JPH06121002A - Dynamic timeout control method - Google Patents

Abstract

(57) [Summary] [Purpose] A dynamic timeout control method that sets a timeout for determining whether the time from sending a message to receiving a response is normal. Set the timeout time to. [Configuration] In a message communication between terminal devices 1 and 2, which performs a message resending process when a response to a message transmission cannot be received within a timeout time set in a timeout time setting unit 3, a transmission time measuring unit 4
Is provided, the time from the first message transmission to the response reception is measured, and a margin is added to the measured time to set it in the timeout time setting unit 3 as an initial value of the timeout time. In the message communication process,
The update setting of this timeout time can be set.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a dynamic time-out time for setting a time-out time for determining whether a time from a message transmission to a response reception is normal in accordance with a time required for transmission between terminal devices. Regarding control method.
In message communication between terminal devices, when a message is transmitted from one terminal device to another terminal device, the receiving terminal device returns a response because the message is received. Therefore, the terminal device on the transmitting side can determine that the transmission of one message has ended by receiving the response. In this case, when the time from the message transmission exceeds the timeout time, the terminal device on the receiving side judges that the message is not received and performs the message resend process, so do not repeat this resend process. Therefore, it is necessary to set the optimum timeout time.

[0002]

2. Description of the Related Art When message communication is performed between terminal devices, it is common that the terminal devices are connected via a relay device such as an exchange. Therefore, each time the terminal devices are connected and message communication is performed, the number of relay devices such as exchanges connected between the terminal devices may be different or the transmission route may be different, resulting in a transmission delay. The time will be different.

FIG. 7 is an explanatory diagram of a relay communication system.
A relay device 73-1 such as an exchange is provided between the terminal devices 71 and 72.
73-n are connected, a message is transmitted from the terminal device 71, and the terminal device 72 receives the message,
The response is transmitted to the terminal device 71. When the terminal device 71 receives this response, it determines that the transmission of one message has ended, and transmits the next message. The message and the response are transmitted in the packet format, which causes a processing delay time Tr in the relay devices 73-1 to 73-n. Further, the terminal device 72 requires the processing time Tt required from the message reception to the response transmission.

Therefore, the time Td required for the terminal device 71 to send a message and receive a response.
When the number of relay devices 73-1 to 73-n is n, Td = 2 × n × Tr + Tt (1) When the response cannot be received within this time Td, it is assumed that the message has been lost and the retransmission processing is performed.

[0005]

As described above, the terminal device 71 performs the retransmission process when the response cannot be received within the time Td of the equation (1).
It is necessary to know the number n of the relay devices 73-1 to 73-n in order to obtain However, in a switching network including a relay device that connects the terminal devices 71 and 72, the route is often changed due to congestion or the like on the way. That is,
The number n of relay devices connected between the terminal devices 71 and 72 is not always constant. Therefore, it is possible to notify the number n of relay devices connected between the terminal devices 71 and 72 from the switching network side, but the switching network side does not have such a function.

Therefore, the terminal devices 71 and 72 set the time Td obtained from the expected maximum number of relay device connections to the timeout time, and when the response cannot be received within this timeout time from the message transmission, the retransmission process is performed. Is configured to do. However, when the number of relay devices connected is small, the time from message loss to retransmission processing becomes long, and the overall transmission efficiency decreases. On the contrary, if the timeout time is shortened, the number of times of retransmission processing increases when the number of relay device connections is large. In addition, since the timeout time for a short-distance terminal device and the timeout time for a long-distance terminal device are significantly different, the timeout time is set based on the expected maximum number of relay device connections for a large number of terminal devices. However, there is a drawback that the process becomes complicated. An object of the present invention is to measure the time from the actual message transmission to the response reception and set an optimal timeout time.

[0007]

SUMMARY OF THE INVENTION The dynamic time-out of the present invention.
The time control method is a message from terminal device 1 to terminal device 2.
Message is sent and the terminal device 2 receives the message
The response is transmitted to the terminal device 1, and the terminal device 1
End of one message transmission by receiving
And the response is set in the timeout time setting unit 3.
If you cannot receive the message within the specified timeout period,
In the message communication method for resending messages,
Measure the time from sending a message to receiving a response
The transmission time measurement unit 4 is provided, and the transmission time measurement unit 4
From sending the first measured message to receiving a response.
Time measured at and the margin added to this time
To the timeout time setting section 3
Set as a value.

After setting the timeout time as an initial value in the timeout time setting unit 3, the transmission time measuring unit 4 measures the time from the message transmission to the response reception for each message transmission, and based on the measurement result, It is determined whether or not the response can be received within the timeout time set in the timeout time setting unit 3. When the response can be received within the timeout time, update setting is performed so that the timeout time is shortened, and within the timeout time. If it cannot be received, update setting is made so that the timeout time becomes longer.

After setting the timeout time as an initial value in the timeout time setting unit 3, the transmission time measuring unit 4 measures the time from the message transmission to the response reception every time a message is transmitted, and the measured result is counted a predetermined number of times. An average value is obtained, and the time obtained by adding a margin to this average value is set in the timeout time setting unit 3 as a new timeout time.

When a response cannot be received within the time-out time set in the time-out time setting unit 3,
The time obtained by multiplying the timeout time by a coefficient or adding a fixed number is set in the timeout time setting unit 3 as a new timeout time.

[0011]

The time required until the first message is transmitted from the terminal device 1 to the terminal device 2 and the response from the terminal device 2 to the message is received is measured by the transmission time measuring unit 4, and the measured time is set to the measured time. A margin is added to obtain a timeout time, which is set in the timeout time setting unit 3 as an initial value. Therefore, even when the number of relay devices connected between the terminal devices 1 and 2 may differ, the timeout time can be set as a value corresponding to the number of connections.

After setting the timeout time as the initial value in the timeout time setting unit 3, the time until the response is received is measured by the transmission time measuring unit 4 for each message transmission, and the response can be received within the initial timeout time. When the initial value is too large, on the contrary, when the response cannot be received within the initial timeout time, it means that the initial value is too small. Optimize.

Further, the average value of the time from the message transmission to the response reception by the transmission time measuring unit 4 is obtained, and the time of this average value is set as a new timeout time. As a result, it is possible to optimize the timeout time by averaging the time variation due to the route change etc. for each message transmission.

Further, if the resend processing is performed when the response cannot be received within the timeout time, the time from the previous message transmission to the response reception cannot be measured. Therefore, the timeout time set in the timeout time setting unit 3 is multiplied by a predetermined coefficient or a fixed number is added to obtain a new timeout time longer than the previous timeout time, and the timeout time setting unit 3 is determined. Set.

[0015]

FIG. 2 is an explanatory view of an embodiment of the present invention, showing a main part of a terminal device, 11 is a control unit, 12 is a message transmitting unit, 13 is a message receiving unit, 14 is a timer, 15
Is a timer control unit, 16 is a margin addition unit, 17 is a register, 18 is a timeout detection unit, and 19 is a retransmission processing unit. The register 17 corresponds to the timeout time setting unit 3, and the timer 14 and the timer control unit 15 correspond to the transmission time measuring unit 4.

The timer control unit 15 clears the timer 14 by starting communication, starts the timer 14 in response to a message transmission instruction to be added to the message transmitting unit 12, and stops the timer 14 by receiving a response reception notification from the message receiving unit 13. The control is performed. Therefore, the value of the timer 14 indicates the measurement time from the message transmission to the response reception.

The margin adding unit 16 adds a margin to the timer value to obtain a time-out time, and the register 1
The margin can be set to a predetermined fixed time or 20% of the timer value, for example. The timeout detector 18 compares the timer value with the timeout time set in the register 17,
When a response cannot be received within the timeout period,
When it is determined that the message has been lost, the resending processing unit 19 is activated and the resending process of the message is performed.

FIG. 3 is a flow chart of an embodiment of the present invention, in which communication start and communication are performed from step (a1) to step (a1).
(A5) and (b1) to (b5), the timer controller 15 controls the timer 14 as described above when communication is started.
Is cleared (a1). Then, the timer 14 is started by the message transmission instruction (a2). It is determined whether a response to this message transmission is received (a
3) Upon reception of a response, that is, upon receipt of a notification from the message receiving unit 13, the timer control unit 15 stops the timer 14 (a4). The timer value of the timer 14 is read, the margin is added by the margin adding unit 16, and the result is written into the register 17 (a5). That is, the initial value of the timeout time is set.

After setting the timeout time in the register 17 and before sending the message, the timer control unit 15 clears the timer 14 (b1), starts the timer 14 according to the message sending instruction (b2), and determines whether or not a response is received. Is judged (b3), and if a response is received within the timeout time, the processing ends normally. When the response cannot be received, the timer value tt and the timeout time tr set in the register 17 (b4) are checked to see if tt <tr (b5). If tt <tr, the time is within the timeout time tr, and the process proceeds to step (b3). If tt ≧ tr, it means that the response cannot be received within the timeout time tr. Therefore, the retransmission processing unit 19 is activated to perform the message retransmission processing. Then, the process proceeds to step (b1).

Therefore, the time from the message transmission to the response reception is measured, the margin is added to make the timeout time, and this timeout time is set in the register 17, and the response reception for the subsequent message transmission is monitored and the message is received. Retransmission processing can be performed by the loss judgment. In that case, the timeout time is set by measuring the time from message transmission to response reception when the other terminal device is connected, so even if the number of relay device connections differs for each communication, it is actually Since the timeout time is set on the basis of the transmission time of, it is not necessary to consider the number of connected relay devices.

In the above-described embodiment, the transmission time is measured at the start of communication, the timeout time is set, and then the message loss determination is performed using this timeout time. Since the timer 14 is started for each message transmission and the time until the response is received is measured, when the response is received within the timeout time, the timeout time is updated so that the timeout time becomes shorter. If the response cannot be received within the timeout time,
By updating the timeout time so that the timeout time becomes longer, the optimization of the timeout time can be promoted in the process of continuing the message communication.

FIG. 4 is an explanatory view of another embodiment of the present invention, in which 21 is a control unit of the terminal device, 22 is a message transmission unit, 23 is a message reception unit, 24 is a timer, 25 is a timer control unit, 26. Is an adder, 27 is a register, 28 is a timeout detection unit, 29 is a retransmission processing unit, 31 is an adder,
32 is a divider (÷ m), 33-1 to 33-m are flip-flops FF1 to FFm, 34 is a multiplier, 35 is an adder, and the flip-flops 33-1 to 33-m, the adder 31 and the divider 31 With the instrument 32, the average value of the measurement times of m times is obtained. Therefore, another known configuration for calculating the average value or the arithmetic function of the processor can be used.

Further, the timer control unit 25 controls the clearing, starting and stopping of the timer 24, and starts the timer 24 in response to a transmission instruction to the message transmitting unit 22 in the same manner as in the above-mentioned embodiment, and the message receiving unit 23 The timer 24 is stopped by the reception notification of the response, and the time from the message transmission to the response reception is measured.

Also, a margin is added to the measurement time from the first message transmission to the response reception at the start of communication to make the initial value of the timeout time, and this is used as the register 27.
Set to. The time until the reception of a response for each subsequent message transmission is measured by the timer 24, and is stored in the shift register including the flip-flops 33-1 to 33-m.
By inputting the timer value of the measurement time, adding the outputs of the respective stages of the shift register to the adder 31, adding them, and dividing by the divider 32, the average value of m times can be obtained.

A margin is added to this average value by the adder 26, and a new timeout time is set in the register 27. Accordingly, even if the congestion of the relay device is encountered and a change that temporarily lengthens the transmission time occurs, the update setting of the timeout time can be optimized.

FIG. 5 is a flowchart of another embodiment of the present invention, in which the timer control unit 25 clears the timer 24 at the start of communication (c1), and the timer 24 is started by a message transmission instruction to the message transmission unit 22. (C2), the message receiving unit 23 determines whether or not a response is received (c3), and the timer 24 is stopped by the reception notification by the response reception (c).
4). Then, the timer value is read and the flip-flop F
Write to F1 to FFm, calculate average value, add margin, and write timeout time to register 27 (c
5). This shows the case where the value obtained by adding the margin to the average value of m times is used as the initial value of the timeout time. However, as in the above-described embodiment, the margin is added based on the result of the first transmission time measurement. It can be the initial value of the timeout time.

FIG. 6 is a flowchart of another embodiment of the present invention during communication, in which the timer 24 is cleared (d1) before message transmission and the timer 24 is instructed by a message transmission instruction.
Is started (d2), it is determined whether or not a response is received (d3), and when the response is received, the contents of the flip-flops FF1 to FFm are sequentially shifted, and the first stage flip-flop FFm is measured by the timer 24. The timer value is set, the average value is calculated, the margin is added, and a new timeout time is calculated and written in the register 27 (d4).

When the response cannot be received, the timer value is set to tt, the register value is set to tr (d5), and tt <t.
It is determined whether or not r (d6). If tt <tr, it means that it is within the time-out period. Therefore, the process proceeds to step (d3). When tt <tr is not satisfied, that is, when the response cannot be received within the timeout time, it is determined that the message has been lost, the retransmission processing unit 29 is activated, the message is retransmitted, and tt × α is set to a new tr or tt + β is set as a new tr in the register 27 (d7). That is,
The timer value tt when the timeout is detected indicates the timeout time (tr) set in the register 27.
This is multiplied by a coefficient α (> 1) by the multiplier 34 and set in the register 27 as a new register value tr (timeout time). Alternatively, the constant value β is set to the timer value tt
(= Register value tr) is added by the adder 35 and set in the register 27 as a new register value tr (timeout time).

Therefore, even if congestion occurs in the relay device on the way during message communication and the transmission time fluctuates, the timeout time corresponding to it can be set. Is no longer set, and message loss can be detected early and retransmission processing can be performed.

[0030]

As described above, according to the present invention, the message is transmitted from the terminal device 1 to the terminal device 2,
The time from the message transmission to the response reception is measured by the transmission time measuring unit 4, a margin is added to the measurement result, and the result is set as the timeout time in the timeout time setting unit 3.
Even if the connection paths between the terminal devices 1 and 2 are different, there is an advantage that the transmission time corresponding to the connection path can be measured and the optimum timeout time can be automatically set based on the result. Therefore, it is possible to detect the loss of the message at an early stage and perform the retransmission processing, and there is an advantage that the overall transmission efficiency can be improved.

Further, the initial value of the timeout time is set based on the first measurement result of the transmission time, and the timeout time is updated and set based on the measurement result of the transmission time for each subsequent message transmission. There is an advantage that an optimal time-out period can be set and unnecessary retransmission processing can be prevented even if the transmission time fluctuates due to congestion of the relay device.

[Brief description of drawings]

FIG. 1 is a diagram illustrating the principle of the present invention.

FIG. 2 is an explanatory diagram of an embodiment of the present invention.

FIG. 3 is a flowchart of an embodiment of the present invention.

FIG. 4 is an explanatory diagram of another embodiment of the present invention.

FIG. 5 is a flowchart of another embodiment of the present invention.

FIG. 6 is a flowchart of another embodiment of the present invention.

FIG. 7 is an explanatory diagram of a relay communication system.

[Explanation of symbols]

 1, 2 terminal device 3 timeout time setting unit 4 transmission time measuring unit

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Katsumasa Kinoshita 1015 Kamiodanaka, Nakahara-ku, Kawasaki City, Kanagawa Prefecture, Fujitsu Limited (72) Inventor, Kiichi Matsuda 1015, Kamedotachu, Nakahara-ku, Kawasaki City, Kanagawa Prefecture, Fujitsu Limited

Claims (4)

[Claims] 1. A message is transmitted from one of the terminal devices (1) and (2) to the other terminal device, and the other terminal device returns a response upon receipt of the message,
When the one terminal device receives the response, it is determined that the transmission of one message has ended, and when the response cannot be received within the timeout time set in the timeout time setting unit (3), the message resend processing is performed. In the message communication system for performing the above, a transmission time measuring unit (4) for measuring the time from the message transmission to the reception of the response to the message is provided, and from the first message transmission measured by the transmission time measuring unit (4) A dynamic time-out time control method characterized in that a time until a response is received is measured, and a time obtained by adding a margin to the time is set as the initial value of the time-out time in the time-out time setting unit (3). 2. After setting the timeout time as an initial value in the timeout time setting unit (3), the transmission time measuring unit (4) measures the time from message transmission to response reception for each message transmission. Based on the result of the measurement, it is determined whether or not the response can be received within the time-out time set in the time-out time setting unit (3). When the response can be received within the time-out time, the time-out is performed. 2. The dynamic timeout time control method according to claim 1, wherein updating is set so that the time becomes short, and when reception is not possible within the timeout time, updating is set so that the timeout time becomes long. 3. After setting the timeout time as an initial value in the timeout time setting unit (3), the transmission time measuring unit (4) measures the time from message transmission to response reception for each message transmission. 3. The average value of a predetermined number of times of the measurement results is obtained, and a time obtained by adding a margin to the average value is set in the timeout time setting unit (3) as a new timeout time. Dynamic Timeout Control Method for Internet. 4. When the response cannot be received within the timeout time set in the timeout time setting unit (3), the timeout time is multiplied by a coefficient or a fixed number is added as a new timeout time to perform the timeout. The dynamic time-out time control method according to claim 1, wherein the time-out setting unit (3) sets the time-out unit.