WO2011048740A1 - Data transmission system, transmission rate controlling method, receiving terminal and transmitting terminal - Google Patents

Abstract

Provided is a data transmission system (100) wherein: a transmitting terminal (2) and a receiving terminal (1) are connected to each other by a network (3); the transmitting terminal (2) transmits data packets to the receiving terminal (1) via the network (3); and the receiving terminal (1) transmits back a control packet to the transmitting terminal (2) upon receiving the data packets from the transmitting terminal (2). A network data measuring unit (11) of the receiving terminal (1) measures network data on the basis of which the congestion level of the network can be predicted. A control packet transmission rate determining means predicts, in each period having a predetermined length, the congestion level of the network by using the network data that has been measured over the period by the network data measuring unit (11), and determines, in accordance with the congestion level thus predicted, the transmission rate of a control packet to be transmitted from the receiving terminal (1). This configuration makes it possible to reliably follow the congestion level of the network.

Description

Data transmission system, transmission speed control method, receiving terminal, transmitting terminal

The present invention relates to a data transmission technique between terminals connected via a network.

In recent years, the scale of data centers that provide various services such as WWW (World Wide Web) search, data communication, content distribution, server hosting, and thin client (Thin Client) has been increasing. For example, a data center owned by a general Internet service provider whose core business is a WWW search service, a data communication service, and a content distribution service has thousands of data center components (servers, storage, routers, switches, etc.) Network equipment).

The data center needs to provide various services to users without interruption. To achieve this, the operation management server installed in the data center can constantly monitor the operating status of thousands of data center components, and can quickly detect failures and malfunctions in data center components. Operation management is performed. Specifically, for example, the operation management server collects operation status log files generated by the data center constituent devices via an operation management network in the data center. The operation management server analyzes the collected operation status log file to grasp the operation status of the data center constituent devices.

In an operation management network, transmission of operation status log files of various sizes generated by thousands of data center components occurs randomly, so communication traffic and network congestion (for example, “network bandwidth-communication traffic”). ") Changes in a short period. Even if a large amount of data packets are transmitted in a state where the network congestion is high (congestion state), transmission delay and packet loss frequently occur due to network congestion, and high-speed data transmission cannot be performed. On the other hand, in order to transmit data at high speed in a state where the network congestion is low (normal state), it is necessary to determine the amount of data packets that can be transmitted within a range where transmission delay and packet loss do not occur.

For this reason, in order to realize high-speed data transmission, it is required to transmit data packets at an appropriate speed following the ever-changing network congestion.

In Patent Document 1, transmission of an acknowledgment packet that is returned from a terminal that receives the data packet (hereinafter referred to as a receiving terminal) to a terminal that transmits the data packet (hereinafter referred to as a transmitting terminal) according to the reception speed of the data packet. A technique for setting a speed and a transmission speed of a data packet to be subsequently transmitted from the transmitting terminal to the receiving terminal is disclosed.

Specifically, when the receiving terminal receives the data packet, it measures the receiving speed and generates an acknowledgment packet including the measured receiving speed and returns it to the transmitting terminal. At this time, the receiving terminal sets the transmission rate of the acknowledgment packet according to the reception rate. The transmission terminal acquires the reception speed included in the acknowledgment packet from the reception terminal, and sets the subsequent data packet transmission speed according to the reception speed.

That is, this technique uses the reception speed of the data packet at the reception terminal as an index of the degree of network congestion, and accordingly, the transmission speed of the acknowledgment packet from the reception terminal and the transmission terminal performed after reception of the data packet. The data packet transmission speed from is adjusted.

Japanese Patent Laid-Open No. 2004-80070

However, the technique of Patent Document 1 sets the transmission rate of the acknowledgment packet currently transmitted by the receiving terminal and the transmission rate of the data packet currently transmitted by the transmitting terminal based on the past receiving rate of the data packet at the receiving terminal. is doing. Although the receiving speed of the data packet at the receiving terminal can be an indicator of the degree of network congestion when the data packet is received, when the receiving terminal transmits an acknowledgment packet, the transmitting terminal transmits subsequent data packets. It may not be an index of the degree of network congestion.

Suppose, for example, a network in which the network congestion level changes in a short cycle, such as the operation management network in the data center described above. In the network, the network congestion level at the time of receiving a data packet at the receiving terminal, the network congestion level when the receiving terminal receives the data packet and sets the transmission speed for the data packet, and the transmitting terminal The network congestion level when setting the transmission rate of the data packet after receiving the response confirmation packet may differ greatly. In such a case, the technique disclosed in Patent Document 1 cannot transmit data packets at an appropriate transmission rate that follows the current network congestion level.

The present invention has been made in view of the above circumstances, and provides a data packet transmission technique that can reliably follow the degree of network congestion.

One aspect of the present invention is a method for controlling a transmission rate in a data transmission system. In a data transmission system, a transmission terminal and a reception terminal are connected by a network. The transmitting terminal transmits a data packet to the receiving terminal via the network, and the receiving terminal returns an acknowledgment packet to the transmitting terminal in response to reception of the data packet from the transmitting terminal.
In the control method according to the aspect of the present invention, in the data transmission system as described above, the network data capable of predicting the congestion degree of the network is measured by the receiving terminal, and the data is transmitted every predetermined period. A network congestion degree is predicted using the measured network data, and a transmission rate of an acknowledgment packet transmitted from the receiving terminal is determined according to the predicted congestion degree.

Note that a method in which the method of the above aspect is replaced with a data transmission system, a reception terminal or a transmission terminal in the data transmission system, and a program that causes a computer to execute a part of the method are also effective as an aspect of the present invention. .

According to the technology of the present invention, data transmission can be performed at an appropriate speed that reliably follows the degree of network congestion.

It is a figure which shows the data transmission system concerning the 1st Embodiment of this invention. It is a flowchart which shows the flow of a process by the data transmission system shown in FIG. It is a figure which shows an example of the reception condition of the data packet in the receiving terminal side. It is a figure which shows an example of a network congestion degree model. It is a figure which shows the structural example of the control packet which the receiving terminal produced in the data transmission system shown in FIG. FIG. 3 is a diagram for explaining an example of a control packet used when determining a data packet transmission rate in the data transmission system shown in FIG. 1 (part 1); FIG. 3 is a diagram for explaining an example of a control packet used when determining a data packet transmission rate in the data transmission system shown in FIG. 1 (part 2); FIG. 3 is a diagram for explaining a method for determining a data packet transmission rate in the data transmission system shown in FIG. 1 (part 1); FIG. 3 is a diagram for explaining a technique for determining a data packet transmission rate in the data transmission system shown in FIG. 1 (part 2); It is a figure for demonstrating another method at the time of determining a data packet transmission rate in the data transmission system shown in FIG. It is a figure which shows the data transmission system concerning the 2nd Embodiment of this invention. It is a flowchart which shows the flow of a process by the data transmission system shown in FIG. It is a figure for demonstrating the data packet shaping part in the transmission terminal of the data transmission system shown in FIG. It is a figure which shows the structure of the control packet which the receiving terminal of the data transmission system shown in FIG. 1 produced.

For clarity of explanation, the following description and drawings are omitted and simplified as appropriate. Each element described in the drawings as a functional block for performing various processes can be configured by a CPU, a memory, and other circuits in terms of hardware, and a program loaded in the memory in terms of software. Etc. Therefore, it is understood by those skilled in the art that these functional blocks can be realized in various forms by hardware only, software only, or a combination thereof, and is not limited to any one.
<First Embodiment>

FIG. 1 shows a data transmission system 100 according to a first embodiment of the present invention. The data transmission system 100 includes a receiving terminal 1 and a transmitting terminal 2. The receiving terminal 1 and the transmitting terminal 2 are a computer or a workstation having a data communication function, a data processing function by program control, an information storage function, and the like, and are connected by a network 3 so as to be able to perform data communication with each other.

The receiving terminal 1 receives a data packet from the transmitting terminal 2 and returns a control packet to be described later to the transmitting terminal 2. The receiving terminal 1 includes a network data measurement unit 11, a control packet transmission rate determination unit 12, a network congestion level model storage unit 13, a control packet creation unit 14, and a control packet transmission rate control unit 15.

When the receiving terminal 1 receives a data packet from the transmitting terminal 2, the network data measuring unit 11 measures network data for predicting the future network congestion. These network data are data indicating, for example, the presence / absence of a lost packet, the reception time of the data packet, and the like. The network data measuring unit 11 measures the network data by confirming the sequence number of the received data packet.

The control packet creation unit 14 creates an acknowledgment packet to be sent back to the transmission terminal 2. In the following description, the acknowledgment packet is also referred to as a control packet.

The network congestion degree model storage unit 13 is a network congestion degree model indicating the relationship between the packet loss rate calculated from network data, the data packet reception interval jitter (dispersion value of the data packet reception interval), and the network congestion degree at a future time. Is remembered.

The control packet transmission rate determination unit 12 determines the transmission rate of the control packet. Specifically, first, the control packet transmission rate determination unit 12 calculates a packet loss rate, a data packet reception interval jitter, and the like using the network data measured by the network data measurement unit 11. Then, the control packet transmission rate determination unit 12 refers to the network congestion model stored in advance in the network congestion model storage unit 13 and determines the future time based on the calculation result of the packet loss rate and the data packet reception interval jitter. Predict network congestion. Thereafter, the control packet transmission rate determination unit 12 determines the optimal transmission rate of the control packet based on the prediction result of the network congestion degree at the future time. That is, in the present embodiment, the control packet transmission rate determining unit 12 includes a first congestion degree predicting unit (not shown) in the claims. Regarding the determination of the transmission rate of the control packet, the control packet transmission rate determination unit 12 performs, for example, as follows. Since the data packet transmission rate needs to be finely controlled in the congestion state, the control packet transmission rate determination unit 12 increases the control packet transmission rate. On the other hand, in the normal state, since there is a margin in the network bandwidth, the transmission efficiency is not greatly affected even if the data packet transmission rate is not strictly controlled. Therefore, the control packet transmission rate determination unit 12 reduces the control packet transmission rate. .

The control packet transmission rate control unit 15 transmits the control packet created by the control packet creation unit 14 to the transmission terminal 2 via the network 3 at the transmission rate decided by the control packet transmission rate decision unit 12.

The transmission terminal 2 transmits a data packet to the reception terminal 1, and includes a data packet transmission rate control unit 21, a loss packet retransmission unit 22, a data packet storage unit 23, a control packet analysis unit 24, and a data packet transmission rate determination unit. 25.

The data packet storage unit 23 stores data packets transmitted from the transmission terminal 2 to the reception terminal 1.

The control packet analysis unit 24 analyzes the control packet received from the receiving terminal 1 and extracts the sequence number and packet loss time of the lost packet. Note that the loss packet is a data packet that cannot be normally received by the receiving terminal 1 among all data packets transmitted from the transmitting terminal 2 to the receiving terminal 1. The packet loss time is an estimated reception time of a lost packet. Further, the control packet analysis unit 24 deletes all data packets transmitted before the first loss packet from the data packet storage unit 23 based on the analysis result.

The data packet transmission rate determining unit 25 determines the transmission rate of the data packet to be transmitted to the receiving terminal 1. The data packet transmission rate determination unit 25 determines not only the transmission rate of the data packet transmitted to the receiving terminal 1 for the first time, but also the transmission rate when resending the lost packet. Although details will be described later, the data packet transmission rate determination unit 25 predicts the congestion level of the network and determines the transmission rate of the data packet according to the predicted congestion level. That is, in the present embodiment, the data packet transmission rate determining unit 25 includes a second congestion degree predicting unit (not shown) in the claims.

The data packet transmission rate control unit 21 transmits data packets (including data packets corresponding to loss packets) to the receiving terminal 1 at the transmission rate determined by the data packet transmission rate determination unit 25.

The loss packet retransmission unit 22 refers to the analysis result by the control packet analysis unit 24 and erases the data packet that has been confirmed to have been normally transmitted to the receiving terminal 1 from the data packet storage unit 23. On the other hand, for the lost packet, the lost packet retransmission unit 22 reads the data packet corresponding to the sequence number of the lost packet from the control packet analysis unit 24 from the data packet storage unit 23 and outputs it to the data packet transmission rate control unit 21. And have them resend.

Next, the flow of processing by the data transmission system 100 of this embodiment will be described with reference to the flowchart of FIG.
The operation will be described on the assumption that the receiving terminal 1 is ready to receive data packets at the start of data transmission by the transmitting terminal 2. This assumption is established by, for example, a method in which a program responsible for data packet reception processing (step D1 in FIG. 2) is always started as a server program in the receiving terminal 1 so that the receiving terminal 1 can always receive data packets. To do.

The sending terminal 2 starts data transmission when, for example, a data packet generated from a log file is stored in the data packet storage unit 23 in a certain amount or more (step S1). The transmission terminal 2 reads the data packet from the data packet storage unit 23 and starts data packet transmission to the reception terminal 1 (step S11). Further, the transmitting terminal 2 starts receiving control packets in order to receive the control packet transmitted by the receiving terminal 1 (step S21).

The receiving terminal 1 measures the network data (step D2) while receiving the data packet transmitted from the transmitting terminal 2 over a certain period (step S12, D1). Next, the receiving terminal 1 calculates a packet loss rate, a data packet reception interval jitter, and the like from the measured network data, and refers to the network congestion model stored in the network congestion model storage unit 13 at a future time. Predict network congestion. Then, the receiving terminal 1 determines the control packet transmission rate based on the predicted network congestion at the future time (step D3). Thereafter, the control packet creation unit 14 creates a control packet that describes the sequence number of the lost packet, the packet loss time, and the like (step D4). The control packet transmission rate control unit 15 transmits the control packet created by the control packet creation unit 14 to the transmission terminal 2 in accordance with the control packet transmission rate determined in step D3 (step D5).

In the transmission terminal 2, the control packet analysis unit 24 analyzes the control packet received from the reception terminal 1, extracts the sequence number of the lost packet, the packet loss time, and the like (steps S22 and S23), and the control packet will be described later. The data packet corresponding to the sequence number before the sequence number described in the last received packet field is deleted from the data packet storage unit 23.

The lost packet retransmission unit 22 reads out a data packet corresponding to the sequence number of the lost packet notified from the control packet analysis unit 24 from the data packet storage unit 23, and outputs the data packet to the data packet transmission rate control unit 21 for retransmission ( Step S12). In parallel, the data packet transmission rate determination unit 25 determines the data packet transmission rate based on the sequence number of the lost packet extracted by the control packet analysis unit 24, the packet loss time, and the packet size of the lost packet (step S13). ).

The data packet transmission rate control unit 21 sequentially transmits all data packets including the loss packet to the receiving terminal 1 according to the data packet transmission rate determined in step S13, and all the data packets stored in the data packet storage unit 23 are normal. The process ends when the data is transmitted (steps S14 and S2).

Next, the data transmission system 100 will be described in more detail with specific examples.
The reception terminal 1 and the transmission terminal 2 are realized by a computer, a workstation, a dedicated terminal, or the like that includes an arithmetic processing unit such as a CPU (Central Processing Unit), an input / output interface, and a storage unit. The storage unit is, for example, a RAM (Random Access Memory), a ROM (Read Only Memory), an HDD (Hard Disk Drive), or the like. The receiving terminal 1 and the transmitting terminal 2 include a network communication unit that supports Ethernet (registered trademark), a wireless LAN, and the like for transmitting data packets and control packets.

It is desirable that the receiving terminal 1 and the transmitting terminal 2 use a UDP (User Datagram Protocol) transport protocol for data packet transmission. This is because the UDP transport protocol does not guarantee the arrival order of packets, retransmission of lost packets, etc., but can transmit data at high speed. In the data packet generated from the operation status log file of the data center component device assumed in the first embodiment, all data packets are transmitted without packet loss when the data transmission is completed (step S2 in FIG. 2). For example, changing the order of arrival of data packets in the transmission process is not a problem.

Meanwhile, it is desirable that the receiving terminal 1 and the transmitting terminal 2 use a TCP (Transmission Control Protocol) transport protocol for control packet transmission. This is because, in the first embodiment, it is desirable that the control packet transmitted by the receiving terminal 1 arrives at the transmitting terminal 2 without changing the arrival order. This is because the network data described in the control packet used by the transmitting terminal 2 to determine the data packet transmission rate is time-series data based on the packet loss time, and when the control packet arrival order is changed, This is because the data packet transmission rate cannot be determined.

When the data transmission is started, the transmission terminal 2 extracts the data packet stored in the data packet storage unit 23 and starts transmission to the reception terminal 1. In the data packet, at least a sequence number indicating the transmission order of the data packet is described. However, since the data packet needs to be retransmitted when the data packet is lost, the data packet is stored in the data packet storage unit even after the transmission terminal 2 extracts the data packet from the data packet storage unit 23. 23 remains stored.

In the receiving terminal 1, the network data measuring unit 11 confirms the sequence number of the data packet received from the transmitting terminal 2 and measures network data such as the occurrence / non-occurrence of a loss packet and the reception time of the data packet. Then, the control packet transmission rate determination unit 12 determines the control packet transmission rate using the network data measured by the network data measurement unit 11 over a certain period. The network data measurement period is set, for example, in units of time or the number of received data packets.

The determination of the control packet transmission rate by the control packet transmission rate determination unit 12 will be described in detail. The situation shown in FIG. 3 will be described as an example. In FIG. 3, the solid line frame indicates a data packet that can be normally received, and the dotted line frame indicates a data packet (loss packet) that cannot be normally received. The number in each frame indicates the sequence number of the data packet, and indicates the past time as the subscript of time T becomes smaller.

In the example shown in FIG. 3, data packets with sequence numbers 1, 3, 6, 7, and 8 (hereinafter referred to as data packets 1, 3, 6, 7, and 8) are time T1, T3, T6, T7, The data packets are normally received at T8, and the data packets 2, 4, and 5 are loss packets that cannot be normally received. Note that the time T8 is the current time, and the reception times of the data packets 2, 4, and 5 that are lost packets are unknown due to reception failure.

In the situation shown in FIG. 3, the network data measurement unit 11 uses the sequence number (2, 4, 5) of each lost packet and the reception time (T1, T3, T6, T7, T8) are acquired.

The control packet transmission rate determination unit 12 calculates a packet loss rate and a data packet reception interval jitter from the network data acquired by the network data measurement unit 11. In the present embodiment, the control packet transmission rate determination unit 12 calculates the ratio of the number of lost packets to the total number of data packets as the packet loss rate. Further, according to the equation (1), the control packet transmission rate determination unit 12 calculates the dispersion value V of the reception interval of the data packets that can be normally received as the data packet reception interval jitter.

In the example shown in FIG. 3, since 3 lost packets occur among 8 data packets, the packet loss rate is 37.5%. In addition, the dispersion value V of the reception intervals T3-T1, T6-T3, T7-T6, and T8-T7 of the five data packets 1, 3, 6, 7, and 8 that can be normally received is the data packet reception interval jitter. Is calculated as

The control packet transmission rate determination unit 12 refers to the network congestion level model storage unit 13 in order to determine the control packet transmission rate. The network congestion level model storage unit 13 stores a state transition model of the network congestion level according to the packet loss rate and the data packet reception interval jitter. FIG. 4 shows an example of a network congestion model.

In the model shown in FIG. 4, the network congestion level is roughly divided into “normal state” and “congestion state”. The initial state at the start of data transmission is the normal state. The normal state is a state where the available network bandwidth is sufficiently left, and the congestion state is a state where the available network bandwidth is not left.

As shown in FIG. 4, in the model, in the normal state, the packet loss rate a [%] is lower than the packet loss rate threshold value ath [%], and the data packet reception interval jitter b is equal to or greater than the data packet reception interval jitter threshold value bth. If the condition (condition 1 in FIG. 4) is satisfied, it is determined that network congestion has occurred and the state has changed to a congestion state. When the above condition 1 is not satisfied, it is determined that the normal state continues.

On the other hand, in the congestion state, the packet loss rate a [%] is not less than the packet loss threshold ath [%] and the data packet reception interval jitter b is shorter than the data packet reception interval jitter threshold bth (condition 2 in FIG. 4). When is established, it is determined that the network congestion has been resolved and the transition to the normal state has occurred. When the above condition 2 is not satisfied, it is determined that the congestion state continues.

The control packet transmission rate determination unit 12 changes the transmission timing of the control packet to a short cycle when the network congestion degree transitions from the normal state to the congestion state because the condition 1 is satisfied in the normal state. That is, the control packet transmission rate determination unit 12 increases the transmission rate of the control packet.

On the other hand, when the condition 2 is established in the congestion state, and the network congestion level changes from the congestion state to the normal state, the control packet transmission rate determination unit 12 changes the transmission timing of the control packet to a long cycle. That is, the control packet transmission rate determination unit 12 reduces the control packet transmission rate.

It should be noted that the control packet transmission rate at high speed and the control packet transmission rate at low speed may be set in advance for the receiving terminal 1.

Next, the creation of the control packet by the control packet creation unit 14 will be described in detail.
FIG. 5 shows an example of the control packet. The control packet includes at least a final received packet field and a lost packet list field. In the final received packet field, the control packet creation unit 14 describes the maximum value of the sequence number of the data packet that can be normally received by the receiving terminal and the reception time. In the loss packet list field, the control packet creation unit 14 describes the sequence numbers and packet loss times of all data packets (loss packets) that the receiving terminal 1 has failed to receive. Since the reception time of the lost packet is unknown, the control packet creation unit 14 estimates the reception time of each lost packet from the time of the data packet that has been successfully received before and after the lost packet, and sets the loss as the packet loss time. Describe in the packet list field.

For example, in the case of the control packet corresponding to the situation shown in FIG. 3, the sequence number 1 and the reception time T1 of the data packet T1 are described in the final reception packet field, and the data packets 2, 4 are stored in the loss packet list field. 5 sequence numbers (2, 4, 5) and their packet loss times are described. The packet loss time of the data packet 2 is estimated from the reception times T1 and T3 of the data packets 1 and 3, and the packet loss time of the data packets 4 and 5 is estimated from the reception times T3 and T6 of the data packets 3 and 6. Is done. Specifically, for example, the control packet creation unit 14 interpolates at the reception times T1 and T3 according to the equation (2) to calculate the packet loss time T2 of the data packet 2.
T2 = T1 + (T3-T1) / 2 (2)

The control packet transmitted from the receiving terminal 1 to the transmitting terminal 2 is analyzed by the control packet analyzing unit 24 of the transmitting terminal 2.

The control packet analyzer 24 first extracts the last received packet field and the lost packet list field of the control packet. Then, the control packet analysis unit 24 deletes the data packet corresponding to the sequence number included in the last received packet field and each data packet corresponding to the sequence number before the data packet from the data packet storage unit 23. As a result, the data packet transmitted before the first packet loss occurs is deleted from the data packet storage unit 23. These data packets can be normally received by the receiving terminal 1.

Further, the control packet analysis unit 24 notifies the loss packet retransmission unit 22 of the sequence number of the lost packet described in the lost packet list field, and transmits the description contents of the last received packet field and the lost packet list field to the data packet. It outputs to the speed determination part 25.

The lost packet retransmission unit 22 reads out the data packet corresponding to the sequence number of each lost packet notified from the control packet analysis unit 24 from the data packet storage unit 23 and outputs the data packet to the data packet transmission rate control unit 21 for retransmission. .

The data packet transmission rate determining unit 25 determines the data packet transmission rate based on the contents of the last received packet field and the lost packet list field received from the control packet analyzing unit 24. Here, it is desirable that the data packet transmission rate determination unit 25 determines the data packet transmission rate based on a plurality of final received packet fields and a lost packet list field. That is, when the control packet analysis unit 24 receives a predetermined number of control packets set in advance and passes the last received packet field and the lost packet list field of the predetermined number of control packets to the data packet transmission rate determination unit 25. The data packet transmission rate determination unit 25 preferably determines the data packet transmission rate.

FIG. 6 shows that when the predetermined number of control packets set for the control packet analysis unit 24 is 3, when the control packet analysis unit 24 receives three control packets, the data packet transmission rate determination unit 25 Specific examples of the last received packet field and the lost packet list field that are output are shown. Note that a control packet with a smaller number is received earlier.

As shown in FIG. 6, the sequence packet number and the reception time T1 of the data packet 1 are described in the last received packet field of the control packet 1, and the data packets 2, 4 to be lost packets are stored in the lost packet list field. 5 sequence number and packet loss time.

Further, the sequence packet number and the reception time T4 of the data packet 4 are described in the last received packet field of the control packet 2, and the data packets 5, 6, 10, 11, 15 sequence numbers and packet loss times are described.

Similarly, the last received packet field of the control packet 3 describes the sequence number of the data packet 15 and the reception time T15, and the lost packet list field contains data packets 16, 18, 20, 24 that are lost packets. , 25, 30 sequence numbers and packet loss times are described.

The data packet transmission rate determining unit 25 determines the data packet transmission rate by predicting the network congestion degree at a future time based on the time variation of the number of lost packets. A process in which the data packet transmission rate determination unit 25 calculates the time variation of the number of lost packets will be described.

The data packet transmission rate determination unit 25 first identifies a loss packet (hereinafter referred to as a duplicate loss packet) that is duplicated in the loss packet list of the three control packets. In the case of FIG. 6, the lost packet 5 described in the lost packet list of the control packet 1 and the control packet 2 corresponds. Then, the data packet transmission rate determination unit 25 excludes the duplicate loss packet from the entire loss packet list excluding the “loss packet list in which the duplicate loss packet is listed last”. In the example of FIG. 6, the lost packet 5 is deleted from the lost packet list of the control packet 1. As a result, as shown in FIG. 7, among the three lost packet lists, the lost packet 5 is described only in the lost packet list of the control packet 2.

Next, the data packet transmission rate determination unit 25 refers to the sequence numbers of the respective loss packets described in the three loss packet lists shown in FIG. 7, and stores each loss packet stored in the data packet storage unit 23. The data size of a data packet having the same sequence number as the sequence number is acquired. Then, the data packet transmission rate determining unit 25 sets the reception time described in the three last received packet fields and the packet loss time described in the lost packet list field shown in FIG. The two-dimensional graph of FIG. The lost packet cumulative value is a total data size of lost packets described in the lost packet list at a time before the packet loss time.

For example, as shown in FIG. 7, since the lost packets 5 and 6 exist before the packet loss time T10 of the lost packet 10, the total lost packet size value at the packet loss time T10 is the lost packets 5, 6, 10 (Since control packet 2 is referred to, the sequence number of the last received packet is 4, the size of lost packets 2 and 4 described in control packet 1 is not included).

Next, the data packet transmission rate determination unit 25 approximates the straight line graph composed of the packet loss time and the total lost packet size value shown in FIG. 8 using an approximation algorithm such as a least square method for each control packet. The result of linear approximation of the straight line graph for each of the control packets 1, 2, and 3 in FIG. 8 is shown as straight lines 1, 2, and 3 in FIG. The slopes 1 (= Sz1 / (T4-T1)), 2 (= Sz2 / (T15-T4)), and 3 (= Sz3 / (T30-T15)) of each of the straight lines 1, 2, and 3 are the time widths ( The straight line 1 corresponds to the packet loss traffic (loss packet traffic) in the data packet traffic transmitted from the transmitting terminal 2 to the receiving terminal 1 in T1 to T4).

Then, the data packet transmission rate determination unit 25 determines the transmission rate of the data packet by obtaining the dispersion values of the inclinations 1, 2, and 3 of the three straight lines shown in FIG. Specifically, for example, when the variance value is smaller than a preset threshold value, it is determined that “the packet loss has occurred but the network congestion is low”, and the data packet transmission rate is increased by a predetermined amount. Turn into. This predetermined amount may be set in advance. On the other hand, if the variance values of the slopes 1, 2, and 3 are greater than or equal to a preset threshold value, it is determined that the network congestion level is increasing and corresponds to the maximum value of the slopes 1, 2, and 3. Decrease the data packet transmission rate by as much.

The data packet transmission rate determination unit 25 may determine the data packet transmission rate using a method other than the above. For example, the data packet transmission rate determination unit 25 linearly approximates the slopes of the three straight lines 1, 2, and 3 shown in FIG. 9 that are lost packet traffic using an approximation algorithm such as the least square method, and creates the straight line shown in FIG. To do. Thereafter, the data packet transmission rate determination unit 25 predicts the loss packet traffic of the control packet (in this case, the control packet 4) received by the transmitting terminal 2 that is not currently received and at a future time using the straight line in FIG. To do. If the slope of the straight line is positive, the data packet transmission rate determination unit 25 determines that the network congestion is high, and reduces the data packet transmission rate by an amount corresponding to the lost packet traffic 4. On the other hand, when the slope of the straight line is negative, the data packet transmission rate determination unit 25 determines that the network congestion level is low and corresponds to the minimum value of the lost packet traffic 1, 2, 3, 4 The data packet transmission speed is increased by the amount of data.

The data packet transmission rate control unit 21 transmits all data packets including the lost packet to be retransmitted at the data packet transmission rate set by the data packet transmission rate control unit 21.

As described above, the data transmission system 100 according to the present embodiment measures the network data by the receiving terminal, predicts the network congestion degree at the future time based on the network data measured by the receiving terminal, and predicts the network congestion. The control packet transmission rate of the receiving terminal and the data packet transmission rate of the transmitting terminal are determined according to the degree. By doing so, it is possible to set the transmission speed of the control packet and the data packet by following the degree of congestion of the network, and to prevent transmission delay, packet loss and the like due to network congestion. Therefore, data transmission can be efficiently performed even in a network in which the network congestion degree changes in a short period.
<Second Embodiment>

FIG. 11 shows a data transmission system 200 according to the second embodiment of the present invention. The data transmission system 200 includes a receiving terminal 5 and a transmitting terminal 6 connected by a network 3. In the data transmission system 200 of FIG. 11, the same reference numerals are given to components having the same configuration or function as those of the data transmission system 100 of the first embodiment shown in FIG. In the description of, detailed description of these components is omitted.

The receiving terminal 5 includes a network data measurement unit 11, a data packet transmission rate determination unit 55, a control packet transmission rate determination unit 52, a control packet creation unit 54, and a control packet transmission rate control unit 15. The data packet transmission rate determination unit 55 is not provided in the receiving terminal 1 of the data transmission system 100. Further, the control packet transmission rate determination unit 52 determines the transmission rate of the control packet by a method different from that of the control packet transmission rate determination unit 12 of the receiving terminal 1, and the control packet creation unit 54 14 creates a control packet that is different from the control packet created by 14. Further, the receiving terminal 5 is not provided with the control packet transmission rate determining unit 12 provided in the receiving terminal 1.

The data packet transmission rate determination unit 55 uses the network data measured by the network data measurement unit 11 to determine the transmission rate of the data packet transmitted by the transmission terminal 6 at a future time. The determination method is the same as that by the data packet transmission rate determination unit 25 in the transmission terminal 2 of the data transmission system 100, and detailed description thereof is omitted here.

The control packet transmission rate determination unit 52 transmits the data packet transmission rate determination unit 55 from the reception terminal 5 to the transmission terminal 6 according to the network congestion level obtained when the transmission rate of the data packet by the transmission terminal 6 is determined. Determine the transmission rate of control packets.

The control packet created by the control packet creation unit 54 is further transmitted in addition to the last received packet field and the lost packet list field that are the same as the control packet created by the control packet analysis unit 24 in the receiving terminal 1 of the data transmission system 100. Contains a velocity field. The control packet creation unit 54 describes the data packet transmission rate determined by the data packet transmission rate determination unit 55 in the data packet transmission rate field.

The control packet created by the control packet creation unit 54 is transmitted to the transmission terminal 6 by the control packet transmission rate control unit 15 at the transmission rate determined by the network data measurement unit 11.

The transmission terminal 6 includes a data packet transmission rate control unit 21, a data packet shaping unit 66, a lost packet retransmission unit 22, a data packet storage unit 23, and a control packet analysis unit 64. The data packet shaping unit 66 is not provided in the transmission terminal 2 of the data transmission system 100. In the present embodiment, the control packet transmitted from the receiving terminal 5 to the transmitting terminal 6 is different from the control packet transmitted from the receiving terminal 1 to the transmitting terminal 2 in the data transmission system 100. For this reason, the processing target of the control packet analysis unit 64 is a control packet different from the analysis target of the control packet analysis unit 24 in the transmission terminal 2. Further, the transmission terminal 6 is not provided with the data packet transmission rate determination unit 25 provided in the transmission terminal 2.

The control packet analysis unit 64 further extracts the data packet transmission rate described in the data packet transmission rate field of the control packet, in addition to the processing performed by the control packet analysis unit 24 in the transmission terminal 2 of the data transmission system 100. The control packet analysis unit 64 outputs the extracted data packet transmission rate to the data packet transmission rate control unit 21.

The data packet shaping unit 66 shapes all data packets to be transmitted (including lost packets to be retransmitted) so as to have a preset packet size.

The data packet transmission rate control unit 21 transmits the data packet shaped by the data packet shaping unit 66 to the receiving terminal 5 at the data packet transmission rate output from the control packet analysis unit 64.

Next, the flow of processing by the data transmission system 200 of this embodiment will be described with reference to the flowchart of FIG.
The operations of the data packet transmission rate control unit 21 and the loss packet retransmission unit 22 in the transmission terminal 6 of the data transmission system 200 indicated by steps S1, S11, S12, S14, S21, S22, and S2 in FIG. Since the operations of the data packet transmission rate control unit 21 and the loss packet retransmission unit 22 in the transmission terminal 2 are the same, the description thereof is omitted.

The operations of the network data measuring unit 11 and the control packet transmission rate control unit 15 in the receiving terminal 5 of the data transmission system 200 shown in steps D1, D2, and D5 of FIG. Since the operations of the data measuring unit 11 and the control packet transmission rate control unit 15 are the same, the description thereof is omitted.

In the data transmission system 100, the size of the data packet transmitted from the transmitting terminal 2 to the receiving terminal 1 is different for each data packet, and the data packet size is not uniform. In the data transmission system 200 of the present embodiment, the data packet shaping unit 66 of the transmission terminal 6 performs shaping so that the size of the data packet read from the data packet storage unit 23 becomes a predetermined size set in advance. (Step S15). Specifically, when the size of the data packet is smaller than the predetermined size, the data packet shaping unit 66 sets the size to a predetermined size by padding dummy data and then controls the data packet transmission rate control. The data is output to the unit 21 and transmitted to the receiving terminal 5. On the other hand, when the size of the data packet is larger than the predetermined size, the data packet shaping unit 66 divides the data packet into a plurality of packets having a predetermined size or less. When the size of the divided data packet is smaller than the predetermined size, the data packet shaping unit 66 sets the size to a predetermined size by padding or the like.

In the data transmission system 100, the transmission rate of the control packet is determined by the control packet transmission rate determination unit 12 of the receiving terminal 1 using the network congestion model stored in the network congestion model storage unit 13. Further, the data packet transmission rate is determined based on the analysis result of the control packet by the data packet transmission rate determination unit 25 of the transmission terminal 2. In the data transmission system 200 of the present embodiment, both the control packet transmission rate and the data packet transmission rate are determined on the receiving terminal 5 side. Specifically, the data packet transmission rate determining unit 55 of the receiving terminal 5 determines the data packet transmission rate (step D6), and the control packet transmission rate determining unit 52 determines the control packet transmission rate. Further, the network congestion model is not used when determining either the control packet transmission rate or the data packet transmission rate.

In the data transmission system 100, when determining the data packet transmission speed (S13 in FIG. 2), it is necessary to acquire the size of the lost packet to be retransmitted. In the data transmission system 200 of the present embodiment, all data packets are shaped to the same predetermined size before being transmitted to the receiving terminal 5, so when determining the data packet transmission rate in step D6 of FIG. The data packet transmission rate may be determined using a predetermined size set in advance.

When the control packet transmission rate determining unit 52 determines the control packet transmission rate (step D3 in FIG. 12), the control packet transmission rate determining unit 52 determines the network determined by the data packet transmission rate determining unit 55 in step D6. The control packet transmission rate is determined based on the degree of congestion. Specifically, for example, when the data packet transmission rate determination unit 55 determines that the network congestion level is a congestion state, the control packet transmission rate determination unit 52 increases the control packet transmission rate. On the other hand, when the data packet transmission rate determining unit 55 determines that the network congestion level is the normal state, the control packet transmission rate determining unit 52 decreases the control packet transmission rate.

The control packet creation unit 54 describes the description contents in the last received packet field and the lost packet list field. At the same time, in order to notify the transmission terminal 6 of the data packet transmission rate determined by the data packet transmission rate determination unit 55, the control packet creation unit 54 further describes the control packet by describing the data packet transmission rate in the transmission rate field. Create (step D4).

The control packet transmission rate control unit 15 transmits the control packet created by the control packet creation unit 54 in Step D4 to the transmission terminal 6 at the control packet transmission rate decided by the control packet transmission rate decision unit 52 in Step D3 (Step S4). D5).

In addition to the processes performed by the control packet analysis unit 24 in the transmission terminal 2 of the data transmission system 100, the control packet analysis unit 64 of the transmission terminal 6 further receives the data packet transmission rate from the data packet transmission rate field in the received control packet. Is extracted and set in the data packet transmission rate control unit 21 (step S23).

Next, differences in the data transmission system 200 from the data transmission system 100 will be described in more detail with specific examples.
In the data transmission system 200, a predetermined data packet size is set for the receiving terminal 5 and the transmitting terminal 6 in order to unify the data packet size. The transmission terminal 6 shapes the data packet so that the data packet to be transmitted to the reception terminal 5 has the predetermined size. With reference to FIG. 13, an example of a data packet shaping process by the transmission terminal 6 will be described.

FIG. 13 shows an example in which the transmission terminal 6 shapes the data packet 1 smaller than the set predetermined size and the data packet 2 larger than the predetermined size into data packets of a predetermined size. The transmission terminal 6 first divides the data packet 2 into two data packets 2a and 2b. At this time, the sum of the sizes of the data packet 2a and the data packet 1 is set to a predetermined size. Then, the transmission terminal 6 connects the data packet 1 and the data packet 2a to create the shaped data packet 1. For the other data packet 2b obtained by dividing the data packet 2, the transmitting terminal 6 sets the size to a predetermined size by padding dummy data, and obtains the shaped data packet 2.

In the example shown in FIG. 13, the size of the data packet 2b obtained by the division is smaller than the predetermined size. If the size of the data packet 2b is larger than the predetermined size, the data packet 2b may be further shaped, for example, by dividing it into a data packet of a predetermined size and another data packet.

FIG. 14 shows the configuration of the control packet created by the control packet creation unit 54 in the receiving terminal 5 of the data transmission system 200. As can be seen from comparison with the configuration of the control packet in the data transmission system 100 (FIG. 5), the control packet in the data transmission system 200 has a data packet transmission rate field in which the data packet transmission rate determined on the receiving terminal 5 side is described. Have Further, in the data transmission system 200, the receiving terminal 5 side determines the data packet transmission rate and notifies the transmitting terminal 6 and the transmitting terminal 6 side transmits at the notified data packet transmission rate. For this reason, in the last received packet field of the control packet in the data transmission system 200, only the maximum value of the sequence number of the data packet successfully received by the receiving terminal is described, and the reception time of the data packet is described as unnecessary. It has not been. Similarly, in the lost packet list field, only the sequence number of each lost packet is described, and the packet loss time is not described as unnecessary.

The data transmission system 200 of the present embodiment has the following advantages in addition to the effects obtained by the data transmission system 100.

In the data transmission system 200, when the transmission terminal 6 transmits a data packet (including a loss packet to be retransmitted) to the reception terminal 5, the transmission terminal 6 shapes the data packet so as to have the same size. Thereby, it is not necessary to describe the reception time and the packet loss time of the last received packet in the control packet returned from the receiving terminal 5 to the transmitting terminal 6, and the size of the control packet can be reduced. Further, since the control packet transmission rate determination unit 52 and the data packet transmission rate determination unit 55 determine the control packet transmission rate and the data packet transmission rate by the same method, the control packet transmission rate determination unit 52 determines the data packet transmission rate determination. Part of the calculation results obtained by the unit 55 can be used, and the process for determining the transmission rate of the control packet can be speeded up. These have an advantageous feature that high-speed data transmission can be realized even when the number of transmitting terminals is increased.

Further, since the receiving terminal 5 determines the control packet transmission rate and the data packet transmission rate without using the network congestion model, the storage capacity can be saved by the size of the network congestion model.

The present invention has been described above based on the embodiments. The embodiment is an exemplification, and various modifications, increases / decreases, combinations of the first and second embodiments, and the like may be made without departing from the gist of the present invention. It will be understood by those skilled in the art that modifications to which these changes, increases / decreases, and combinations are also within the scope of the present invention.

For example, in the data transmission system 100 according to the first embodiment, a model in which the network congestion level is roughly divided into a normal state and a congestion state as illustrated in FIG. 4 is used, but other states of the network congestion level are used. Alternatively, a model to which state transition conditions are added may be used.

Further, in the data transmission system 100, the network congestion model is used for determining the control packet transmission rate, and the network congestion model is not used for determining the data packet transmission rate. Conversely, the control packet transmission rate determination unit determines the control packet transmission rate without using the network congestion degree model, and the data packet transmission rate determination unit determines the data packet transmission rate using the network congestion degree model. Also good.

Of course, the network congestion model may be used when determining either the control packet transmission rate or the data packet transmission rate.

Further, the method for determining the data packet transmission rate is not limited to the above-described methods.

This application claims priority based on Japanese Patent Application No. 2009-240226 filed on October 19, 2009, the entire disclosure of which is incorporated herein.

The technology according to the present invention can be used in a data transmission system that collects log files for data center operation management, a content distribution system such as video and audio, and a data communication system in a thin client service.

DESCRIPTION OF SYMBOLS 100 Data transmission system 1 Reception terminal 2 Transmission terminal 3 Network 5 Reception terminal 6 Transmission terminal 11 Network data measurement part 12 Control packet transmission rate determination part 13 Network congestion degree model memory | storage part 14 Control packet creation part 15 Control packet transmission rate control part 21 Data packet transmission rate control unit 22 Loss packet retransmission unit 23 Data packet storage unit 24 Control packet analysis unit 25 Data packet transmission rate determination unit 200 Data transmission system 52 Control packet transmission rate determination unit 54 Control packet creation unit 55 Data packet transmission rate determination 64 Control packet analysis unit 66 Data packet shaping unit

Claims (13)

1. A transmitting terminal and a receiving terminal are connected by a network, the transmitting terminal transmits a data packet to the receiving terminal via the network, and the receiving terminal confirms with the transmitting terminal in response to reception of the data packet from the transmitting terminal A transmission rate control method in a data transmission system for returning a response packet,
   The receiving terminal measures network data that can predict the degree of congestion of the network,
   For each period of a predetermined length, the network data measured over the period is used to predict the degree of network congestion,
   A transmission rate control method, comprising: determining a transmission rate of an acknowledgment packet transmitted from the receiving terminal according to the predicted degree of congestion.
2. The transmission rate control method according to claim 1, further comprising: determining a transmission rate of a data packet to be transmitted from the transmission terminal according to the predicted congestion level of the network.
3. A transmitting terminal and a receiving terminal are connected by a network, the transmitting terminal transmits a data packet to the receiving terminal via the network, and the receiving terminal confirms with the transmitting terminal in response to reception of the data packet from the transmitting terminal A data transmission system for returning a response packet,
   The receiving terminal is
   Network data measuring means for measuring network data capable of predicting the degree of congestion of the network;
   First congestion degree prediction means for predicting the congestion degree of the network using the network data measured by the network data measurement means over the period for each predetermined length period;
   A data transmission system comprising: an acknowledgment packet transmission rate determining unit that determines a transmission rate of the acknowledgment packet according to the congestion level predicted by the first congestion level predicting unit.
4. 4. The data transmission system according to claim 3, wherein the network data includes information that can indicate an occurrence state of a lost packet among data packets transmitted from the transmitting terminal.
5. The receiving terminal further includes:
   A data packet transmission rate determining unit for determining a transmission rate at which the transmitting terminal transmits a data packet to the receiving terminal according to the congestion degree predicted by the first congestion degree predicting unit;
   An acknowledgment packet creating means for creating the acknowledgment packet;
   The confirmation response packet creation means includes the data packet transmission speed determined by the data packet transmission speed determination means in the confirmation response packet,
   The data transmission system according to claim 3 or 4, wherein the transmission terminal transmits a data packet to the reception terminal at the data packet transmission rate included in the acknowledgment packet returned from the reception terminal. .
6. The receiving terminal further comprises an acknowledgment packet creating means for creating the acknowledgment packet;
   The transmitting terminal is
   A second congestion degree prediction means for predicting the congestion degree of the network;
   Data packet transmission rate determining means for determining a transmission rate for transmitting data packets to the receiving terminal according to the congestion degree predicted by the second congestion degree predicting means,
   The acknowledgment packet creating means is information required by the second congestion degree predicting means and the data packet transmission rate determining means, and at least a part of the network data measured by the network data measuring means or the The information that is data calculated from network data is included in the acknowledgment packet,
   The second congestion degree prediction means predicts the congestion degree of the network from the information included in a plurality of control packets received within the period for each predetermined length period,
   5. The data transmission system according to claim 3, wherein the data packet transmission rate determining unit determines a data packet transmission rate according to the congestion level predicted by the second congestion level prediction unit. .
7. A receiving terminal that is connected to a transmitting terminal that transmits a data packet by a network and returns an acknowledgment packet to the transmitting terminal in response to reception of the data packet from the transmitting terminal;
   Network data measuring means for measuring network data capable of predicting the degree of congestion of the network;
   First congestion degree prediction means for predicting the congestion degree of the network using the network data measured by the network data measurement means over the period for each predetermined length period;
   A receiving terminal comprising: an acknowledgment packet transmission rate determining unit that determines a transmission rate of the acknowledgment packet according to the congestion level predicted by the first congestion level predicting unit.
8. The receiving terminal according to claim 7, wherein the network data includes information that can indicate an occurrence state of a lost packet among data packets transmitted from the transmitting terminal.
9. A data packet transmission rate determining unit for determining a transmission rate at which the transmitting terminal transmits a data packet to the receiving terminal according to the congestion degree predicted by the first congestion degree predicting unit;
   An acknowledgment packet creating means for creating the acknowledgment packet;
   9. The receiving terminal according to claim 7, wherein the confirmation response packet creating unit includes the data packet transmission rate determined by the data packet transmission rate determining unit in the confirmation response packet.
10. An acknowledgment packet creating means for creating the acknowledgment packet;
    The confirmation response packet creating means is information necessary for predicting the degree of congestion of the network by the transmitting terminal and determining the transmission rate of the data packet according to the predicted degree of congestion, and measured by the network data measuring means The receiving terminal according to claim 7, wherein the information that is at least a part of the network data or data calculated from the network data is included in the acknowledgment packet.
11. A transmission terminal connected to a reception terminal via a network, transmitting a data packet to the reception terminal, and receiving an acknowledgment packet returned by the reception terminal in response to reception of the data packet;
    The acknowledgment packet includes the transmission rate of the data packet determined according to the degree of network congestion predicted by the receiving terminal,
    A transmitting terminal that transmits a data packet to the receiving terminal at a transmission speed of the data packet included in the acknowledgment packet.
12. The transmitting terminal is
    12. The transmission terminal according to claim 11, further comprising data packet shaping means for shaping a data packet to be transmitted to the reception terminal so as to have a predetermined size.
13. A transmission terminal connected to a reception terminal via a network, transmitting a data packet to the reception terminal, and receiving an acknowledgment packet returned by the reception terminal in response to reception of the data packet;
    A second congestion degree prediction means for predicting the congestion degree of the network;
    Data packet transmission rate determining means for determining a transmission rate for transmitting data packets to the receiving terminal according to the congestion degree predicted by the second congestion degree predicting means,
    The acknowledgment packet is information required by the second congestion degree predicting means and the data packet transmission rate determining means, and at least a part of the network data measured on the receiving terminal device side or the information The information that is data calculated from network data is included,
    The second congestion degree prediction means predicts the congestion degree of the network from the information included in a plurality of control packets received within the period for each predetermined length period,
    The transmission terminal characterized in that the data packet transmission rate determination means determines the transmission rate of the data packet according to the congestion degree predicted by the second congestion degree prediction means.

Images