

Returning to Figure 2 (a), the rightmost line “Communication line (decimal)”
'' column indicates the communication line determined by referring to the table in FIG. 2(b) based on the hash calculation results.

To add an explanation, in FIG. 1, the processing unit PU#1 of the gateway device GW#1, the gateway device ? &G
The processing unit PU#2 of W#2 determines the communication partner gateway device for each communication frame FR received from the end system ES of the local area network LAN, and also determines the communication line LN to be used there. The source network address and destination network address in the frame FR are taken out and a hash calculation is performed using the folding method. Based on the result of the hash calculation, a communication line is selected based on a correspondence table with communication lines to be allocated according to the value range.

The hash operation using the folding method is an operation in which the authoritative digit string of an address is folded into 8-bit units, and each partial digit string (8 bits) separated by the crease is exclusive-ORed. The amount of memory used does not increase due to shifts, additions, etc., and processing can be easily performed in a uniform time independent of the total number of communication lines.

The result of this hash calculation is the same even if the correspondence between the source network address and the destination network address is reversed, that is, the gateway device on the other side. Therefore, both gateway devices GW#1 and GW#
The same communication line can be allocated on either side of 2 using the same communication line correspondence table based on the same information (hash calculation result).

Note that in this embodiment, the allocation is performed by processing the folding method, but the processing may be performed by the base conversion method, the square-taking method, or the like.

FIG. 3 is a block diagram showing another embodiment of the present invention.

In this embodiment, the gateway device GW is duplicated as a connection form between local area networks LAN, and therefore both sides have 4
This figure shows an example of a configuration in which there are two gateway devices GW and a cross connection is made between them.

In the same figure, local area! 1iLAN#I is a gateway device GW#11 and a gateway device GW#1.
Local area network LAN #2 accommodates two gateway devices Gw, gateway device GW #21 and gateway device GW #22.
to accommodate. When the end system ES in the local area network LAN #1 and the end system ES in the local area network LAN #2 communicate, the combination of the gateway device GW #1 and the gateway device GW #21, the gateway device GW #11 and gateway device G
Combination of W#22, combination of gateway device GW#12 and gateway device GW#21, gateway device [G
Combination of W#12 and gateway device 2cw#22, 4
There are two possible combinations.

In the conventional method of selecting a communication line to be used by the gateway device Gw based only on the source network address of the end system ES accommodated in one local area network LAN, it is not possible to select one communication line. Furthermore, since the destination network address cannot be used for management, it is not possible to use the same communication line if the communication direction is different.

Therefore, in the present invention, as already mentioned many times, selection processing is performed using both the source and destination network addresses in the communication frame.

In this way, even when the gateway devices GW are duplicated as a connection form between local area 1iiLANs, and four gateway devices GW are cross-connected on both sides, without depending on the connection form of the communication line between the gateway devices GW, It becomes possible to select one specific communication line from a plurality of communication lines in the same direction.

Each gateway device G hashash computing unit and corresponding table.
The arrangement at W and the method of selecting a communication line accordingly are also the same as in the embodiment described in FIG.

[Effects of the Invention] According to the present invention, each gateway device GW connected between a plurality of local area networks LANs performs the same hash calculation processing, so that the end systems ES of both local area networks LANs Since the same communication line is automatically assigned for both sending and receiving, there is no need to impose any restrictions on the end system ES when allocating communication lines, and it does not depend on the method of assigning network addresses. It is possible to efficiently and inexpensively select the same specific communication line for each end system ES from among a plurality of communication lines without imposing any restrictions on addressing of the system ES.

[Brief explanation of drawings]

FIG. 3 is an explanatory diagram of the system configuration country showing another embodiment of the present invention. Description of symbols ES...end system, LAN...local area network, AU...access unit, GW...gateway device, PU...processing unit, CU...line control unit, LN...・Communication line, DA...destination network address field, SA...source network address field

Claims (1)

[Claims] 1) a first local area network accommodating a plurality of end systems, a second local area network accommodating a plurality of end systems, a public network including a plurality of communication lines, and a first local area network accommodating a plurality of end systems, a public network including a plurality of communication lines, A network system comprising: a first gateway device that connects between the first local area network and the public network; and a second gateway device that connects between the second local area network and the public network. In this, the gateway device is equipped with a hash calculation unit and a correspondence table between the calculation result and the corresponding communication line, and the gateway device is provided with a hash calculation unit and a correspondence table between the calculation result and the corresponding communication line, and the first end system accommodated in the first local area network , to a certain second end system accommodated in the second local area network, with the address of the first end system as the source address and the address of the second end system as the destination address. When transmitting data in the assigned frame configuration, the first gateway device that receives the frame from the first end system uses the hash calculator to perform a hash operation using the source address and destination address as input data. and determines the line to be used in the public network by referring to the correspondence table based on the calculation result, and conversely transmits the address of the second end system from the second end system to the first end system. When transmitting data with the frame configuration given as the source address and the address of the first end system as the destination address, the second gateway device receiving the frame from the second end system A hash calculation unit is used to perform a hash calculation using the source address and destination address as input data, and since the input data is the same as the calculation performed by the first gateway device, the same calculation result is obtained. The communication line selection method is characterized in that the same used line in the public network is selected and determined by referring to the correspondence table based on the calculation result.