JP2001136209A - Communication apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To suppress the decrease of effective throughput and also to reduce the discard rate of packet in the case a bit rate takes place in a communication apparatus used in a communication system utilizing a radio communication line. SOLUTION: This communication apparatus which transmits and receives data including user information to/from the opposite station through the radio communication line, also receives an acknowledgment packet from the opposite station to perform acknowledgment and limits a continuously transmissible data amount by a window size without the acknowledgment, is provided with a discard cause specifying means 32 which inputs a signal from a communication line and specifies the discard cause of data transmitted by the self-station on the basis of a signal transmitted by another station, and a window size controlling means 20 which changes controls over the window size between in the case of the congestion occurrence of the discard cause specified by the discard cause specifying means and in the case of the bit error occurrence when the discard of the data transmitted by the self-station is detected and makes the window size in the case of the bit error occurrence at least larger than that in the case of the congestion occurrence.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a communication apparatus for transmitting / receiving data including user information to / from a partner station via a wireless communication line.

[0002]

2. Description of the Related Art Various communication systems (wireless LANs, mobile communication networks, satellite networks, etc.) using wireless communication lines have been widely used in recent years because of their high mobility.

For example, a wireless LAN (Local Area Network)
An example of k) is “Nikkei Communications, pp.81-
87, April 19, 1999, and an example of a mobile communication network in which a portable personal computer is connected to a mobile telephone is described in "Nikkei Communications, pp. 99-1".
17, March 1, 1999, and an example of a satellite network is described in "TCP / IP Enhancemen by N. Ghani et al."
ts for Satellite Networks ”, IEEE Communications Ma
gazine, pp. 64-72, July 1999].

[0004] In such a wireless link type network, bit errors occur due to deteriorating weather conditions and fading, so that it is inevitable that some errors occur in data being transferred. If a bit error occurs in the data being transferred, generally, a checksum, CRC (cyclic
The occurrence of an error is detected by a relay device or a receiving device by an error detection process using redundancy check) or the like. Then, the erroneous data is discarded by the relay device or the receiving device.

[0005] When a relay device such as a router exists in the middle of a communication path, data is discarded even when congestion occurs in the relay device. When performing data communication between computers such as file transfer, e-mail, Web, etc., highly reliable data transfer is required.
When data is discarded during transfer, it is necessary to perform control such as retransmission.

For example, in Reference 1 “WRStevens,“ TCP / IP
Illustrated, Volume 1: The Protocol ”, Addison-We
sley Publishing Company, 1994 ", the transport protocol TCP (Transmission Control Protocol).
l) In the case of the receiving station, the data containing user information
(A data packet) is notified to the transmitting side when it is received without error (delivery confirmation function), and when the data packet is discarded due to a bit error or congestion on the transmission line, the transmitting station is discarded Executes the function of retransmitting the packet (error recovery function).

In order to prevent packet overflow on the receiving side and congestion in the network, the transmitting side executes a function (flow control function) for limiting the number of data packets that can be transmitted without confirming delivery. In TCP, a sliding window flow control method (hereinafter, referred to as a window flow control method) is employed in order to simultaneously realize a flow control function and a delivery confirmation function. In the window flow control method, the number of data packets that can be transmitted by the transmitting station without acknowledgment is called a window size, and the transmitting station controls the window size to transmit traffic to the receiving station (that is, transmit traffic to the network). Control.

[0008] In the TCP protocol, when the discard of a data packet is detected by utilizing the property of the window flow control system, it is assumed that congestion has occurred between the transmitting station and the receiving station, and the window size is determined. Is changed to a small value (pages 306 to 316 of the aforementioned document 1).

[0009]

As described above, in a communication system using a wireless communication line, packets are discarded not only when congestion occurs but also when a bit error occurs in the wireless communication line. .

However, in a dynamic window protocol such as TCP, regardless of the cause of the packet discard,
Detecting packet discard reduces the window size. When the window size is reduced, the amount of data transmitted by the transmitting station per unit time is reduced, so that the effective communication throughput is reduced. Also, reducing the amount of data transmitted by the transmitting station when a bit error occurs in a wireless communication line does not necessarily reduce the number of discarded packets. However, it is necessary to reduce the window size for occurrence of congestion.

According to the present invention, a communication apparatus used in a communication system using a wireless communication line suppresses a decrease in effective throughput when a bit error occurs and reduces a packet discard rate when a bit error occurs. The purpose is to:

[0012]

According to the first aspect of the present invention, there is provided a communication apparatus comprising:
Sends / receives data including user information to / from the partner station via the wireless communication line, receives a delivery confirmation packet from the partner station for the data transmitted by the own station, confirms the delivery, and performs no delivery confirmation. In a communication device that limits the amount of data that can be transmitted continuously to the window size,
A discarding cause specifying means for specifying a cause of discarding of data transmitted by the own station based on a signal transmitted by another station; and a discarding cause specifying means for detecting discarding of data transmitted by the own station when the discarding of data transmitted by the own station is detected. Window size control means for distinguishing the control over the window size between a case where a congestion occurs and a case where a bit error occurs, and increasing the window size when a bit error occurs compared to the window size when a congestion occurs. It is characterized by having been provided.

In the first aspect, the discard cause specifying means specifies the cause of discard of the data transmitted by the own station based on the signal transmitted by the other station. The window size control means controls the window size depending on whether the cause of the discard specified by the discard cause specifying means is a congestion or a bit error when detecting the discard of the data transmitted by the own station. For discrimination, the window size when a bit error occurs is made larger than that when congestion occurs.

In the case where congestion occurs, the window size may be reduced as in the prior art. In the case where a bit error occurs, window control for the occurrence of congestion may be completely stopped. The window size may be changed so as to be larger.

According to the first aspect, since the window size when data is discarded due to the occurrence of a bit error is larger than that when congestion occurs, a decrease in effective throughput can be suppressed. If congestion occurs, the window size can be changed to a size large enough to avoid it. Further, since it is not necessary to generate new traffic in accordance with this control, it is possible to prevent an increase in traffic.

In the case of the TCP protocol, for example, an ICMP (Internet Control Message) transmitted by a relay device is used.
ge Protocol) It is possible to identify whether a congestion has occurred or a bit error has occurred in the current connection based on the content of the error packet. According to a second aspect of the present invention, in the communication device of the first aspect, when the discard of the data transmitted by the own station due to occurrence of a bit error is detected, the maximum user data field length of the data packet transmitted by the own station is determined. User data field length control means is provided to limit the maximum user data field length when the frequency of occurrence of bit errors is reduced.

The packet discarding rate caused by the occurrence of a bit error is almost proportional to the packet length. That is,
The longer the packet length, the higher the packet discard rate. According to the second aspect, when a bit error occurs, the maximum user data field length of the data packet transmitted by the own station is limited to a short length, so that an increase in the packet discard rate can be suppressed.

The relationship between the packet length and the packet loss rate is described in reference 3 (M. Schwartz, “Telecommunications”).
tion Network Protocols, Modeling and Analysis ”, A
ddison-Wesley Publishing Company, pp. 131-134, 199
4) is shown. According to a third aspect of the present invention, in the communication apparatus of the second aspect, the maximum user data field length of the data packet transmitted by the own station is limited to a value of an integral multiple of a predetermined fixed length.

According to the third aspect, the length of the user data to be transferred is limited to an integral multiple of the fixed length.
Data management in a buffer circuit or the like used for retransmission is facilitated, and the circuit configuration is simplified. According to a fourth aspect of the present invention, in the communication apparatus according to the third aspect, each time the user data field length control means detects discard of the data transmitted by the own station due to occurrence of a bit error, the data packet transmitted by the own station. The maximum user data field length is changed to a value reduced by the fixed length, and each time the transmission of the data packet transmitted by the own station is confirmed to be successful, the maximum user data field length is increased by the fixed length. It is characterized by changing.

In the present invention, when a packet is discarded due to the occurrence of a bit error, the maximum user data field length is reduced by the fixed length, and when the packet is successfully transmitted, the maximum user data field length is increased by the fixed length. Therefore, control for suppressing an increase in the packet loss rate is simple. According to a fifth aspect of the present invention, in the communication device according to the third aspect, when the user data field length control means detects discard of the data transmitted by the own station due to occurrence of a bit error, the data transmitted by the own station. The maximum user data field length of the packet is changed to a value equal to the fixed length, and each time the transmission of the data packet transmitted by the own station is confirmed to be successful, the maximum user data field length is increased by the fixed length. It is characterized by changing.

According to a fifth aspect of the present invention, when a packet is discarded due to a bit error, the maximum user data field length becomes a value equal to the fixed length, and when the packet is successfully transmitted, the maximum length is increased by the fixed length. Become. Therefore, control for suppressing an increase in the packet loss rate is simple. According to a sixth aspect of the present invention, in the communication apparatus according to the third aspect, when the user data field length control means detects discard due to occurrence of a bit error in the data transmitted by the own station, the data transmitted by the own station. The maximum user data field length of the packet is changed to a maximum value that does not exceed half of the current maximum user data field length among the integral multiples of the fixed length, and the data packet transmitted by the own station is changed. Each time transmission success is confirmed, the maximum user data field length is changed to a value increased by the fixed length.

According to a sixth aspect of the present invention, when a packet is discarded due to a bit error, the maximum user data field length is an integer multiple of the fixed length and the current maximum user data field length is set to an integer multiple of the fixed length. Is the largest value that does not exceed half. When the packet transmission is successful, the maximum user data field length is increased by the fixed length. Therefore, control for suppressing an increase in the packet loss rate is simple.

According to a seventh aspect of the present invention, in the communication apparatus according to the third aspect, each time the user data field length control means confirms the success of transmitting a predetermined number of data packets transmitted by the own station, the maximum user data field length is controlled. The field length is changed to a value increased by the fixed length. In the present invention, when the transmission of a predetermined number of packets is successful, the maximum length of the user data field is increased by the fixed length. Therefore, control for suppressing an increase in the packet loss rate is simple.

According to an eighth aspect of the present invention, in the communication apparatus of the third aspect, the user data field length control means controls the number of the fixed length number corresponding to the current maximum user data field length of the data packet transmitted by the own station. Each time the transmission of a data packet is confirmed to be successful, the maximum user data field length is changed to a value increased by the fixed length.

Each time a packet corresponding to the current maximum user data field length is successfully transmitted, the maximum user data field length increases by the fixed length. Therefore, control for suppressing an increase in the packet loss rate is simple.

[0026]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a communication device according to the present invention will be described below with reference to FIGS. This form corresponds to all claims. FIG. 1 is a block diagram showing the configuration of the mobile terminal of this embodiment. FIG. 2 is a block diagram illustrating a configuration example of a communication system. FIG. 3 is a schematic diagram showing the state transition of the connection state. FIG. 4 is a flowchart showing the operation (1) of the transmission side TCP processing unit of this embodiment. FIG. 5 is a flowchart showing the operation (1) of the ACK packet processing unit of this embodiment.

FIG. 6 is a schematic diagram showing a configuration example (1) of the transmission side TCP connection management table. FIG. 7 is a schematic diagram showing a configuration example (2) of the transmission-side TCP connection management table. In each drawing, the contents of each claim are concretely shown as follows. FIG. 4 includes the contents of claims 1 and 2, FIG. 8 includes the contents of claims 3 and 4, FIG. 9 includes the contents of claim 5, and FIG.
0 includes the contents of claim 6, FIG. 11 includes the contents of claim 7, and FIG. 12 includes the contents of claim 8.

FIG. 8 is a flowchart showing the operation (2) of the transmission side TCP processing unit of this embodiment. FIG. 9 is a flowchart showing the operation (3) of the transmission side TCP processing unit of this embodiment. FIG. 10 is a flowchart showing the operation (4) of the transmitting side TCP processing unit of this embodiment. FIG. 11 is a flowchart showing the operation (2) of the ACK packet processing unit of this embodiment. FIG. 12 is a flowchart showing the operation (3) of the ACK packet processing unit of this embodiment.

In this embodiment, the discard cause specifying means in claim 1 corresponds to the connection state detecting section 32 or the ICMP error packet processing section 80, and the window size control means in claim 1 corresponds to the transmission side TCP processing section 20. I do. The user data field length control means in claim 2 corresponds to step S11 in FIG.

In this embodiment, it is assumed that the communication apparatus of the present invention is implemented as the mobile terminal 1 used in the communication system as shown in FIG. The communication device of the present invention does not need to be a mobile terminal, and the present invention can be applied to a wired terminal as long as a communication system has a wireless link on a communication path as described later. Referring to FIG. 2, a mobile terminal 1 is connected to a wireless base station 3 via a wireless line 4.
Communication with is possible. The wireless base station 3 is connected to the fixed terminal 2 via a wired communication network 5. Therefore, when the mobile terminal 1 communicates with the fixed terminal 2, the wireless line 4 and the wired communication network 5 are used.

In such a communication system, the radio line 4
There is a possibility that a bit error may occur at a relatively high frequency due to deterioration of weather conditions or fading in the part. The communication device of the present invention is useful for suppressing a decrease in throughput when a bit error occurs. Therefore, the communication apparatus of the present invention is applied not only to the mobile terminal 1 directly connected to the wireless line 4 as shown in FIG. 2, but also to a communication terminal used in a communication system including at least one wireless line 4 in a communication path. May be.

In this embodiment, TCP is used as a communication protocol between the mobile terminal 1 and the fixed terminal 2. The reason why the TCP is adopted here is that the present invention facilitates the application of the present invention because the TCP treats the data amount or the like of which delivery has been confirmed as a value in byte units. Here, TC
As the version of P, it is assumed that 4.4 BSD (Berkeley Software Distribution) described in Document 1 is used.

In connection type protocols such as TCP, it is necessary to control connection establishment and release during communication. However, since the main control contents of the communication device of the present invention relate to the data transfer phase after the connection is established, the description of the connection establishment and release processing is omitted here.

As shown in FIG. 1, the mobile terminal 1 includes a transmission-side upper layer processing unit 10, a transmission-side TCP processing unit 20, a reception-side TCP processing unit 30, a transmission-side TCP connection management table 40, and a transmission-side IP process. Unit 51, receiving-side IP processing unit 5
2, a transmission line processing unit 60, a reception line processing unit 70, and an ICMP error packet processing unit 80 are provided. Here, it is assumed that there is only one-way communication for transmitting data from the mobile terminal 1 to the fixed terminal 2, but
If a function on the receiving side is added to the mobile terminal 1, bidirectional communication can be supported.

The transmission-side upper-layer processing unit 10 is an element that performs processing of an application layer relating to applications such as file transfer, electronic mail, and WWW (World Wide Web). In this example, the transmission-side upper layer processing unit 10 is realized by software executed by a computer. It is. Sender upper layer processing unit 1
0 includes an upper layer transmission processing unit 11 for processing an application and a message buffer 12 for holding a message generated by the application.

The transmission side TCP processing section 20 is a section for executing the processing on the transmission side of TCP, and in this example, is realized by software executed by a computer. Sending TC
The P processing unit 20 includes a window flow control unit 21 and a DT
A packet creation unit 22, a retransmission packet buffer 23, and a retransmission timer management unit 24 are provided. The window flow control unit 21 controls a window size to avoid congestion, and is realized by software executed by a computer in this example. DT packet generator 2
2 creates a DT packet. DT packet is TCP
A packet corresponding to the segment of the term. The retransmission packet buffer 23 holds the transmitted packet in preparation for retransmission of the packet. The retransmission timer management unit 24 monitors the time when there is no response from the partner station after transmitting the packet, in order to manage whether or not retransmission is necessary.

The transmitting TCP processing unit 20 also includes a window monitoring timer for checking whether the window size of the receiving station has been updated, but is omitted in FIG. 1 because it is not related to the present invention. ing. Receiving TCP
The processing unit 30 is a part that executes processing on the receiving side of TCP, and is realized by software executed by a computer in this example. The receiving side TCP processing unit 30 includes A
The CK packet processing unit 31 is included. ACK
The packet processing unit 31 is provided with a connection state detection unit 32.

Transmission side TCP connection management table 4
0 holds the state value of each TCP connection.
The transmission-side TCP connection management table 40 holds data as shown in FIG. 6 or FIG. 7 for each connection. When only the window size is controlled in accordance with the cause of the packet discard, it is sufficient to provide the information shown in FIG. 6 on the transmission side TCP connection management table 40. However, when further control is to be performed, the information shown in FIG. It is necessary to add information held in the side TCP connection management table 40.

Each TCP connection set between the terminals is identified by a connection identifier including a source IP address 401, a destination IP address 402, a source port number 403, and a destination port number 404 shown in FIG. In the example of FIG. 6, the transmission side TCP connection management table 40 stores VT (S) 4
05, VT (A) 406, maximum data length D1, and connection state D2.

In the example of FIG. 7, the transmittable data length D3
DT packet transmission success confirmation number D4 is further added. VT (S) 405 is the transmission sequence number (unit: byte) of the DT packet to be transmitted next. Therefore, the number of bytes transmitted so far by the transmitting station is (VT (S)-
It is represented by 1). The initial value is 1.

VT (A) 406 is a reception sequence number (unit: byte) which is expected to be received next in order by the receiving station of the other party. Therefore, the receiving station of the other party is (VT (A)
This means that the number of bytes of the transmission sequence number up to -1) has been received. The initial value is 1. Maximum data length D1
Is the length of the user data field of the largest DT packet that can be transmitted in one transfer. This is the maximum segment size (MSS) in TCP terms.
Is equivalent to

The connection state D2 holds information indicating any of "normal", "congestion occurrence", and "bit error occurrence". The initial value is “normal”. Here, “normal”, “congestion occurrence”, and “bit error occurrence” mean a state where no congestion or bit error has occurred, a state where congestion has occurred, and a state where bit error has occurred in the relay device or link on the connection, respectively. It shows the state that it is doing.

The transmittable data length D3 is the transmittable DT
It indicates the length of the maximum user data field of the packet. The initial value of the transmittable data length D3 is the maximum data length D1. The number of DT packet transmission success confirmations D4
Means the number of DT packets for which transmission (including retransmission) has been confirmed to be successful after changing the transmittable data length D3. D
The initial value of the T packet transmission success confirmation number D4 is 0.

The window flow control unit 21 of the transmission-side TCP processing unit 20 controls the number of bytes that can be transmitted without acknowledgment of transmission from the partner station, that is, the window size (WS). The transmitting station permits transmission up to the number of bytes having a value up to (VT (A) + WS-1). This window size includes the number of bytes that can be processed (called the receiving window size) notified from the receiving station or the number of bytes that can be sent to avoid the occurrence of congestion in the network (called the congestion window size). , The smallest value is assigned.

In the DT packet creation unit 22, when permitted by the window flow control unit 21, the transmittable data length D3 of the transmission side TCP connection management table 40 is calculated based on the message stored in the message buffer 12 or the retransmission packet buffer 23. Retrieve data for the number of bytes not exceeding. And a DT having a user data field and a sequence number field.
Create a packet. The data extracted from the message buffer 12 is assigned to the user data field of the DT packet, and the transmission side TCP is assigned to the sequence number field.
VT (S) 4 held in the connection management table 40
Set the value of 05. Then, VT (S) 405 is changed to D
It is increased by the length of the user data field of the T packet.

When the discard of the DT packet is detected due to the timeout of the retransmission timer management unit 24, the window size is controlled. This control will be described later. When the discard of the DT packet is detected, the DT packet not received the acknowledgment from the receiving station is taken out from the retransmission packet buffer 23 and retransmitted. However,
The DT packet at the time of retransmission includes the D
The same packet as the T packet is retransmitted.

The transmitting-side IP processing unit 51 performs transmission-side TCP processing.
The IP of the TCP packet created by the processing unit 20
(Internet Protocol) header is added. The packet to which the IP header is added is input to the transmission-side line processing unit 60 and output to the upstream wireless line 91 via the upstream wireless line transmission packet buffer 61 and the upstream wireless line transmission control unit (corresponding to a line driver) 62. Is done.

Uplink radio line transmission packet buffer 61
Is a buffer memory for holding data waiting for transmission processing. The uplink radio channel transmission control unit 62 controls transmission to the uplink radio channel 91. The packet input from the downlink radio line 92, that is, the received packet is input to the reception-side line processing unit 70, and is transmitted to the reception-side IP processing unit 52 via the downlink radio-line reception control unit 72 and the downlink radio-line reception packet buffer 71. Is input to

The downlink radio channel reception control section 72 controls reception of data input from the downlink radio channel 92. The downlink wireless channel received packet buffer 71 is a buffer memory that holds received data waiting to be processed. The receiving-side IP processing unit 52 performs IP processing of the received packet, that is, processing such as error detection and rearrangement of the IP packets. The ACK packet processing unit 31 included in the receiving TCP processing unit 30 performs processing of the received acknowledgment packet, that is, an ACK (Acknowledgement) packet.

Upon receiving the ACK packet, the ACK packet processing unit 31 receives the acknowledgment sequence number (acknowledgement number) field value and the window size field value (indicating that the ACK packet has been correctly received by the other receiving station). Is set to VT (A) 406, which is the state value of the transmission-side TCP connection management table 40. Also, there is (VT (A) -1) in the packet buffer 23 for retransmission.
The data up to the number of bytes of the transmission sequence number corresponding to the value of is deleted.

The ACK packet processing unit 31 sends a delivery confirmation order number and an AC
The reception window size indicated in the window size field of the K packet is notified. When an ACK packet is received, the transmission-side TCP processing unit 20 identifies whether or not the same ACK packet (called a duplicate ACK) is continuously received. When the receiving station on the other end receives the DT packet out of sequence, it sends out a duplicate ACK. The transmission side TCP processing unit 20 detects this.

When the transmission-side TCP processing unit 20 receives the ACK packet, the window flow control unit 21 controls the window size. This control will be described later. When the window size is changed, the number of DT packets created and transmitted thereafter is limited to the updated window size.

When all three consecutive ACK packets are duplicate ACKs, the transmission-side TCP processing unit 20 may determine that congestion has occurred or a bit error has occurred in the partner station or the relay station. Is considered to be Further, the transmitting side TCP processing unit 20 performs the duplicate ACK
When the reception of the DT packet is detected, the DT packet for which the delivery confirmation has not been received from the other receiving station is retransmitted.

In the case of the conventional communication device, if all three consecutive ACK packets are duplicate ACKs, it is considered that congestion has occurred in the partner station or the relay station, and the congestion is unconditionally determined. Control is performed to reduce the window size. However, in a communication system such as that shown in FIG.
Even if a bit error occurs in the
An ate ACK occurs or a timeout occurs in the retransmission timer management unit 24.

When congestion actually occurs in a relay station or the like, controlling the number of packets to be transmitted by the transmitting station by performing congestion control is useful for avoiding congestion. However, in the case of performing the conventional control, when a bit error occurs in the wireless channel 4, the number of packets to be transmitted is suppressed due to the limitation of the window size, although congestion is not required to be avoided. Throughput decreases.

Therefore, in the mobile terminal 1 of FIG. 1, when detecting the discard of the transmitted DT packet, the connection state detecting unit 32 specifies the cause of the discard of the packet and performs control according to the difference between the causes. It has become. In this example, the relay station (for example, the radio base station 3) or the receiving station on the other side transmits information that can be used to identify the cause of the packet discard using at least one of the following techniques. Is assumed to be transferred to

For example, according to the technique described on page 317 of Document 1, when a large number of congestions or bit errors occur in the relay device, the relay station transmits the relay to the transmitting station of the connection passing through the relay device. If the device has an ICMP (Inte
rnet Control Message Protocol) Notification is made using an error packet (ICMP error packet method).

Also, reference 2 (M. Allan et al., “Ongoing T
CP Research Related to Satellites ”, http://www.ie
tf.org/internet-drafts/draft-ietf-tcpsat-res-issue
s-10.txt, August, 1999), when the occurrence of congestion is detected in a relay device, information of "occurrence of congestion" is recorded in the header of an IP packet passing through the relay device, and the IP The receiving station on the other side receiving the packet is T
An ACK packet in which the information is recorded in the CP header is returned to the transmitting station (forward congestion occurrence notification method: explicit con
gestion notification). That is, the receiving station transmits in the reverse direction of the packet flow.

According to the technique disclosed in Document 2, when the relay apparatus detects an increase in the packet discard rate due to a link bit error, information on "packet discard due to bit error (abnormal bit error occurrence)" is transmitted. The received ICMP error packet is transmitted to the receiving station of the connection passing through the relay device (forward bit error abnormality occurrence notification method). That is, the packet is transmitted in the forward direction of the flow of the packet. The receiving station that has received the ICMP error packet notifies the TCP transmitting station of the information using the “TCP option” in the TCP header.

Therefore, the transmitting station (mobile terminal 1) receives an ICMP error packet transmitted from a relay station (radio base station 3 or the like), an ACK packet from a receiving station (fixed terminal 2), or a receiving station (fixed terminal 2). By examining the "TCP option" from (1), it is possible to identify the presence or absence of "bit error occurrence" and the presence or absence of "congestion occurrence".

The ICMP error packet processing unit 80 shown in FIG. 1 processes an ICMP error packet input from the downlink radio line 92, and determines whether or not “abnormal bit error” and “congestion occurs” based on the information contained therein. Is identified. Further, according to the identification result, the ICMP error packet processing unit 80 changes the contents of the connection state D2 of the transmission side TCP connection management table 40 according to the state transition shown in FIG.

The connection state detection unit 32 included in the ACK packet processing unit 31 examines the TCP header of the received ACK packet, and identifies the presence / absence of “abnormal bit error” and the presence / absence of “congestion”. Also, according to the identification result, the connection state detection unit 32
Connection status D in the P connection management table 40
2 is changed according to the state transition shown in FIG.

That is, the content of the connection state D2 is one of "normal", "congestion occurrence", and "bit error abnormality occurrence". The initial value is “normal”. "Normal"
This indicates a state in which the relay device and the link on the connection are not in a congestion occurrence state and no bit error abnormality has occurred. “Congestion has occurred” indicates a state in which congestion has occurred in the relay device or link on the connection. Further, "abnormal bit error occurrence" indicates a state in which packet discarding due to a bit error frequently occurs.

Note that “reception of an ACK packet in a normal state” in FIG. 3 means reception of an ACK packet that does not include any information of “abnormal occurrence of bit error” and “occurrence of congestion”. ACK packet processing unit 31 of this mode
The operation (1) will be described with reference to FIG. When the mobile terminal 1 receives the ACK packet, the processing in FIG. 5 is executed. The processing shown in FIG. 5 corresponds to claims 1 and 2.

In step S20, the contents of the header of the received ACK packet are referred to. If the information of "abnormal bit error occurrence" or "congestion occurrence" is included in the header, the process proceeds from step S21 to S22. Furthermore,
If an ACK packet including the information of “abnormal bit error occurrence” is received, the process proceeds from step S22 to S24,
“Error bit error occurrence” is set in the connection state D2 of the transmission-side TCP connection management table 40.

If an ACK packet containing information of "occurrence of congestion" is received, the process proceeds from step S22 to step S23.
Then, "congestion occurrence" is set in the connection state D2 of the transmission-side TCP connection management table 40. On the other hand, when an ACK packet that does not include any information of “abnormal bit error occurrence” and “occurrence of congestion” is received,
The process proceeds from step S21 to S25, where "normal" is set in the connection state D2 of the transmission-side TCP connection management table 40.

In the next step S26, the transmission side TC
Transmittable data length D in P connection management table 40
Increase 3 However, if the transmittable data length D3 exceeds the upper limit D3max, steps S27 to S28
To the transmittable data length D3 and the upper limit value D3ma
x is set. The transmittable data length D3 of the transmission-side TCP connection management table 40 is a value used when the DT packet creation unit 22 creates a DT packet, and regulates the maximum user data field length.

In this example, the transmittable data length D3 becomes smaller when an “error in bit error” occurs, as described later, but the ACK not including any information of “error occurrence of bit error” and “occurrence of congestion”. When a packet is received, the transmittable data length D3 can be returned to the original size in the process of step S26. When the connection state D2 is normal, it is assumed that congestion has been alleviated as in the case of the conventional device. That is, since step S51 in FIG. 5 is executed, the congestion window size WS increases.

On the other hand, the transmission side TCP processing unit 20 detects the timeout in the retransmission timer
When the discard of the DT packet is detected by receiving the icate ACK, the processing shown in FIG. 4 is executed. The processing in FIG. 4 corresponds to claims 1 and 2. In step S10, the connection state is identified with reference to the connection state D2 of the transmission-side TCP connection management table 40. That is, if the connection state D2 is “abnormal bit error”, the process proceeds to step S11, and if the connection state D2 is “congestion occurs”, the process proceeds to step S14.

That is, when the DT packet is discarded due to the occurrence of congestion in the relay station or the like, the congestion window size (WS) becomes smaller each time step S14 is executed, as in the conventional apparatus. However, step S
By the processes of S15 and S16, the congestion window size is restricted to a predetermined lower limit value WSmin or more. Further, when a DT packet is discarded due to frequent occurrence of bit errors in the wireless channel 4, the transmittable data length D3 of the transmission-side TCP connection management table 40 becomes smaller each time step S11 is executed. However, step S1
2, the transmittable data length D3 is restricted to a predetermined minimum value D3min or more.

Therefore, the window size used for the packet transmission control of the mobile terminal 1 is DT due to the occurrence of congestion.
When the packet is discarded, the value is changed to a small value. However, when the DT packet is discarded due to frequent occurrence of bit errors, the window size does not change. For this reason,
When many bit errors occur, a drop in throughput due to a change in window size is prevented. If the DT packet is discarded due to frequent occurrence of bit errors, the transmittable data length D3 is changed to a small value in step S11. As shown in the literature 3, the packet discard rate due to a bit error increases as the packet length increases (when the bit error rate is sufficiently small, the packet discard rate is proportional to the packet length).

In this example, the DT packet creation unit 22 regulates the maximum length of the user data field of the DT packet to be created according to the transmittable data length D3 of the transmission-side TCP connection management table 40. When the transmittable data length D3 decreases, the length of the user data field of the DT packet transmitted from the mobile terminal 1 decreases. Therefore, an increase in the packet discard rate is suppressed.

When the connection state D2 returns to "normal" after the length of the user data field of the DT packet is shortened due to the occurrence of many bit errors, it is considered that the abnormal occurrence of the pit error has been alleviated, and steps S25 to S25 in FIG. S28
Is performed, the user data field length gradually returns to the original length every time an ACK packet is received.

When only the window size is controlled when a packet is discarded, the processing of steps S11 to S13 in FIG. 4 and steps S25 to S28 in FIG. 5 may be omitted. Even in that case, it is possible to suppress a decrease in throughput when the DT packet is discarded. "Congestion occurrence" and "Bit error abnormality occurrence"
Is provided in the relay station, the connection state D2 cannot be “normal” when step S10 is executed. However, if a relay station that does not have a function of notifying of “occurrence of congestion” and “occurrence of bit error” is used, step S1
When 0 is executed, the connection state D2 may become “normal”. In that case, the connection state D
It is desirable to execute congestion control even if 2 is “normal”.

In the above description, it is assumed that only one connection is allocated between mobile terminal 1 and fixed terminal 2. However, if information on a plurality of connections identified by connection identifiers is stored in the transmission-side TCP connection management table 40, communication control to which a plurality of connections are assigned can be handled.

The processes shown in FIGS. 8, 9 and 10 are all modifications of the process shown in FIG. Further, the processes shown in FIGS. 11 and 12 are both modifications of the process shown in FIG. The same processes are denoted by the same step numbers. The processing shown in FIG. 8 corresponds to claims 3 and 4, the processing shown in FIG. 9 corresponds to claim 5, and the processing shown in FIG.
The processing shown in FIG. 0 corresponds to claim 6, the processing shown in FIG. 11 corresponds to claim 7, and the processing shown in FIG. 12 corresponds to claim 8.

In each of the processes shown in FIGS. 8, 9, and 10, the value of the transmittable data length D3 is set to a predetermined reference value in order to facilitate management of the management addresses and the like of the message buffer 12 and the retransmission packet buffer 23. It is restricted to an integral multiple of the value Lo. As the reference value Lo, for example, 256 bytes which is an integral multiple of the reference unit (32 bits) of the TCP header may be used.

In the modification shown in FIG. 8, in step S11B, processing is performed to reduce the value of the transmittable data length D3 by the reference value Lo. When the processing in FIG. 8 is performed, it is necessary to change the content of step S26 in FIG. 5 so that the value of the transmittable data length D3 is increased by the reference value Lo. In this case, the lower limit D3min is replaced with the reference value Lo.
The same value as is assigned. Therefore, in the example of FIG. 8, the value of the transmittable data length D3 is restricted to only a value that is an integral multiple of the reference value Lo.

In the modification of FIG. 9, in step S11C, the reference value Lo is set to the transmittable data length D3. When the processing in FIG. 9 is performed, step S in FIG.
26, the value of the transmittable data length D3 is set to the reference value Lo.
Only need to be changed to increase. In this case, the lower limit D3min is assigned the same value as the reference value Lo. Therefore, in the example of FIG. 9, the value of the transmittable data length D3 is restricted to only a value that is an integral multiple of the reference value Lo.

In the modified example of FIG. 10, in step S11D, the maximum value Lx that does not exceed half the transmittable data length D3 among integer multiples of the reference value Lo is set as the transmittable data length D3. . When the processing in FIG. 10 is performed, it is necessary to change the content of step S26 in FIG. 5 so that the value of the transmittable data length D3 is increased by the reference value Lo. In this case, the lower limit D3min is replaced with the reference value L.
The same value as o is assigned. Therefore, in the example of FIG. 10, the value of the transmittable data length D3 is restricted to only a value that is an integral multiple of the reference value Lo.

Next, a modification of FIG. 11 will be described.
In step S31, VT (A), which is the state value held in the transmitting-side TCP connection management table 40, is obtained from the delivery confirmation sequence number field value of the received ACK packet.
The value obtained by subtracting 406 is set in the register Y. The value set in the register Y indicates the number of newly confirmed delivery bytes.

In the next step S32, the value of the number of DT packet transmission success confirmations D4 is increased by the value of the register Y. When the value of the DT packet transmission success confirmation number D4 is equal to or greater than the constant Np, the process proceeds from step S33 to S34. In step S34, the value of the transmittable data length D3 in the transmission-side TCP connection management table 40 is increased by the reference value Lo. In step S35, the DT packet transmission success confirmation number D4 in the transmission-side TCP connection management table 40 is set.
Is cleared to 0. The value of the DT packet transmission success confirmation number D4 in the transmission-side TCP connection management table 40 is initialized to 0 when a TCP connection is established.

Therefore, each time the transmission confirmation of the DT packet corresponding to the constant Np is successful, the value of the transmittable data length D3 increases by the reference value Lo. Therefore, the transmittable data length D
The value of 3 can be restricted to a value that is an integral multiple of the reference value Lo.

Next, a modification of FIG. 12 will be described.
In this example, steps S31 and S32 are executed as in FIG. In step S33B, when the number of DT packet transmission success confirmations D4 in the transmission-side TCP connection management table 40 becomes equal to or more than (D3 / Lo), the process proceeds to step S34. In step S34, the value of the transmittable data length D3 in the transmitting side TCP connection management table 40 is set to the reference value L.
Increase by o. Therefore, the value of the transmittable data length D3 increases at a parabolic rate.

[0085]

As described above, according to the present invention, even when packet discarding is detected, the window size does not become smaller when many bit errors occur, so that a decrease in throughput is suppressed. it can. In addition, when the number of bit errors frequently occurs, by reducing the maximum user data field length of the packet, it is possible to suppress an increase in the packet discard rate.

Also, by restricting the maximum length of the user data field of the packet to only a value that is an integral multiple of the reference value, it becomes easy to manage the address of the buffer memory, etc.
It is possible to prevent the configuration or processing of the device from becoming complicated. Furthermore, in the present invention, it is not necessary to generate new traffic in order to improve the throughput, and the present invention does not have any special influence on the control of the device on the receiving side, so that the device can be easily mounted.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a mobile terminal according to an embodiment.

FIG. 2 is a block diagram illustrating a configuration example of a communication system.

FIG. 3 is a schematic diagram showing a state transition of a connection state.

FIG. 4 is an operation (1) of a transmission-side TCP processing unit according to the embodiment;
It is a flowchart which shows.

FIG. 5 is a flowchart illustrating an operation (1) of the ACK packet processing unit according to the embodiment;

FIG. 6 is a schematic diagram illustrating a configuration example (1) of a transmission-side TCP connection management table.

FIG. 7 is a schematic diagram illustrating a configuration example (2) of a transmission-side TCP connection management table.

FIG. 8 shows an operation (2) of a TCP processing unit on the transmission side according to the embodiment;
It is a flowchart which shows.

FIG. 9 is an operation (3) of the transmission-side TCP processing unit according to the embodiment;
It is a flowchart which shows.

FIG. 10 is a flowchart illustrating an operation (4) of the transmitting-side TCP processing unit according to the embodiment;

FIG. 11 is a flowchart illustrating an operation (2) of the ACK packet processing unit according to the embodiment;

FIG. 12 is a flowchart illustrating an operation (3) of the ACK packet processing unit according to the embodiment.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Mobile terminal 2 Fixed terminal 3 Radio base station 4 Wireless line 5 Wired communication network 10 Transmission upper layer processing unit 11 Upper layer transmission processing unit 12 Message buffer 20 Transmission TCP processing unit 21 Window flow control unit 22 DT packet creation unit 23 Retransmission packet buffer 24 Retransmission timer management unit 30 Reception side TCP processing unit 31 ACK packet processing unit 32 Connection state detection unit 40 Transmission side TCP connection management table 51 Transmission side IP processing unit 52 Receiving side IP processing unit 60 Transmission side line processing unit Reference Signs List 61 Uplink wireless line transmission packet buffer 62 Uplink wireless line transmission control unit 70 Reception side line processing unit 71 Downlink wireless line reception packet buffer 72 Downlink wireless line reception control unit 80 ICMP error packet processing unit 91 Uplink wireless line 92 Downlink wireless line 401 Transmission Yuan I Address 402 Destination IP address 403 source port number 404 destination port number 405 VT (S) 406 VT (A) D1 maximum data length D2 connection state D3 transmittable data length D4 DT packet transmission success acknowledgment number

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 5K014 AA02 BA03 FA13 GA02 5K030 GA03 GA13 HA08 HC01 HC09 JL01 JL07 JT03 JT09 KA03 KA07 KA13 LA01 LC03 LC11 MA04 MB05 5K033 AA01 BA08 CB04 DA01 DA19 DB20 EA02 EA06 EA03 EA06 EA06 5K03 DD03 FF11 FF13 HH01 HH02 HH10 MM14 MM16

Claims (8)

[Claims] An apparatus transmits and receives data including user information to and from a partner station via a wireless communication line, receives a transmission confirmation packet from the partner station for data transmitted by the own station, and performs transmission confirmation. A communication device for limiting the amount of data that can be continuously transmitted without acknowledgment by a window size, a discard cause specifying means for specifying a cause of discard of data transmitted by the own station based on a signal transmitted by another station, When the discard of the data transmitted by the own station is detected, the control for the window size is distinguished between the case where the cause specified by the discard cause specifying means is the occurrence of congestion and the case where the bit error occurs, and the case where the bit error occurs. A window size control means for making the window size of the window larger than the window size when congestion occurs. . 2. The communication apparatus according to claim 1, wherein when detecting a discard due to a bit error occurring in the data transmitted by the own station, the maximum user data field length of the data packet transmitted by the own station is determined. A communication device, comprising: a user data field length control unit that limits the maximum user data field length when the frequency of occurrence of bit errors is reduced. 3. The communication device according to claim 2, wherein the maximum length of the user data field of the data packet transmitted by the own station is limited to an integral multiple of a predetermined fixed length. 4. The communication apparatus according to claim 3, wherein the user data field length control means detects discard of the data transmitted by the local station due to occurrence of a bit error.
The maximum user data field length of the data packet transmitted by the own station is changed to a value reduced by the fixed length, and each time the transmission of the data packet transmitted by the own station is confirmed to be successful,
A communication device, wherein a maximum user data field length is changed to a value increased by the fixed length. 5. The communication apparatus according to claim 3, wherein said user data field length control means detects data discarded by the own station when discarding the data transmitted by the own station due to occurrence of a bit error. Each time the maximum user data field length of the packet is changed to a value equal to the fixed length and the transmission of the data packet transmitted by the own station is confirmed,
A communication device, wherein a maximum user data field length is changed to a value increased by the fixed length. 6. The communication device according to claim 3, wherein said user data field length control means detects a discard of the data transmitted by the own station due to occurrence of a bit error. The maximum user data field length of the packet is changed to a maximum value that does not exceed half of the current maximum user data field length among the integral multiples of the fixed length, and the data packet transmitted by the own station is changed. A communication apparatus characterized in that the maximum user data field length is changed to a value increased by the fixed length every time transmission success is confirmed. 7. The communication apparatus according to claim 3, wherein the user data field length control means confirms the success of transmitting a predetermined number of data packets transmitted by the local station each time.
A communication device, wherein a maximum user data field length is changed to a value increased by the fixed length. 8. The communication apparatus according to claim 3, wherein said user data field length control means controls said fixed length number of data packets corresponding to a current maximum user data field length of a data packet transmitted by the own station. A communication apparatus characterized in that the maximum user data field length is changed to a value increased by the fixed length every time transmission success is confirmed.