JP5448793B2 - Communication equipment - Google Patents

Description

  The present invention relates to a communication device that is connected to a terminal and transmits a data packet from the terminal to an external network by a token bucket method.

  2. Description of the Related Art Conventionally, communication devices that are connected to a terminal such as a PC (Personal Computer) and perform wireless communication are known. In such a communication device, there is a method called a token bucket that controls a communication speed (throughput) so as not to use excessive network resources.

  The token bucket mechanism includes a packet buffer that accumulates input packets and a token buffer (token bucket) that accumulates tokens. When a predetermined amount of tokens accumulates in the token buffer, packets are output from the packet buffer, and the token buffer tokens decrease. In this way, the throughput of packets output to the external network is controlled by the amount of tokens replenished in the token buffer per unit time (token rate). This token rate dynamically changes depending on predetermined parameters and the network environment, and is adjusted to be low when the network is congested and high when the network is free.

  For example, Patent Document 1 proposes a method of calculating a token rate based on packet throughput at a predetermined timing based on token increase / decrease states.

  One of the congestion control methods of TCP (Transmission Control Protocol) is ECN (Explicit Congestion Notification). ECN is a method for avoiding congestion by limiting the transmission speed by allowing the TCP sender to detect congestion and notifying the TCP sender of congestion information.

JP 2007-221529 A

  By the way, when the communication device adopts the token bucket method, when the congestion avoiding process by ECN is performed on the terminal side (TCP sender), the input throughput of packets from the terminal to the communication device decreases. However, according to the method of Patent Document 1, the calculation is performed at a predetermined timing based on the increase / decrease state of the token, and the calculation result may not be the token rate based on the input throughput itself after the change due to congestion avoidance. .

  Therefore, in the method of Patent Document 1, there is a delay until the decrease in input throughput due to congestion avoidance is reflected in the token rate, and the output throughput after change is not used, so that the output throughput can be sufficiently controlled. There wasn't.

  As a result, for example, if packets are accumulated in the packet buffer when congestion occurs, this packet may be transmitted and discarded by the congested external network regardless of congestion avoidance by ECN.

  An object of the present invention is to provide a communication device capable of controlling the output throughput of data packets without delay even when congestion occurs.

Communication device according to the present invention is a communication device for transmitting an interface section connected to an external device, the data inputted from the interface unit in the token bucket scheme to the external network, received from the external network and with reference to the information, a determining section whether the occurrence of congestion in the external network, and throughput detection unit for detecting an input throughput of data from the interface unit, the congestion by the determination unit If it is determined that there is no occur, based on the statistical value of the detected said input throughput by the throughput detection unit, determines a token rate before Symbol token bucket scheme, occurrence of the congestion by the determination unit If it is determined that there is, the token rate is determined based on the input throughput immediately after the determination. And a determination unit for determining a.
In addition, the determination unit determines whether or not congestion has occurred in the external network by referring to header information of a packet received from the external network and determining whether or not there is data indicating the occurrence of congestion in the external network. It is preferable to do.

Moreover, it is preferable that the determination unit calculates an exponential smoothed average value of the input throughput as the statistical value.
In addition, when the determination unit determines that the congestion occurs, the determination unit preferably rewrites the exponential smoothed average value to the input throughput immediately after the determination.

Further, it is preferable that the determination unit calculates a maximum throughput or a final throughput in a predetermined period as the statistical value.

A communication device according to the present invention is a communication device that outputs data input from an external device to an external network, and a determination unit that determines whether or not congestion has occurred in the external network, and is input from the external device A throughput detection unit that detects the input throughput of data, and a control unit that controls the throughput of outputting data to an external network based on the input throughput immediately after the determination when the determination unit determines that the congestion occurs And comprising.

  According to the present invention, the output throughput of data packets can be controlled without delay even when congestion occurs.

It is a block diagram which shows the function structure of the PC card which concerns on embodiment of this invention. It is a block diagram which shows the structure of the upstream data flow in the token bucket mechanism which concerns on embodiment of this invention. It is a sequence diagram of congestion control by ECN according to an embodiment of the present invention. It is a figure which shows the example of the time transition of the exponential smoothing average value which concerns on embodiment of this invention. It is a flowchart which shows the monitoring process of the throughput which concerns on embodiment of this invention. It is a flowchart which shows the determination process of the token rate which concerns on embodiment of this invention.

  Hereinafter, an example of an embodiment of the present invention will be described. In the present embodiment, a PC card 1 connected to a PC 2 (terminal) will be described as an example of a communication device. In addition to the PC card 1, the communication device is connected to an external terminal such as a mobile phone, a PHS (registered trademark), a USB terminal, an in-vehicle communication module, a vending machine communication module, or the like for wireless data communication. It may be various communication devices that provide functions.

FIG. 1 is a block diagram showing a functional configuration of the PC card 1 according to the present embodiment.
The PC card 1 includes a control unit 10, a storage unit 20, an interface unit 30, a communication unit 40, and a token bucket mechanism 50.

  The control unit 10 controls the entire PC card 1, and performs predetermined control on the interface unit 30, the communication unit 40, the token bucket mechanism 50, and the like, for example. In addition, the control unit 10 receives input from the interface unit 30, the communication unit 40, and the like, and executes various processes. And the control part 10 controls the memory | storage part 20 in the case of a process execution, reads various programs and data, and writes data.

  Further, the control unit 10 includes a throughput detection unit 11, a token detection unit 12, a determination unit 13, and a determination unit 14. And the control part 10 determines the token rate in the token bucket mechanism 50 according to a communication condition, and controls the communication speed of PC2. In addition, the function of each part with which the control part 10 is provided is mentioned later.

  The storage unit 20 includes, for example, a working memory and is used for arithmetic processing by the control unit 10. Further, in the present embodiment, a program for executing processing to be described later, various data, and the like are stored. The storage unit 20 may also serve as a removable external memory.

  The interface unit 30 is an interface that transmits and receives data packets to and from the PC 2. The interface unit 30 inputs a packet received from the PC 2 to the token bucket mechanism 50 and transmits a packet output from the token bucket mechanism 50 to the PC 2.

  The communication unit 40 communicates with an external device (base station) in a predetermined frequency band (for example, 2 GHz band or 800 MHz band). The communication unit 40 demodulates the signal received from the antenna 41, supplies the processed signal to the token bucket mechanism 50, modulates the signal supplied from the token bucket mechanism 50, and Send to external device.

  The token bucket mechanism 50 controls the packet throughput of the upstream data flow from the PC 2 to the external network (base station) and the downstream data flow from the external network (base station) to the PC 2.

  FIG. 2 is a block diagram showing the configuration of the upstream data flow in the token bucket mechanism 50 according to this embodiment.

  First, input packets from the PC 2 are stored in the packet buffer 51. Further, tokens are added to the token buffer 52 at regular intervals (for example, 30 msec).

  If there is a sufficient amount of tokens in the token buffer 52, the token bucket mechanism 50 outputs the packets stored in the packet buffer 51 to the external network, and decreases the tokens in the token buffer 52. Further, when there are no more tokens in the token buffer 52, packets are not output until the tokens are accumulated, and continue to accumulate in the packet buffer 51.

  The configuration of the downstream data flow in the token bucket mechanism 50 is the same, and “input from PC 2” in FIG. 2 is read as “input from network”, and “output to network” is “output to PC 2”. Is read as

Returning to FIG. 1, the function of each part of the control part 10 which concerns on this embodiment is explained in full detail.
The throughput detector 11 detects the input throughput of the data packet from the PC 2 that passes through the interface unit 30 to the token bucket mechanism 50. Then, the detected input throughput is notified to the determination unit 14.

  The token detection unit 12 detects a change in the token amount held in the token buffer 52. Specifically, the token detection unit 12 detects that the increase in the token held in the token buffer 52 has continued for a predetermined time or more and that the token has shifted from the token decrease state to the token depletion state. Then, the detection result is notified to the determination unit 14.

  The determination unit 13 refers to the TCP header (header information) of the packet received from the external network by the communication unit 40 and determines the presence / absence of data indicating the occurrence of congestion in the external network (an ECE flag described later).

Here, the congestion control by ECN will be described.
FIG. 3 is a sequence diagram of congestion control by ECN according to the present embodiment. In this example, a TCP sender (for example, PC2) and a TCP receiver communicate with each other via a router.

  The router determines that the network is congested when the amount of use of the buffer for storing packets waiting for transfer exceeds a certain threshold. When the router determines that the network is congested, it sets a CE flag indicating that it is experiencing congestion in the ToS field of the IP header of the packet, and forwards the packet to the TCP receiver (step S1).

  When the TCP receiver receives a packet in which the CE flag is set in the IP header, the TCP receiver sets an ECE (ECN-Echo) flag in the TCP header and transmits an acknowledgment packet (Ack) to the TCP sender. This is continued until a packet having a CWR (Congestion Window Reduced) flag in the TCP header is received from the TCP sender (step S2).

  When a TCP sender receives a packet with an ECE flag set in the TCP header, the TCP sender determines that congestion has occurred. Then, the TCP sender reduces the congestion window size to the slow start threshold value to lower the transmission rate, and sets the CWR flag in the TCP header of the packet to be transmitted after that (step S3).

  When the TCP receiver receives a packet having the CWR flag set in the TCP header, the TCP receiver stops setting the ECE flag in the acknowledgment packet (step S4).

  The determination unit 13 determines whether there is an ECE flag in the TCP header in step S2 as data indicating the occurrence of congestion in the external network.

  When the determination unit 13 determines that there is no ECE flag as data indicating the occurrence of congestion, the determination unit 14 determines the token rate of the token buffer 52 based on the statistical value of the input throughput detected by the throughput detection unit 11. decide. On the other hand, when the determination unit 13 determines that the ECE flag is present, the determination unit 14 determines the token rate based on the input throughput immediately after the determination.

  Specifically, the determination unit 14 determines the token rate by the following calculation method (1) or (2), for example.

(1) In the case of simple calculation The determination unit 14 acquires, as the above statistical value, the throughput at the time when the token reduction state ends or the maximum throughput in the token reduction state, and sets an allowable coefficient for the acquired throughput. The multiplied value is determined as the updated token rate.
It is assumed that the ratio allowed for the token rate with respect to the input throughput is called an allowable coefficient, and the allowable coefficient is set in advance and stored in the storage unit 20.

  Further, the determination unit 14 obtains, as the above statistical value, the throughput when a certain time has elapsed from the time when the token increase state started, or the maximum throughput during this time, and multiplies the obtained throughput by an allowable coefficient. The value is determined as the updated token rate.

  Here, when the determination unit 13 determines that the ECE flag is present, at the time of determination, the determination unit 14 determines a value obtained by multiplying the input throughput immediately after the determination by the allowable coefficient as the updated token rate. To do.

(2) In the case of exponential smoothing averaging process The determination unit 14 performs exponential smoothing averaging process on the throughput periodically sampled during the token reduction state. When the token reduction state ends, the determination unit 14 acquires the maximum value or the last value after the exponential smoothing averaging process as the above statistical value, and multiplies the acquired statistical value by an allowable coefficient. Determined as the updated token rate.

Here, if the n-th sample value X _n throughput is obtained at time t _n, the values A _n after exponential smoothing averaging process,
A _n = (1-C) A _n-1 + CX _n
C = 1−exp (−1 / T) Δt
Δt = t _n −t _n−1
As required. T is an arbitrary constant.
Further, the obtained exponential smoothed average value (A _n ) may be stored in the storage unit 20.

  Further, the determination unit 14 performs exponential smoothing averaging processing on the throughput periodically sampled during the token increase state. The determination unit 14 acquires the maximum value or the last value after the exponential smoothing averaging process as the above statistical value when a certain time has elapsed from the time when the token increase state starts, and the acquired statistical value A value obtained by multiplying the allowable coefficient is determined as the updated token rate.

Here, when it is determined by the determination unit 13 that the ECE flag is present, at the time of determination, the determination unit 14 determines the exponential smoothed average value (A _n ) stored in the storage unit 20 immediately after this determination. Rewrite to the input throughput of. Then, the determination unit 14 determines a value obtained by multiplying the rewritten exponential smoothed average value (A _n ) by an allowable coefficient as the updated token rate.

FIG. 4 is a diagram illustrating an example of temporal transition of the exponential smoothed average value according to the present embodiment.
When there is no congestion in the external network, the exponential smoothing average value transitions the curve indicated by A in the figure. In this case, even if there is a sudden change in throughput, the past throughput is smoothed, and the calculated value changes smoothly.

On the other hand, congestion occurs in the external network, when it is determined that the ECE flag exists by the determination unit 13 at time t _1, the exponential smoothing average, transitions curve indicated by B in FIG. In this case, the exponential smoothing average value changes rapidly at time t1, but the determination unit 14 can determine the token rate corresponding to the current throughput changed for avoiding congestion. Thereafter, the curve B gradually approaches the curve A with time.

  Furthermore, it is desirable that the storage unit 20 stores a maximum token rate and a minimum token rate in advance. In this case, the determination unit 14 determines the maximum token rate when the maximum token rate is exceeded and the minimum token rate when the maximum token rate is exceeded, regardless of the calculation methods (1) and (2) described above. Determined as the updated token rate.

Hereinafter, the processing flow of the control unit 10 using the exponential smoothing average throughput will be described.
FIG. 5 is a flowchart showing throughput monitoring processing according to the present embodiment.

  In step S <b> 1, the control unit 10 (throughput detection unit 11) detects the input throughput from the PC 2 to the token bucket mechanism 50.

  In step S <b> 2, the control unit 10 (determination unit 14) calculates an exponential smoothed average throughput using the input throughput detected in step S <b> 1 and the previously calculated value stored in the storage unit 20.

  In step S3, the control unit 10 stores the current input throughput detected in step S1 and the exponential smoothed average throughput calculated in step S2 in the storage unit 20.

  In step S4, the control unit 10 detects the presence or absence of a predetermined event such as the end of communication, and determines whether or not to end the monitoring process. If this determination is YES, the control unit 10 ends the process. If the determination is NO, the control unit 10 returns the process to step S1 and continues the monitoring process.

FIG. 6 is a flowchart showing token rate determination processing according to the present embodiment.
In step S11, the control unit 10 (determination unit 13) refers to the TCP header of the packet received from the external network by the communication unit 40, and determines the presence or absence of the ECE flag, that is, whether or not congestion is being avoided. If this determination is YES, the control unit 10 moves the process to step S12. If the determination is NO, the control unit 10 moves the process to step S13.

  In step S12, the control unit 10 (determination unit 14) determines that the input throughput from the PC 2 is suppressed to avoid congestion, and the index stored in the storage unit 20 by the throughput monitoring process (FIG. 5). Rewrite the smoothed average throughput with the current input throughput.

  In step S13, the control unit 10 (determination unit 14) determines whether or not it is a token rate determination timing. Specifically, the control unit 10 detects that the token detection unit 12 has continued to increase the number of tokens held in the token buffer 52 for a predetermined time or has transitioned from a token depletion state to a token depletion state. It is determined whether or not it has been done. When this determination is YES, the control unit 10 moves the process to step S14, and when the determination is NO, the control unit 10 returns the process to step S11.

  In step S <b> 14, the control unit 10 (determination unit 14) calculates and determines a token rate based on the exponential smoothed average throughput stored in the storage unit 20.

  In step S15, the control unit 10 detects the presence or absence of a predetermined event such as the end of communication, and determines whether or not to end the throughput determination process. If this determination is YES, the control unit 10 ends the process. If the determination is NO, the control unit 10 returns the process to step S11 and continues the process.

  As described above, according to the present embodiment, the ECN congestion avoidance operation can be prevented from being delayed by the token bucket mechanism 50. Therefore, the PC card 1 can control the output throughput of data packets without delay even when congestion occurs in the external network.

  Further, when using the exponential smoothed average throughput, the PC card 1 determines the token rate by resetting the exponential smoothed average throughput to the current value only when congestion occurs. Thereafter, the PC card 1 can determine the token rate at a normal timing by using the exponentially smoothed average throughput that is periodically updated without monitoring the completion of the congestion control.

  As mentioned above, although preferred embodiment was described, this invention can be implemented with a various form, without being limited to embodiment mentioned above. In addition, the effects described in the above-described embodiments are merely a list of the most preferable effects resulting from the present invention, and the effects according to the present invention are not limited to those described above.

  In the above-described embodiment, the determination of whether or not congestion is being avoided (step S11) is regularly repeated in the token rate determination process, but is not limited thereto. After the exponential smoothing average throughput is reset to the current value (step S12) and the token rate is calculated (step S13), the judgment (step S11) during congestion avoidance is made until the next timing in step S13. It may be omitted.

  In the above-described embodiment, the case where the input throughput varies due to the ECN congestion avoidance process has been described. However, the input throughput may vary greatly due to other reasons. The control unit 10 may monitor the amount of fluctuation of the input throughput, and may determine the token rate based on the input throughput after the fluctuation at the timing of the fluctuation when there is a fluctuation exceeding a predetermined fluctuation threshold. . As a result, the control unit 10 can quickly reflect the fluctuation of the input throughput on the token rate.

1 PC card (communication equipment)
2 PC (terminal)
DESCRIPTION OF SYMBOLS 10 Control part 11 Throughput detection part 12 Token detection part 13 Determination part 14 Determination part 20 Storage part 30 Interface part 40 Communication part 41 Antenna 50 Token bucket mechanism 51 Packet buffer 52 Token buffer

Claims (6)

And an interface section connected to an external device, a communication device for transmitting the data inputted from the interface unit in the token bucket scheme to the external network,
Said with reference to the information received from the external network, determining unit whether the occurrence of congestion in the external network,
And throughput detection unit for detecting an input throughput of data from the interface unit,
If it is determined that there is no occurrence of the congestion by the determination unit, based on the statistics of the input throughput detected by the throughput detection unit, determines a token rate before Symbol token bucket scheme, the determination unit communication device and a determination unit configured to determine the token rate based on an input throughput immediately after the case of occurrence of congestion is determined that there, the determination by. The determination unit determines whether or not congestion has occurred in the external network by referring to header information of a packet received from the external network and determining whether or not there is data indicating the occurrence of congestion in the external network. Item 2. The communication device according to Item 1. The determining unit, as the statistical value, the communication apparatus according to claim 1 or 2 for calculating the exponential smoothing average value of said input throughput. The determination unit is configured when it is determined that the occurrence of the congestion is present by judging unit, a communication device according to claim 3 in which the exponential smoothing average, rewrites the input throughput immediately after the determination. The determining unit, as the statistical value, the communication apparatus according to claim 1 or 2 calculates the maximum throughput or final throughput in a predetermined period.   A communication device that outputs data input from an external device to an external network,
  A determination unit for determining whether or not congestion occurs in the external network;
  A throughput detector that detects an input throughput of data input from the external device;
  A communication device comprising: a control unit that controls a throughput for outputting data to an external network based on an input throughput immediately after the determination when the determination unit determines that the congestion occurs.

Images