JPH0591144A - Gateway - Google Patents

Abstract

PURPOSE:To reduce the probability of occurrence of a problem relating to overflow of a built-in communication buffer and to improve the data transmission efficiency. CONSTITUTION:An idle buffer size of a communication buffer 14 is obtained by a buffer size management means 11. A header analysis means 12 reads a window size of a header of a protocol of a received data message. A window size correction means 13 corrects a window size when the received data message is sent according to the idle buffer size. Thus, when an end system receiving a data message from a relevant gateway sends data message to the other end system via the gateway, overflow of the communication buffer of the gateway is prevented.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a gateway for passing data messages of a protocol having a header including a window size and connecting different networks or communication lines, and more particularly to overflow of a built-in communication buffer. The present invention relates to a gateway that can reduce problems and improve data transmission efficiency.

[0002]

2. Description of the Related Art A local area network (local ar
ea network, hereinafter also referred to as LAN) is 0.1 to 10
It is a network that transmits digital data at a transmission rate of about 0.1 to 100 MBPS within a range of about km.
Further, the LAN has a certain kind of exchange function as an integrated communication system, and it is possible to organically combine distributedly installed computers, computer peripheral devices, communication terminal devices, and the like.

Therefore, according to such a LAN, the hardware (peripheral devices etc.) of distributed computers, databases, etc. are effective within a limited range such as the premises of offices, factories, research laboratories, and universities. Can be shared.

In addition, digital data public network (digital data
exchange (hereinafter, also referred to as a DDX network) enables data communication not only in Japan but also in foreign countries. In this DDX network, CCITT (international telegrath
and telephone consultative committee) Recommendation X. 25
There is a packet switching network that complies with standards such as.

[0005] The packet exchange is to add a header containing destination information and the like to a block called a packet in units of predetermined bytes, and multiplex packets having different destinations using the same transmission path. Such a packet-switched network has features such that extremely high transmission quality can be obtained.

Further, in recent years, online data exchange is performed by connecting different networks or communication lines, such as connecting different LANs with a DDX network.

FIG. 8 is a block diagram of an example in which different LANs are connected by a DDX network.

In FIG. 8, the networks N10 and N14 are LANs such as CSMA / CD Ethernet defined by IEEE802.3 (hereinafter, IEEE802.3 will be described as an example). or,
The network N12 is a DDX network, X. It is a 25 packet network.

A total of five end systems 2 are connected to the network N10. Also, a total of five end systems 2 are connected to the network N14. These end systems 2 are digital devices such as independent computers and communication terminal devices.

The network N10 and the network N
12 is connected by a gateway 10a, and the network N12 and the network N14 are connected by a gateway 10b. Therefore, the network N10 and the network N14 are connected online.

The transmission rate of the networks N10 and N14 is 10 MBPS, and the transmission rate of the network N12 is 2.4 to 48 KBPS. That is, there is a transmission speed difference of at least 200 times or more between the networks N10 and N14 and the network N12.

Further, there is a difference in protocol and the like between the networks N10 and N14 and the network N12.

Therefore, the gateways 10a and 10b
Is used for connecting different networks having such transmission speed differences and protocol differences.

FIG. 9 shows the IP, T over IEEE802.3.
It is a frame diagram which mounts CP protocol.

The frame indicated by reference numeral A4 in FIG. 9 is used during data communication in the networks N10 and N14.

The frame indicated by the code A4 is a TCP header H for the user data D indicated by the code A1.
t is assembled into a frame such as code A2. For this A2 frame, an IP header Hi is assembled into a frame such as code A3, and a LAN header (IEEE802.3 MAC header, etc.) Ha Are assembled and generated.

In FIG. 9, reference character La is the reception frame length of the frame indicated by reference sign A4, that is, the data length.

FIG. 10 is a data structure diagram of the TCP header.

In FIG. 10, the TC shown in FIG. 9 is used.
The data structure of the P header Ht is mainly shown.

In FIG. 10, the TCP header Ht
Is the source port number Ht 1 and the receiver port number Ht
2, sequence number Ht 3, ACK number Ht 4,
Window size W, checksum Ht 5, argent pointer Ht 6 and other TCP options H
and t 7. Also, this TCP header H
User data D follows t.

The source port number Ht 1 is written with a port number for identifying the application program that transmitted the frame. A port number for identifying an application program that receives the frame is written in the reception source port number Ht2.

The sequence number Ht 3 is written with the serial number of the transmitted frame at each source port.

The ACK number Ht 4 is written with the sequence number to be transmitted next to the source port of the frame by the source port receiving the frame. That is, when the frame in which the ACK number is written as "10" is transmitted from the end system A to the end system B, the sequence number of the frame from the next end system B to the end system A is "10". .. That is, L
Data drop (frame drop) during data communication with AN
Is checked for these sequence numbers Ht 3 and ACK
It is managed by the number Ht 4.

The window size W is written with the size of the communication buffer of the transmission source port, which is used when the transmission source port transmitting the frame becomes the reception side.

The checksum Ht 5 is the TCP header Ht
A checksum including the user data D and the user data D is written.

The description of the argent pointer Ht 6 and other TCP options Ht 7 will be omitted.

FIG. 11 is a diagram showing an X.20 format when transmitting a frame of IP or TCP protocol mounted on IEEE802.3. It is a data block diagram of the head packet of 25 packets.

In FIG. 11, symbols Hi and Ht are
It is the same as the one with the same reference numeral in FIG. 9 described above. As shown in FIG. At the beginning of 25 packets,
X. 25 headers Hx are assembled.

X. The packet length of 25 packets is usually
It is 128, 256, 1024 bytes / packet.
There are also 4096 bytes / packet as a long packet.

As described above, since the packet length Lb is constant, it may not be possible to transmit all the user data in one packet. For example, in FIG. 11, since the frame of FIG. 9 having the frame length La cannot be transmitted in one packet, the user data D of FIG.
1 is only the head portion of the user data D shown in FIG.

FIG. 12 shows the X.264 protocol when transmitting the frames of the IP and TCP protocols mounted on the IEEE802.3. It is a data block diagram of the 2nd packet after 25 packets.

In FIG. 12, the reference numeral Hx is the same as the same reference numeral in FIG.

As shown in FIG. 12, when all the user data cannot be sent in one packet,
The second packet is X. 25 header Hx and user data D2. The configuration of the third, fourth, etc. packets is the same, and these user data are D3, D4, etc., respectively. Also, the frame length La
9 is transmitted, the necessary number of packets are generated and transmitted until all the user data D of the frame can be transmitted.

FIG. 13 is a sequence diagram showing a data communication sequence between the different networks.

The gateway 10a and the gateway 10b are one implementation method of an apparatus called an interworking unit. Therefore, hereinafter, these gateways 10a and 10b are also referred to as IWUa and IWUb.

In FIG. 13, an end system 2 on the network N10 shown in FIG. 8 serves as a file server, and an end system 2 on the network N14 serves as a client. It is shown. After that,
The file server is called end system A and the client is called end system B.

In FIG. 13, first, the request RT of the file A in the end system A which is a file server is sent from the end system B which is a client to the IWUb.
(A) is transmitted via the network N14.

The IWUb relays the received request RT (A) via the network N12 and transmits it to the IWUa.

The IWUa relays the received request RT (A) via the network N10 and transmits it to the end system A which is a file server.

Upon receiving the file request RT (A), the end system A reads the corresponding file A from, for example, the hard disk device and sends it as data DT (A) to the IWUa via the network N10.

When the data DT (A) is received, IWUa
Divides it into a total of four packet data DT (1) to DT (4) according to the packet length of the network N12, and sequentially transmits the packet data to IWUb via the network N12.

When all the packet data DT (1) to DT (4) are received, the IWUb concatenates them into one frame and sends the data DT via the network N14.
(A) Send to end system B.

As described above, the gateway (IW
U) and the network N12 (X.25 packet network) are used to connect the networks N10 and N14, thereby enabling online data exchange between the end systems 2 in the different networks N10 and N14.

[0044]

However, when connecting different networks or communication lines using the conventional gateway (IWU), the capacity of the communication buffer of the end system on each network to be connected, and If the capacity of the communication buffer of the gateway (IWU) is not matched, there is a problem that the data transmission efficiency is lowered, such as overflow of the communication buffer built in the gateway.

This is a problem caused by the matters listed below.

(1) Different network N1 to be connected
0, N14 and the network N1 used for this connection
There is a very large difference in the transmission speed between the two.

(2) The load may be concentrated on the gateway (IWU) used for this connection, such as when data communication requests are issued from a plurality of end systems.

(3) The above-mentioned protocols such as TCP / IP support only the flow control for the communication buffer of the receiving end system.

Therefore, there is a possibility that the communication buffer built in the gateway will overflow, and in this case, the data transmission efficiency will decrease.

FIG. 14 is a sequence diagram showing a data communication sequence between different networks when the communication buffer of the gateway (IWUb) overflows.

In FIG. 14, reference numerals IWUa, IW
Ub, RT (A), DT (A), DT (1) to DT
(4) is the same as the one with the same reference numeral in FIG.

In FIG. 14, packet data DT
When (3) is received, an overflow has occurred in the IWUb communication buffer.

When an overflow occurs in this manner, the IWUb transmits the packet RNR (3) to the IWUa and starts the IWUa transmission restriction.

After that, when the communication buffer of the IWUb becomes empty, the packet RR (3) is transmitted to the IWUa to release the transmission restriction of the IWUa. That is, the period indicated by the symbol Ts is a period during which transmission from IWUa to IWUb is being restricted.

The RNR means "receive not ready".
, And RR is “receive ready.” Also, RN
R (3) also conveys that the packet data from the IWUa having the sequence number "3" has disappeared due to the overflow of the communication buffer. Also, "RR (3)"
Also transmits a resend request for the packet data of sequence number “3” after releasing the transmission restriction.

Therefore, after the transmission restriction is released, the IWUa retransmits the packet data DT (3) and then the packet data DT
(4) is transmitted.

As described above, when an overflow occurs in the IWU communication buffer, a data communication sequence for controlling transmission between IWUs must be performed, and such an overflow causes the data communication sequence to disappear. The packet data that has been lost must be retransmitted, resulting in a decrease in data transmission efficiency. In addition, X. In the 25 DDX network or the like, there is a cost problem in that an extra charge is incurred for the amount of data retransmission. In addition, X. The 25 charges are packet-based.

The present invention has been made in order to solve the above-mentioned conventional problems, and a gateway capable of reducing the problem relating to the overflow of the built-in communication buffer and improving the data transmission efficiency. The purpose is to provide.

[0059]

SUMMARY OF THE INVENTION The present invention provides a communication buffer in a gateway for connecting a different network or communication line for passing a data message of a protocol having a header containing a window size.
Buffer size managing means for managing the buffer size of the communication buffer, header analyzing means for reading the window size from the header of the received data message, and modifying the window size according to the free buffer size of the communication buffer The above-mentioned object is achieved by modifying the window size of the header when the received data message is transmitted.

[0060]

The present invention has been made by paying attention to the fact that when connecting different networks or communication lines, there is a protocol having a header including a window size as a protocol of a data message passed by this.

For such a protocol, for example, FIG.
There is the TCP protocol described above using 0. That is, as shown by the symbol W in FIG. 10, the TCP protocol has a header including a window size. However, the protocol of the present invention is not limited to this. Further, the window size value is not limited to being at the beginning of the data message, but may be at the end of the data message. Further, the window size value itself may be any value according to the size of the reception buffer, and may simply be information of "large", "medium", and "small".

When the gateway of the present invention relays the data message of the protocol having the header including the window size, if the window size is larger than the free buffer size of the communication buffer of the gateway, the communication of the gateway is performed. The window size is modified according to the free buffer size of the buffer for use.

Therefore, according to the present invention, the end system which finally receives the data message with the window size modified receives an empty communication buffer of the gateway in the middle of transmitting the data message to the other end system. It is possible to recognize the window size that also considers the buffer size. Therefore, the overflow of the communication buffer of the gateway on the way can be reduced.

The free buffer size of the present invention means
Any storage capacity that can be used when the gateway temporarily stores a data message that is transmitted in the opposite direction to a data message that includes a window size that is modified (may not be modified depending on the magnitude of the value) is used. For example, it may be a simple buffer size. Further, even when a data message is still in the data buffer during relay between certain end systems, a free buffer size at each time point may be set in a gateway or the like that can relay between other end systems.

FIG. 1 is a block diagram showing the gist of the present invention.

The gateway of the present invention has a communication buffer 14, a buffer size management means 11, a header analysis means 12, and a window size correction means 13.

The communication buffer 14 is used as a buffer when relaying a data message at the gateway.

The buffer size management means 11 manages the buffer size of the communication buffer 14. For example, the buffer size management means 11 obtains a free buffer size of the communication buffer 14.

The header analysis means 12 reads the window size from the header of the received data message. For example, when the TCP protocol is used, the window size value indicated by the symbol W in FIG. 10 is read.

The window size correction means 13 corrects (determines) the window size value read by the header analysis means 12 according to the free buffer size of the communication buffer 14 obtained by the buffer size management means 11. ) Do.

Therefore, the gateway of the present invention uses the window size modified by the window size modification means 13 as the window size of the header when transmitting the received data message.

Therefore, according to the present invention, it is possible to reduce the problem of overflow of the communication buffer incorporated therein and to improve the data transmission efficiency.

[0073]

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

FIG. 2 is a block diagram of the first embodiment of the present invention.

In FIG. 2, the gateway 10
Only the part of relaying from the network N1 to the network N2 is shown, and the present invention is applied to this part. As described above, the present invention may be applied to at least one direction between networks or communication lines connected by a gateway.

The gateway 10 shown in FIG. 2 includes communication control units 70 and 72, a communication buffer 20, a relay control unit 60, a buffer size management unit 30, a header analysis unit 40, and a window size correction unit 50. It has and.

Network N1 of this gateway 10
And the network N2 are mainly connected to each other by the communication control unit 7
0, 72, the relay control unit 60, and the communication buffer 20.

Further, the correspondence between the respective constituent elements of the present invention shown in FIG. 1 and the respective constituent elements of this embodiment is represented by reference numerals 14 and 20,
Reference numeral 11 corresponds to reference numeral 30, reference numeral 12 corresponds to reference numeral 40, reference numeral 13 corresponds to reference numeral 50, respectively.

FIG. 3 is a block diagram of the second and third embodiments of the present invention.

Although not shown in the figure, the gateway of FIG. 3 includes the communication control units 70 and 72 and the like of the first embodiment, and relays from the network N3 to the network N4. That is, in FIG. 3, the illustration of the components denoted by reference numerals 20, 30, 40, 50 and 60 in the first embodiment is omitted.

Further, in the gateway 10 of FIG. 3, the present invention is also applied to a portion for relaying from the network N4 to the network N3.

The relay from the network N4 to the network N3 is mainly performed by the communication control units 76 and 78, the relay control unit 66, and the communication buffer 26.

From network N4 to network N
Correspondence between each part constituting the part for relaying to No. 3 and each component in FIG. 1 is as follows: reference numeral 14 and reference numeral 26, reference numeral 11 and reference numeral 36, reference numeral 12 and reference numeral 46, reference numeral 13 and reference numeral 56. Each corresponds.

In the gateway 10 of FIG. 3,
Storage means and the like are shared by the communication buffers 20 and 26. Further, the buffer size management units 30 and 36,
Header analysis units 40 and 46, window size correction unit 50
56 and the like may be partially shared.

FIG. 4 is a flow chart of the second embodiment.

That is, in the second embodiment of the present invention, the processing shown in the flow chart of FIG. 4 is performed for the relay from the network N4 to the network N3 of FIG.

The process shown in FIG. 4 is activated each time a data message is received from the network N4.

In FIG. 4, first, at step 102, the buffer size management unit 36 checks the available free buffer size of the communication buffer 26. This empty buffer size is the communication buffer 20 (not the communication buffer 26) of the gateway 10 that can be used for the next or subsequent communication that is transmitted between the end systems in the opposite direction to this data message. Actually, since the storage means are partially shared, they are almost the same). Even if all the received data messages are not completely transmitted and remain in the communication buffers 20 and 26, the gateway 10 of the present embodiment can receive another data message between other end systems, for example. There is. The buffer size management units 30 and 36 allocate a buffer area to each of a plurality of received data messages while managing the available free buffer size.

In step 104, the buffer size management unit 36 checks in step 102,
It is determined whether or not the free buffer size for the data message received this time is equal to or smaller than the predetermined reference value α. That is, when the free buffer size is smaller than the predetermined reference value α,
When the free buffer size is equal to or larger than the predetermined reference value α, the process proceeds to the next step 106, and the process branches to the front of step 112.

In step 106, the header analysis unit 4
6 reads the window size value in the header of the data message received this time. In the received data message which is the TCP protocol, the code W in FIG.
The part indicated by is the window size.

In step 108, the header analysis unit 4
6 determines whether the window size value of the header read in step 106 is larger than a predetermined reference value β. When it is determined that the window size value is large, the process proceeds to the next step 110, and when it is determined that the window size value is not large, the process branches to the front of step 112.

In step 110, the above step 102 is performed.
The window size value of the header when the data message received this time is relayed and transmitted to the network N3 is calculated according to the empty buffer size calculated in step S3.

Incidentally, in step 112, the gateway 1
0 is the processing that has been conventionally performed, and the network N4
This is a process of relaying the data message received from the device and transmitting it to the network N3. Also, this step 112
In the relay process, the corrected value is used when the window size is corrected.

FIG. 5 is a flowchart of the relay process performed by the gateway.

The relay process of FIG. 5 is the process of step 112 of FIG. 4 or step 148 of FIG. 6 which will be described later, and relays the data message received from the network N4, which is conventionally performed by the gateway. Is a relay process of transmitting the data to the network N3,
It is mainly executed by the relay control unit 66.

In FIG. 5, first, in step 120, the frame received from the network N4 conforming to IEEE802.3 is X. It is determined whether transmission is possible with only one packet to be transmitted to the network N3, which is the 25 DDX network. That is, 128, 256, 1024, 40
96 bytes / packet packet, network N4
Determines whether all data messages received from can be sent. If it is determined that the length of the packet to be transmitted according to the received frame will be long, the next step 122
If it is determined that it is not long, go to step 12
Branch ahead of 4.

In step 122, the data message received from the network N4 is divided into a plurality of packets having a predetermined packet length.

In steps 124 and 126, X. 25
The header Hx is assembled and the like, and the plurality of packets divided in step 122 are assembled into a packet that can be transmitted from the network N3, and the packet is transmitted to the network N3.

FIG. 6 is a flow chart of the third embodiment of the present invention.

That is, in FIG. 3, the network N
The third embodiment of the present invention is to perform the processing of FIG. 6 in the processing of relaying the data message received from No. 4 and transmitting it to the network N3.

The second embodiment is for data communication between other end systems before all data messages received from the network N4 for data communication between certain end systems are transmitted from the network N3. When receiving another data message from the network N4, the allocation size of the communication buffers 20 and 26 is appropriately determined. In the third embodiment, the functions in this respect are omitted to simplify the entire processing content.

In FIG. 6, first in step 140,
The window size modification unit 56 reads out the buffer size of the communication buffer 26, which is a predetermined fixed value. In the third embodiment, since the buffer size of the communication buffer 26 allocated to the data message received from the network N4 is constant, the buffer size management unit 36 only controls the flow of the communication buffer 26, The size of the buffer allocated to the received data message is not determined.

Step 142 corresponds to FIG. 4 of the second embodiment.
The same process as step 106 is performed.

At step 144, the step 140 is executed.
According to the buffer size read in step S1, it is determined whether the window size value in the header of the data message received from the network N4 should be modified. That is, in step 144, if the window size value read in step 142 is larger than the buffer size read in step 140, it is determined that the window size value is modified, and step 146 is executed. Go to. On the other hand, if the value of the window size is less than or equal to the buffer size, the process branches to the front of step 148.

In step 146, the window size for relaying the received data message and transmitting it to the network N3 is corrected to the buffer size read in step 140.

The relay process in step 148 is a process conventionally performed in the gateway 10, for example, the relay process in FIG.

FIG. 7 is a sequence diagram showing the data communication sequence of the first to third embodiments of the present invention.

The sequence diagram of FIG. 7 corresponds to that of FIG.
14 corresponds to FIG. Reference numerals IWUa, IWUb, RT (A), DT (1) to DT in FIG.
(4) is the same as the one with the same reference numeral in FIGS. 13 and 14 described above.

In FIG. 7, the code W corresponds to the code W in FIG. 10 and is the window size of the TCP header.

Further, the wind mize "m" shown in FIG.
Is the buffer size of the communication buffer of the end system B when the end system B receives the data message DT from the end system A.

The window size "n" is the buffer size (or empty buffer size) of the IWUb communication buffer when the data message is transmitted from the end system A to the end system B (n <m).

In FIG. 7, first, the end system B, which is the client, requests RT to read the file A of the end system A, which is the file server.
Send (A) to IWUb. The required RT at this time
The value of the window size W of the TCP header in (A) is “m
It has become.

Next, upon receiving the request RT (A) from the end system B, the IWUb relays the request RT (A) and the IWUa
Send to. The value of the window size W of the TCP header of the request RT (A) transmitted from the IWUb to the IWUa is rewritten to the buffer size (or empty buffer size) “n” of the communication buffer of the IWUb by the IWUb ( n <m).

The IWUa which has received the request RT (A) having the window size value of "n" transmits it to the end system A. At this time, the buffer size (or empty buffer size) of the communication buffer when the IWUa receives data from the end system A is “n”.
If "1" is smaller than the buffer size "n", the window size W of the request RT (A) transmitted from the IWUa to the end system A may be rewritten to "n1".

The end system A which has received the request RT (A) transmits the file A required by the end system B in accordance with this request.

At this time, the end system A recognizes the minimum buffer size in the route to the end system B from the value of the window size W of the TCP header of the received request RT (A). Therefore, the end system A transfers the data of the file A to the above-mentioned conventional FIG. 13 or FIG.
As described above, the DT (A1) and the DT (A2) are not transmitted at one time but are transmitted separately.

Therefore, according to the first to third embodiments of the present invention, the buffer size of the communication buffer of the gateway in the route from the end system A to the end system B is the same as that for the reception of the end system B. Even when the size is smaller than the buffer size, the transmission restriction as shown in FIG. 14 of the related art does not occur. Therefore, according to the first to third embodiments, it is possible to reduce the problem related to the overflow of the built-in communication buffer and improve the data transmission efficiency.

Although the window size is corrected in these embodiments, since the checksum in the TCP header is the sum operation of the complement of "1", it is easy to recalculate the checksum when the window size is changed. .. In recent years,
It is expected that the interworking unit will be used as an agent for network management (LAN standard MIB-II). In this case, the TCP header analysis means of these embodiments is effective in that it can be used for TCP header analysis when used as a network management agent.

[0119]

As described above, according to the present invention, it is possible to provide a gateway capable of improving the data transmission efficiency by reducing the problem relating to the overflow of the built-in communication buffer. It is possible to obtain the excellent effect.

[Brief description of drawings]

FIG. 1 is a block diagram showing the gist of the present invention.

FIG. 2 is a block diagram of a first embodiment of the present invention.

FIG. 3 is a block diagram of a second embodiment and a third embodiment of the present invention.

FIG. 4 is a flowchart of the second embodiment.

FIG. 5 is a flowchart of a relay process used in the second embodiment and the third embodiment.

FIG. 6 is a flowchart of the third embodiment.

FIG. 7 is a sequence diagram showing a data communication sequence of the first to third embodiments.

FIG. 8 is a block diagram of an example in which different LANs are connected by a DDX network.

FIG. 9 is an illustration of an I mounted on an IEEE802.3.
FIG. 3 is a frame diagram of P and TCP protocols.

FIG. 10 is a data configuration diagram of a TCP header.

FIG. 11 is a diagram showing a case of transmitting a frame of IP or TCP protocol mounted on IEEE802.3,
The first X. It is a data block diagram of 25 packets.

FIG. 12 is a diagram showing a case of transmitting a frame of IP or TCP protocol mounted on IEEE802.3,
The second and subsequent X. It is a data block diagram of 25 packets.

FIG. 13 is a sequence diagram showing a conventional data communication sequence between different networks.

FIG. 14 is a sequence diagram showing a data communication sequence when the communication buffer of the gateway overflows.

[Explanation of symbols]

2 ... End system, 10, 10a, 10b ... Gateway (or interworking unit, IWU), 11 ... Buffer size management means, 12 ... Header analysis means, 13 ... Window size correction means, 14, 20, 26 ... Communication buffer , 30, 36 ... Buffer size management unit, 40, 46 ... Header analysis unit, 50, 56 ... Window size correction unit, 60, 66 ... Relay control unit, 70, 72, 76, 78 ... Communication control unit, N1 to N4 , N10, N12, N14 ... Network, Hx ... X. 25 header, Hi ... IP header, Ht ... TCP header, D, D1, D2 ... User data, RT (A) ... File request, DT (A), DT (A1), DT (A2), DT (1)
-DT (4) ... Transmission data, IWUa, IWUb ... Interworking unit (gateway), W ... Window size, m ... End system B buffer size, n ... IWUb buffer size, n1 ... IWUa buffer size, La ... Frame length of LAN protocol according to IEEE802.3, Lb ... X. Packet length of 25 DDX network packets.

Claims (1)

[Claims] 1. A gateway for connecting a different network or a communication line for passing a data message of a protocol having a header including a window size, and a buffer size managing means for managing a communication buffer and a buffer size of the communication buffer. And a header analysis unit that reads the window size from the header of the received data message, and a window size correction unit that corrects the window size according to the free buffer size of the communication buffer. A gateway characterized by modifying the window size of the header when sending.