JP5249966B2 - Communication control method and communication system for improving throughput, and program therefor - Google Patents

Description

  The present invention relates to a communication quality control technique in a packet communication network such as the Internet or an intranet, and is particularly suitable for improving the throughput by optimizing the transmission rate of a data communication protocol such as TCP (Transmission Control Protocol). The present invention relates to a TCP parameter setting technique.

  In a packet communication network represented by the Internet, since a plurality of users share the network, the packet transfer quality depends on the use situation of the network.

  However, in recent years, network technology such as DiffServ (Differentiated Services) can guarantee packet transfer quality in units of flows (packets that have the same source and destination address information and port number). It is possible.

  As for the quality assurance (QoS: Quality of Services) technology using DiffServ, for example, there are those described in Patent Document 1 and Patent Document 2.

  In these conventional technologies, the average rate and burst size are specified for the flow for quality assurance in the network, and the traffic inflow is restricted at the edge router (device installed at the entrance of the network) for traffic exceeding the specification. Perform policing (excess packet discard). Commercial products are widely used that implement a token bucket policer for policing (policing or performing policing using a token bucket) (see FIG. 1 described later).

  In communication in such an environment, Patent Document 1 and Patent are applied to the flow for performing UDP (User Datagram Protocol) communication with a substantially constant transmission rate such as video distribution and voice call by appropriate control on the transmission side. Transfer with no packet loss becomes possible by the quality assurance technique of Document 2.

  However, in communications with congestion control (window control) functions such as TCP that gradually increase the transmission rate and decrease the transmission rate at a certain rate when packet loss occurs, packet loss has a large effect on throughput. .

  For example, in Non-Patent Document 1 (described in the third paragraph in “4.1 Simulation results”) and the like, in a network that performs quality assurance as described in Patent Documents 1 and 2, the TCP throughput is a specified rate due to policing. The problem is that it will drop significantly.

  As a technique for solving this problem, there is a shaping technique for smoothing the transmission rate of packets as described in Patent Document 3, for example. In this shaping technique, traffic transmitted from a terminal is passed through a shaping device, and the traffic transmission rate is smoothed within the device so as to be within the network traffic specification.

  However, since the shaping device needs to be installed between the network and the transmission terminal, each user who subscribes to the service needs to have the shaping device. However, since a shaping apparatus is expensive, the use of a quality assurance network based on the shaping apparatus is not practical from the viewpoint of cost.

  Further, for example, Non-Patent Document 2 proposes a new policing mechanism for avoiding TCP performance degradation that occurs when TCP and policing are combined. This is an approach to improve the policing mechanism itself, so that it is not clearly shown how it will be performed, so improving TCP throughput in an environment using existing token bucket policers widely implemented in commercial products It cannot be a means to achieve this. Moreover, since implementation is not widespread as a commercial product, it is difficult to apply to an actual network.

JP 2004-320489 A JP 2005-073106 A Japanese Patent Laid-Open No. 06-268671

P.P.Mishra, “Effect of Leaky Bucket policing on TCP over ATM performance,” Porc. Of IEEE International Conference on Communications (ICC'96), vol.3, pp.1700-1706, 1996. van Haalen, R., Malhotra, R., “Improving TCP performance with bufferless token bucket policing: A TCP friendly policer,” Local & Metropolitan Area Networks, 2007. LANMAN 2007. 15th IEEE Workshop, pp.72-77, 2007.

  The problem to be solved by the present invention is that, in a quality assurance network using a conventional technology, for example, a network technology such as DiffServ, traffic exceeding a secured rate is discarded by policing of an edge router. In a communication protocol that lowers the transmission rate each time a packet loss occurs, in order to prevent a decrease in throughput due to policing, it is conventionally necessary to use a shaping device or replace policing with a new mechanism, but use a shaping device. To this end, it is necessary to provide a shaping device for service subscribers, which is difficult in terms of cost. In addition, in order to introduce a new policing mechanism, a mechanism that is not implemented in commercial products is installed, Carriers need to rebuild the whole It is that a big cost burden to take.

  Accordingly, an object of the present invention is to solve these problems of the prior art and optimally set the parameters of TCP communication in a network environment using a token bucket policer widely used in commercial products as a policing function. Therefore, it is to provide a communication control method and communication system capable of improving the throughput, and a program therefor.

The present invention employs the following configuration in order to achieve the above object.
a) In the communication control method according to the present invention, a reception terminal transmits a reception confirmation packet to the transmission terminal when receiving a data packet from the transmission terminal via a network that performs communication by packet exchange. A communication control method in a communication system that transmits a plurality of packets at a time each time a reception confirmation packet from a receiving terminal arrives, and increases the number of packets transmitted at a time without a reception confirmation every time a reception confirmation packet arrives. Thus, in an environment where policing for restricting traffic inflow is applied in the network, a packet that determines a packet size that does not cause a TCP retransmission timeout using a round-trip delay time with a communication partner and a set value of a policer Size determination procedure and packet size determined by the packet size determination procedure It is characterized by having a communication start procedure to start applying TCP communication.

b) The reception terminal transmits a reception confirmation packet to the transmission terminal when receiving the data packet from the transmission terminal via a network that performs communication by packet exchange, and the transmission terminal receives the reception confirmation packet from the reception terminal. A communication control method in a communication system for transmitting a plurality of packets at a time every time a message arrives and increasing the number of packets transmitted at a time without receiving a confirmation every time a reception confirmation packet arrives. In an environment where policing that restricts is applied, a packet size changing unit that dynamically changes a packet size after the start of communication, a packet size determination procedure that determines a packet size that can improve throughput, and Apply the packet size determined by the packet size determination procedure. It is characterized by having a communication start procedure for starting the TCP communication.

c) The communication control method further includes a throughput measurement procedure for measuring a throughput at a set packet size at regular intervals, and the packet size change procedure allows the packet size to be allowed in the communication environment at regular intervals. The packet size determination procedure is a procedure for determining a packet size that provides the maximum throughput from the throughputs measured at regular intervals by the throughput measurement procedure. It is a feature.

d) In the communication system according to the present invention, the reception terminal transmits a reception confirmation packet to the transmission terminal when receiving the data packet from the transmission terminal via the network that performs communication by packet exchange. A communication system that transmits a plurality of packets at a time each time a reception confirmation packet from a terminal arrives, and increases the number of packets transmitted at a time without a reception confirmation every time a reception confirmation packet arrives. In an environment where policing for restricting traffic inflow is applied, packet size determination means for determining a packet size at which TCP retransmission timeout does not occur using a round-trip delay time with a communication partner and a set value of a policer; TCP communication using the packet size determined by the packet size determining means It is characterized by having a communication start means for starting.

e) The reception terminal transmits a reception confirmation packet to the transmission terminal when receiving a data packet from the transmission terminal via a network that performs communication by packet exchange, and the transmission terminal receives the reception confirmation packet from the reception terminal. Policing that restricts traffic inflow in a network is applied in a communication system that increases the number of packets that are transmitted at one time without receiving confirmation each time a reception confirmation packet arrives. The packet size changing means for dynamically changing the packet size after the start of communication, the packet size determining means for determining the packet size capable of improving the throughput, and the packet size determining means TCP communication that applies the specified packet size It is characterized by having a communication start means.

f) The communication system further includes throughput measuring means for measuring the throughput at the set packet size at regular intervals, and the packet size changing means is allowed to allow the packet size at regular intervals in the communication environment. The packet size determining means is means for determining a packet size for obtaining the maximum throughput from the throughputs measured at regular intervals by the throughput measuring means. It is said.

g) Further, in the communication system, in TCP communication through the network to which the token bucket policer is applied, the communication system further includes a throughput measurement unit that measures the throughput at a set packet size at regular intervals, and the packet size changing unit includes: The packet size determining means is a means for halving the packet size every predetermined time, and the packet size determining means is configured to set the packet size changed by the packet size changing means when a retransmission timeout does not occur within a predetermined time. The throughput is measured by the throughput measuring means, and when the measured throughput is larger than the throughput of each packet size measured before, the packet size is set to the packet size that maximizes the throughput, and the throughput When the throughput is smaller than the previously measured packet size, the new packet size is set to the intermediate value between the packet size that gives the maximum value of the previously measured throughput and the packet size that gives the second largest throughput. It is a means for determining a packet size that maximizes the throughput by using the binary search method.

h) Further, in the communication system, in TCP communication through a network to which a token bucket policer is applied, the communication system further includes a throughput measurement unit that measures a throughput at a set packet size at regular intervals, and the packet size changing unit includes: Means for sequentially decreasing the packet size from the maximum value allowed in the communication environment at regular time intervals, and the packet size determining means is configured to measure the throughput measured at regular time intervals by the throughput measuring means triggered by a retransmission timeout. It is a means for determining a packet size from which the maximum throughput can be obtained.

i) Further, in the communication system, in TCP communication through a network to which a token bucket policer is applied, the communication system further includes a throughput measuring unit that measures a throughput at a set packet size at regular intervals, and the packet size changing unit includes: The packet size deciding means sets the packet size changed by the packet size changing means when the retransmission timeout does not occur within a predetermined time. In this case, the throughput is measured by the throughput measuring means, and when the measured throughput is larger than the throughput in each previously measured packet size, the packet size is set to the packet size that maximizes the throughput, If the throughput is smaller than the throughput at each packet size measured before, the packet size that has obtained the maximum value of the previously measured throughput and the intermediate value of the packet size that has issued the second largest throughput are set as new packets. It is a means for determining a packet size that maximizes throughput by using a binary search method for size.

j) A program according to the present invention is a program for causing a computer to function as each means in the communication system.

  ADVANTAGE OF THE INVENTION According to this invention, when performing TCP communication in the network environment which applied the token bucket policer, the communication control method and communication system which can aim at a throughput improvement by using small packet size, and the same A program can be provided.

FIG. 1 is a diagram illustrating a system configuration example in a token bucket policer application network. FIG. 2 is a diagram illustrating an example of a typical TCP congestion window size when the packet size is 1000 bytes and 250 bytes. FIG. 3 is a diagram illustrating a process flowchart according to the first embodiment. FIG. 4 is a diagram showing an example of the relationship between the packet size and TCP throughput (payload throughput) (in the case of a policing rate of 2 Mbps and a bucket size of 25 KB). FIG. 5 is a diagram illustrating a processing flowchart according to the second embodiment. FIG. 6 is a diagram illustrating an example of the relationship between the packet size and the TCP throughput in the second embodiment. FIG. 7 is a diagram illustrating a processing flowchart according to the third embodiment. FIG. 8 is a flowchart illustrating a process according to the fourth embodiment. FIG. 9 is a flowchart illustrating a process according to the fifth embodiment. FIG. 10 is a flowchart illustrating a process according to the sixth embodiment.

<Points of the present invention>
In TCP communication in a policing environment, the upper limit of the number of packets that can be retransmitted at the time of packet discard is determined by the size of the congestion window. Therefore, by performing communication with a small packet size, the congestion window size is increased and TCP retransmission is performed. Throughput is improved by suppressing the occurrence of timeout.

<Significance of the present invention>
In TCP communication, burst traffic is transmitted to perform an operation of increasing the transmission rate until the transmission side detects an increase in packet loss or delay by autonomous window control.

  As a result, the rate is temporarily exceeded, and the packet is discarded by the policer. Since TCP window control operates when an ACK packet arrives, detection of packet discard is also detected when an ACK packet arrives.

  On the other hand, since there is a time difference between packet discard and ACK arrival due to a transfer delay, the congestion window remains large even immediately after a policer discards a packet, and packets are sent one after another even if there are no surplus tokens in the policer. . The result is burst discard.

  When a packet is discarded in this way, TCP tries to retransmit the discarded packet, but the number of retransmission attempts depends on the size of the congestion window at the timing when the packet discard occurs, and the large window size. If the window size is small, the number of re-transmission attempts of the discard packet decreases. If the number of retransmission attempts is smaller than the number of discarded packets, packet retransmission fails and a retransmission timeout occurs.

  In many implementations, the congestion window size of TCP increases on the basis of the number of transmitted / received packets, not the amount of transmitted / received data. Therefore, even if the same amount of data is transmitted, congestion is reduced by reducing the packet size. The window size can be increased.

  FIG. 2 is a diagram illustrating an example of a typical TCP congestion window size when the packet size is 1000 bytes and 250 bytes. Although both have a transmission amount of 7000 bytes, the congestion window size is greatly different, and the window size in terms of the number of packets is larger when the packet size is 250 bytes. For this reason, the number of retransmittable packets at the time of burst discard also increases.

  In the present invention, in TCP communication under policing application, by reducing the packet size, the congestion window size is increased when the policer discards, and throughput is improved by increasing the number of retransmission attempts of discarded packets. Plan.

  Even if a retransmission timeout occurs when the number of re-transmission attempts of a discarded packet is small, it is possible to prevent a retransmission timeout from occurring by increasing the number of attempts.

  On the other hand, by reducing the packet size, the overhead of the IP header and the TCP header increases, and the payload throughput decreases. Therefore, it is necessary to select an appropriate packet size that maximizes the payload throughput and apply it to TCP communication.

Hereinafter, specific embodiments of the present invention will be described with reference to the drawings.
<Example 1>
FIG. 1 is a diagram illustrating a configuration example of a communication system in a network to which a token bucket policer is applied.
In the figure, 10 is a transmitting terminal, 20 and 40 are access lines, 30 is a DiffServ network, 31 and 32 are edge routers, 311 is a token bucket policer, 33 is a policer controller, and 50 is a receiving terminal.

  A dotted line between the transmission terminal 10 and the policer controller 33 represents a signaling path based on a logical connection rather than a direct connection.

  In principle, the processing is performed by the transmission terminal 10, the policer controller 33, and the token bucket policer 311 before the start of communication, but after the start of communication, the processing is performed only by the transmission terminal.

Next, an outline of processing in the present embodiment will be described.
The transmission terminal 10 is connected to the DiffServ network edge router 31 through the access line 20. The edge router 31 performs traffic policing using a token bucket model that defines an average rate and a burst amount.

  A data packet sent from the sending terminal 10 is transferred from the edge router 31 via the DiffServ network 30 to the receiving terminal 50 via the access line 40 via the edge router 32 connected to the communication partner.

  The transmitting terminal 10 and the receiving terminal 50 notify the DiffServ network 30 of the flow transmission rate before or during the communication start, and request the packet transfer quality to be ensured.

  The edge router 31 to which the transmitting terminal 10 is connected monitors the transfer amount of the target flow when this request is received, and performs policing so that packets exceeding the allowable amount are discarded.

  In the monitoring of the packet transfer amount in the edge router 31, the transmission amount per unit time is monitored, or monitoring by a token bucket meter is performed. The “flow” refers to a packet group having the same source and destination address information and the same port number.

  The transmission terminal 10 determines the packet size and the assumed throughput obtained in advance from the round trip time (RTT) with the reception terminal 50 and the information of the policing rate and bucket size, which are the setting parameters of the token bucket policer 311. By recording the relationship and setting an arbitrary packet size, a TCP flow is established with the receiving terminal 50 and communication is started.

Next, the process in Example 1 is demonstrated in detail using a flowchart.
FIG. 3 is a diagram illustrating a process flowchart according to the first embodiment.

  First, the rate is notified from the transmission terminal 10 to the policer controller 33 of the network (DiffServ network) 30 (step S10). The policer controller 33 of the network sets policer parameters in the token bucket policer 311 (step S11).

  The transmission terminal 10 knows the policer parameters set in the token bucket policer 311 by interacting with the policer controller 33, or searches for policer parameters to be used from preset information, and the policer parameters The transmission terminal 10 determines the transmission packet size (step S12), starts communication with the determined transmission packet size (step S13), and holds this until the communication ends (step S14).

  FIG. 4 is a diagram showing an example of the relationship between the packet size and TCP throughput (payload throughput) (in the case of a policing rate of 2 Mbps and a bucket size of 25 KB).

  The relationship between the packet size and the assumed throughput is obtained in advance from the policing rate, bucket size (policing rate 2 Mbps, bucket size 25 KB in the example of FIG. 4), and RTT, for example, and a packet size of 550 bytes that maximizes the throughput is specified. It is possible to do.

  In addition to maximizing the throughput, a packet size that realizes fairness between flows accepted by the network may be selected, and the setting criteria are not limited to these.

<Example 2>
In the environment of the first embodiment, the transmission terminal 10 establishes a TCP session with the reception terminal 50 as usual. At this time, as an initial value of the packet size, if communication is performed on the Ethernet (registered trademark), it starts with 1500 bytes which is an MTU (Max Transfer Unit). The present embodiment is characterized in that the packet size used for the communication is changed after the communication is started.

Next, the processing in the second embodiment will be described in more detail using a flowchart.
FIG. 5 is a diagram illustrating a processing flowchart according to the second embodiment.

  The transmission terminal 10 notifies the rate to the policer controller 33 of the network (DiffServ network) 30 (step S20). The policer controller 33 of the network sets policer parameters in the token bucket policer 311 (step S21). The transmission terminal 10 sets MTU as the packet size (step S22) and starts communication (step S23).

  After the start of communication, the transmission terminal 10 continues the communication while resetting (changing / maintaining) the packet size at an arbitrary timing during the communication, and this is repeated until the communication ends (steps S24 and S25).

  FIG. 6 is a diagram illustrating an example of the relationship between the packet size and the TCP throughput in the second embodiment.

  As shown in the figure, communication is performed with a 1500-byte packet size at the start of communication, but the packet size is changed at an arbitrary timing. For example, as shown in the figure, the packet size can be 800 bytes after 10 seconds from the start of communication, the packet size can be 200 bytes after 30 seconds from the start of communication, and the packet size can be 500 bytes after 50 seconds from the start of communication.

<Example 3>
In the environment of the first embodiment, the transmission terminal 10 establishes a TCP session with the reception terminal 50 as usual. At this time, as an initial value of the packet size, if communication is performed on the Ethernet (registered trademark), it starts with 1500 bytes which is an MTU (Max Transfer Unit).

  After communication is started, a value obtained by arbitrarily reducing the packet size at a fixed time interval T is set, and the throughput when the packet size is used is measured and stored.

  The packet size is changed from the maximum allowable value to a certain minimum value, and one of the packet sizes that maximizes the throughput is selected, and communication is continued using the packet size in subsequent communication.

Next, the processing in the third embodiment will be described in more detail using a flowchart.
FIG. 7 is a diagram illustrating a processing flowchart according to the third embodiment.

  First, the rate is notified from the transmission terminal 10 to the policer controller 33 of the network (DiffServ network) 30 (step S30). The policer controller 33 of the network sets policer parameters in the token bucket policer 311 (step S31). The transmitting terminal 10 sets MTU as the packet size (step S32) and starts communication (step S23).

  After the start of communication, the transmitting terminal 10 changes the packet size at regular time intervals T and measures and stores the throughput (steps S34 to S36). The transmitting terminal 10 changes / sets the optimum packet size (step S37), and continues communication with the changed / set packet size (step S38).

  In throughput measurement, payload throughput is measured instead of IP level throughput. Further, the minimum value of the packet size is equal to or larger than the value obtained by adding the IP packet header size and the TCP packet header size.

  The packet size value to be decreased may be the same value every time or a different value.

<Example 4>
In the environment of the first embodiment, the transmission terminal 10 establishes a TCP session with the reception terminal 50 as usual. At this time, as an initial value of the packet size, if communication is performed on the Ethernet (registered trademark), it starts with 1500 bytes which is an MTU (Max Transfer Unit).

  After the start of communication, a value obtained by halving the current packet size at an interval of a fixed time T is set, and the throughput during that period is measured and stored. This is repeated in order until a retransmission timeout does not occur within a predetermined time (the predetermined time here is generally different from the predetermined time T), and the throughput at each packet size is stored.

  Next, an optimal packet size is acquired from the stored throughput for each packet size using a binary search method, and communication is performed using the acquired optimal packet size.

Next, the processing in the fourth embodiment will be described in more detail using a flowchart.
FIG. 8 is a flowchart illustrating a process according to the fourth embodiment.

  First, the transmission terminal 10 notifies the rate to the policer controller 33 of the network (DiffServ network) 30 (step S40). The policer controller 33 of the network sets policer parameters in the token bucket policer 311 (step S41). The transmission terminal 10 sets MTU as the packet size (step S42) and starts communication (step S43).

  After the start of communication, the transmission terminal 10 measures and records the throughput every time the fixed time T has elapsed (step S44: Y) (step S45). If a retransmission timeout occurs within a predetermined time (step S46: Y), the current packet size is changed (in this embodiment, the packet size is halved), and the process returns to step S44 (step S47).

  Here, the predetermined time is generally different from the fixed time T, and is longer than the fixed time T (allowing equality).

  If a retransmission timeout does not occur within a predetermined time (step S46: N), the transmitting terminal 10 acquires and sets an optimal packet size using a binary search method from the stored throughput for each packet size (step S46). S48), the subsequent communication is performed to the end with the acquired optimum packet size (step S49).

  Here, a method for obtaining an optimum packet size when a retransmission timeout does not occur within the predetermined time will be specifically described.

  When a retransmission timeout does not occur within the predetermined time, the throughput for the predetermined time is calculated and compared with the throughput of each packet size stored in the past.

  If the throughput when retransmission timeout does not occur is higher than the past throughput, communication may be continued using the packet size in subsequent communications, or a retransmission timeout occurs when the packet size is increased. The maximum packet size in the range not to be searched may be searched, and communication using the packet size may be continued.

  If the throughput when retransmission timeout does not occur is lower than the past throughput, the packet size that maximizes the throughput by the binary search method using the packet size that records the top two of the previously stored throughputs To decide.

  Refer to steps S70 to S74 of the processing flowchart (FIG. 10) of the sixth embodiment described later for a specific processing example for determining the packet size that maximizes the throughput using the binary search method.

  Further, as the censoring condition of the above-described binary search method, a packet size difference is less than a certain value or a throughput difference is within a certain value, but is not limited thereto.

<Example 5>
In the environment of the first embodiment, the transmission terminal 10 establishes a TCP session with the reception terminal 50 as usual. At this time, as an initial value of the packet size, if communication is performed on the Ethernet (registered trademark), it starts with 1500 bytes which is an MTU (Max Transfer Unit).

  After the start of communication, with the occurrence of a retransmission timeout, set a value that is arbitrarily reduced from the current packet size to the communication until the next retransmission timeout occurs, and measure the throughput during that time, Remember. One packet size that maximizes the throughput is selected from these, and communication is continued using the packet size in subsequent communication.

  Note that this embodiment is characterized in that the packet size is switched in response to a retransmission timeout, so that communication using the same packet size is continuously used while the retransmission timeout occurs a plurality of times to measure the throughput. May be performed.

  The packet size value to be decreased may be the same value every time or a different value.

Next, the processing in the fifth embodiment will be described in more detail using a flowchart.
FIG. 9 is a flowchart illustrating a process according to the fifth embodiment.

  First, the rate is notified from the transmission terminal 10 to the policer controller 33 of the network (DiffServ network) 30 (step S50). The policer controller 33 of the network sets policer parameters in the token bucket policer 311 (step S51). The transmission terminal 10 sets MTU as the packet size (step S52) and starts communication (step S53).

  After starting the communication, the transmission terminal 10 sets a value that is arbitrarily changed (decreased) from the current packet size to the communication until the next retransmission timeout occurs when the retransmission timeout occurs, The throughput during that time is measured and stored (steps S54 to S56).

  One packet size that maximizes the throughput is selected from the stored throughputs for each packet size, and communication is performed using the packet size. These processes are repeated until a retransmission timeout does not occur within a predetermined time (the predetermined time here is generally different from the predetermined time T) (step S54: N). The processes in steps S54 to S56 are repeated until the communication is completed (step S57).

<Example 6>
In the environment of the first embodiment, the transmission terminal 10 establishes a TCP session with the reception terminal 50 as usual. At this time, as an initial value of the packet size, if communication is performed on the Ethernet (registered trademark), it starts with 1500 bytes which is an MTU (Max Transfer Unit).

  After the start of communication, when a retransmission timeout occurs, set a value that halves the current packet size for communication until the next retransmission timeout occurs, and measure and store the throughput during that time . This is repeated in order until a retransmission timeout does not occur within a predetermined time, and the throughput at each packet size is stored.

  When a retransmission timeout does not occur within the predetermined time, the throughput for the predetermined time is calculated and compared with the throughput of each packet size stored in the past. If the throughput when retransmission timeout does not occur is higher than the past throughput, communication is continued using the packet size in subsequent communication.

  If the throughput when retransmission timeout does not occur is lower than the past throughput, the packet size that maximizes the throughput by the binary search method using the packet size that records the top two of the previously stored throughputs To decide.

  Note that this embodiment is characterized in that the packet size is switched in response to a retransmission timeout, so that communication using the same packet size is continuously used while the retransmission timeout occurs a plurality of times to measure the throughput. May be performed.

Next, the processing in the sixth embodiment will be described in more detail using a flowchart.
FIG. 10 is a flowchart illustrating a process according to the sixth embodiment.

  First, the transmission terminal 10 notifies the network policer controller 33 of the rate (step S60). The policer controller 33 of the network sets policer parameters in the token bucket policer 311 (step S61). The transmitting terminal 10 sets MTU as the packet size (step S62) and starts communication (step S63).

  After the start of communication, the transmission terminal 10 sets a size that halves the current packet size for communication until the next retransmission timeout occurs when the retransmission timeout occurs, and measures the throughput during that time. Are stored (steps S64 to S66).

  If a retransmission timeout does not occur within the predetermined time (step S64: N), the throughput for the predetermined time is calculated (step S67) and compared with the throughput of each packet size stored in the past (step S68).

  If the throughput when the retransmission timeout does not occur is higher than the throughput recorded in the past (step S68: Y), communication is continued using the packet size in the subsequent communication (step S69).

  If the throughput when the retransmission timeout does not occur is lower than the past throughput (step S68: N), the throughput is determined by the binary search method using the packet size in which the top two stored throughputs are recorded. The packet size that maximizes is determined (steps S70 to S74).

  The details of the binary search method in steps S70 to S74 will be described. When the throughput when the retransmission timeout does not occur is lower than the past throughput (N in step S68), the highest value Th1 of the throughput and the second A packet size of P = (P1 + P2) / 2 is calculated for the order value Th2 and the packet sizes P1 and P2 giving it, and the throughput at this packet size is measured and recorded (step S71).

  Next, if the packet size change termination condition (described later) is not met (step S72: N), the packet size is changed from P ′ to (P1 + P) / 2 (step S73), the throughput is measured, and the throughput is measured. After updating the highest value Th1 and the second value Th2 and the packet sizes P1 and P2 that give them (step S74), the process returns to step S72.

  If the packet size change termination condition is met (step S72: Y), communication is continued according to the packet size at that time (step S75).

  Note that as the packet size change termination condition of the binary search method described above, a packet size difference is less than a certain value or a throughput difference is within a certain value, but is not limited thereto.

  Note that the processing performed by the communication device as the transmission terminal described with reference to the flowcharts of FIGS. 3, 5, 7, 8, 9, and 10 is a CPU or memory built in the computer that constitutes the transmission terminal. This is realized by executing a program corresponding to each process using hardware such as a register. A program corresponding to each of these processes can be distributed to the market via a recording medium such as FD, CD-ROM, or DVD, or a network such as the Internet.

10: Sending terminal 20: Access line 30: DiffServ network 31, 32: Edge router 311: Token bucket policer 33: Policer controller 40: Access line 50: Receiving terminal

Claims (8)

When a receiving terminal receives a data packet from a transmitting terminal via a network that performs communication by packet exchange, the receiving terminal transmits a receiving confirmation packet to the transmitting terminal, and the transmitting terminal receives the receiving confirmation packet from the receiving terminal each time A communication control method in a communication system that transmits a plurality of packets at a time and increases the number of packets to be transmitted at a time without reception confirmation every time a reception confirmation packet arrives.
In an environment where policing for restricting traffic inflow is applied in the network,
A packet size determination procedure for determining a packet size at which TCP retransmission timeout does not occur, using a round-trip delay time with a communication partner and a set value of a policer;
A communication start procedure for starting TCP communication to which the packet size determined by the packet size determination procedure is applied. When a receiving terminal receives a data packet from a transmitting terminal via a network that performs communication by packet exchange, the receiving terminal transmits a receiving confirmation packet to the transmitting terminal, and the transmitting terminal receives the receiving confirmation packet from the receiving terminal each time A communication control method in a communication system that transmits a plurality of packets at a time and increases the number of packets to be transmitted at a time without reception confirmation every time a reception confirmation packet arrives.
In an environment where policing for restricting traffic inflow is applied in the network,
A packet size change procedure for dynamically changing the packet size after communication starts;
A packet size determination procedure for determining a packet size capable of improving throughput;
A communication start procedure for starting TCP communication using the packet size determined by the packet size determination procedure ;
A throughput measurement procedure for measuring the throughput at a set packet size at regular intervals;
I have a,
The packet size changing procedure is a procedure for sequentially decreasing the packet size from a maximum value allowed in the communication environment at regular intervals,
The packet size determination procedure, a communication control method characterized by procedure step der to determine the packet size at which the maximum throughput can be obtained from the throughput measured for each predetermined time by the throughput measurement procedure. When a receiving terminal receives a data packet from a transmitting terminal via a network that performs communication by packet exchange, the receiving terminal transmits a receiving confirmation packet to the transmitting terminal, and the transmitting terminal receives the receiving confirmation packet from the receiving terminal each time A communication system that transmits a plurality of packets at a time and increases the number of packets to be transmitted at a time without receipt confirmation every time a receipt confirmation packet arrives.
In an environment where policing for restricting traffic inflow is applied in the network,
A packet size determining means for determining a packet size that does not cause a TCP retransmission timeout, using a round-trip delay time with a communication partner and a set value of a policer;
Communication start means for starting TCP communication to which the packet size determined by the packet size determination means is applied. When a receiving terminal receives a data packet from a transmitting terminal via a network that performs communication by packet exchange, the receiving terminal transmits a receiving confirmation packet to the transmitting terminal, and the transmitting terminal receives the receiving confirmation packet from the receiving terminal each time In a communication system that transmits a plurality of packets at a time and increases the number of packets to be transmitted at a time without receiving confirmation every time a reception confirmation packet arrives,
In an environment where policing for restricting traffic inflow is applied in the network,
A packet size changing means for dynamically changing the packet size after starting communication;
A packet size determining means for determining a packet size capable of improving the throughput;
Communication start means for starting TCP communication using the packet size determined by the packet size determination means;
A throughput measuring means for measuring the throughput in the set packet size at regular intervals ;
I have a,
The packet size changing means is means for sequentially reducing the packet size from a maximum value allowed in the communication environment at regular intervals,
The packet size determination means, a communication system, wherein the Ru means der to determine the packet size at which the maximum throughput can be obtained from the measured throughput for each predetermined time by the throughput measurement unit. When a receiving terminal receives a data packet from a transmitting terminal via a network that performs communication by packet exchange, the receiving terminal transmits a receiving confirmation packet to the transmitting terminal, and the transmitting terminal receives the receiving confirmation packet from the receiving terminal each time In a communication system that transmits a plurality of packets at a time and increases the number of packets to be transmitted at a time without receiving confirmation every time a reception confirmation packet arrives,
In an environment where policing for restricting traffic inflow is applied in the network,
A packet size changing means for dynamically changing the packet size after starting communication;
A packet size determining means for determining a packet size capable of improving the throughput;
Communication start means for starting TCP communication using the packet size determined by the packet size determination means ;
In TCP communication through a network to which a token bucket policer is applied,
A throughput measuring means for measuring the throughput in the set packet size at regular intervals;
I have a,
The packet size changing means is means for halving the packet size every certain time,
The packet size determining means performs the measurement of the throughput when the packet size changed by the packet size changing means is set by the throughput measuring means when a retransmission timeout does not occur within a predetermined time. When the throughput is larger than the throughput at each packet size measured before, the packet size is set to the packet size that maximizes the throughput, and when the throughput is smaller than the throughput at each packet size measured before, Packet size that maximizes throughput by using a binary search method with a new packet size that is the middle of the packet size that yielded the maximum value of the previously measured throughput and the second largest throughput. Communication system, wherein the determining means der to Apply predicates. When a receiving terminal receives a data packet from a transmitting terminal via a network that performs communication by packet exchange, the receiving terminal transmits a receiving confirmation packet to the transmitting terminal, and the transmitting terminal receives the receiving confirmation packet from the receiving terminal each time In a communication system that transmits a plurality of packets at a time and increases the number of packets to be transmitted at a time without receiving confirmation every time a reception confirmation packet arrives,
In an environment where policing for restricting traffic inflow is applied in the network,
A packet size changing means for dynamically changing the packet size after starting communication;
A packet size determining means for determining a packet size capable of improving the throughput;
Communication start means for starting TCP communication using the packet size determined by the packet size determination means ;
In TCP communication through a network to which a token bucket policer is applied,
A throughput measuring means for measuring the throughput in the set packet size at regular intervals ;
I have a,
The packet size changing means is means for sequentially reducing the packet size from a maximum value allowed in the communication environment at regular intervals,
The packet size determination means, in response to a retransmission timeout, and wherein the Ru means der to determine the packet size at which the maximum throughput can be obtained from the measured throughput for each predetermined time by the throughput measurement unit communication system. When a receiving terminal receives a data packet from a transmitting terminal via a network that performs communication by packet exchange, the receiving terminal transmits a receiving confirmation packet to the transmitting terminal, and the transmitting terminal receives the receiving confirmation packet from the receiving terminal each time In a communication system that transmits a plurality of packets at a time and increases the number of packets to be transmitted at a time without receiving confirmation every time a reception confirmation packet arrives,
In an environment where policing for restricting traffic inflow is applied in the network,
A packet size changing means for dynamically changing the packet size after starting communication;
A packet size determining means for determining a packet size capable of improving the throughput;
Communication start means for starting TCP communication using the packet size determined by the packet size determination means ;
In TCP communication through a network to which a token bucket policer is applied,
A throughput measuring means for measuring the throughput in the set packet size at regular intervals;
I have a,
The packet size changing means is means for halving the packet size triggered by a retransmission timeout,
The packet size determining means performs the measurement of the throughput when the packet size changed by the packet size changing means is set by the throughput measuring means when a retransmission timeout does not occur within a predetermined time. When the throughput is larger than the throughput at each packet size measured before, the packet size is set to the packet size that maximizes the throughput, and when the throughput is smaller than the throughput at each packet size measured before, Packet size that maximizes throughput by using a binary search method with a new packet size that is the middle of the packet size that yielded the maximum value of the previously measured throughput and the second largest throughput. Communication system, wherein the determining means der to Apply predicates. Program for causing a computer to function as each unit in the communication system according to any one of claims 3 7.

Images