JP7482098B2 - COMMUNICATIONS DEVICE AND PROGRAM FOR TRANSMITTING AND RECEIVING DATA USING MULTIPLE STREAMS - Google Patents

Description

The present invention relates to a technology for transmitting and receiving data using multiple streams.

TCP (Transmission Control Protocol) is used to send and receive data over a network. Congestion control is performed when sending and receiving data using TCP. Specifically, a communication device that receives data (hereafter referred to as a receiving device) notifies the communication device that transmits the data (hereafter referred to as a transmitting device) of the reception status of packets from the transmitting device by using ACK or NACK. Note that ACK indicates that a packet has been received normally, and NACK indicates that a lost packet has occurred. Note that a lost packet includes a packet that was not received normally by the receiving device, that is, a packet that did not arrive at the receiving device, and a packet that arrived at the receiving device but contained an error.

In congestion control, the transmitting device manages the window size. The window size specifies the number of packets that the transmitting device can transmit without receiving an ACK from the receiving device. The transmitting device increases or decreases the window size based on the reception status reported by the receiving device, i.e., ACK or NACK. Generally, the transmitting device decreases the window size when it receives a NACK, and increases the window size according to a predetermined algorithm while only an ACK is being received. The number of packets transmitted by the transmitting device according to the window size is adjusted according to the network state, thereby achieving congestion control. Non-Patent Document 1 discloses a congestion control method called CUBIC. According to Non-Patent Document 1, the transmitting device increases the window size according to a three-dimensional function while only an ACK is being received.

Non-Patent Document 2 also discloses MPTCP (Multi-Path TCP), which transmits and receives data using multiple streams in order to increase the data transmission speed from a transmitting device to a receiving device. In MPTCP, congestion control is performed for each stream. Note that a stream may also be called a connection or a path.

"CUBIC: A New TCP-Friendly High-Speed TCP Variant", ACM SIGOPS Operating Systems Review, Volume 42 Issue 5, July 2008 IEEE/ACM Transactions on Networking (Volume: 14, Issue: 6, December 2006), "Multi-Path TCP: A Joint Congestion Control and Routing Scheme to Exploit Path Diversity in the Internet"

As described in Non-Patent Document 2, by using multiple streams, a higher data transmission speed (throughput) can be achieved compared to transmitting data using a single stream. However, considering that packets are lost in each stream, a huge number of streams must be set in order to achieve a higher throughput, which increases the processing load of the communication device.

The present invention provides technology that reduces the processing load on communication devices.

According to one aspect of the present invention, a communication device that transmits packets including data to a receiving device via a network is characterized by comprising: a plurality of streams; a setting means for setting one or more substreams for each of the plurality of streams; an allocation means for allocating the packets to one of the plurality of streams; and a transmission means for transmitting the packets allocated to the stream on each of the one or more substreams of the stream.

This invention makes it possible to reduce the processing load on communication devices.

FIG. 1 is a configuration diagram of a communication system used to explain an embodiment. FIG. 2 is an explanatory diagram of a data transmission and reception method according to an embodiment. FIG. 2 is an explanatory diagram of a data transmission and reception method according to an embodiment. FIG. 4 is a sequence diagram of a process performed at the start of data transmission and reception according to an embodiment. FIG. 4 is a sequence diagram of a process performed during data transmission and reception according to an embodiment. FIG. 2 is a diagram showing a configuration of a transmitting side of a communication device according to an embodiment. FIG. 2 is a diagram illustrating a receiving side configuration of a communication device according to an embodiment.

The following embodiments are described in detail with reference to the attached drawings. Note that the following embodiments do not limit the invention according to the claims, and not all combinations of features described in the embodiments are necessarily essential to the invention. Two or more of the features described in the embodiments may be combined in any desired manner. In addition, the same reference numbers are used for the same or similar configurations, and duplicate descriptions are omitted.

Figure 1 is a configuration diagram of a communication system used to explain this embodiment. A plurality of communication devices are connected to a network 30. Each of the communication devices is capable of transmitting and receiving data. In the following, a communication device that transmits data is referred to as a transmitting device 10, and a communication device that receives data transmitted by the transmitting device 10 is referred to as a receiving device 20. The transmitting device 10 and the receiving device 20 can each be connected to the network by one or more interfaces. The one or more interfaces can be, for example, an interface for connecting to a wireless local area network (LAN), an interface for connecting to a wired LAN, a wireless interface standardized by 3GPP (3rd Generation Partnership Project), etc. The network 30 in Figure 1 includes various types of networks such as a mobile communication network and the Internet.

Figure 2 is an explanatory diagram of data transmission from a transmitting device 10 to a receiving device 20. In this embodiment, as described in Non-Patent Document 1, a packet including data is transmitted by one of multiple streams. In Figure 2, two streams, stream 50 and stream 60, are used, but this is an example, and the number of streams can be three or more.

Each stream is made up of one or more substreams. In FIG. 2, stream 50 is made up of substreams 50-1, 50-2, and 50-3, and stream 60 is made up of substreams 60-1, 60-2, and 60-3. Note that the number of substreams making up each stream may differ for each stream.

Each time transmitting device 10 receives a packet (data) to be transmitted from an application program, it determines which stream to use to transmit the packet, that is, it determines which stream to assign the packet to. Any method can be used for this assignment. For example, transmitting device 10 can assign new packets to each stream in a round-robin format. Transmitting device 10 can also determine the number of packets waiting to be transmitted for each stream and assign new packets to streams with the fewest number of packets waiting to be transmitted. Meanwhile, in this embodiment, transmitting device 10 transmits packets assigned to a stream through all sub-streams that make up that stream.

Figure 3 shows the method of transmitting packets by the above-mentioned transmitting device 10. In Figure 3, the transmitting device 10 uses two streams #1 and #2 to transmit data (packets). Note that stream #1 has two substreams #1-1 and #1-2, and stream #2 has two substreams #2-1 and #2-2. In Figure 3, the transmitting device 10 assigns packets #1 to #6 to each stream in a round-robin format. Therefore, packets #1, #3, and #5 are transmitted in stream #1, and packets #2, #4, and #6 are transmitted in stream #2. Since packets assigned to a stream are transmitted in all substreams of the stream, packets #1, #3, and #5 are transmitted in substreams #1-1 and #1-2, and packets #2, #4, and #6 are transmitted in substreams #2-1 and #2-2. Note that it is the substreams that are actually transmitted to the network 30, and the network 30 does not recognize the streams.

In this embodiment, the transmitting device 10 and the receiving device 20 perform congestion control on a stream-by-stream basis. In FIG. 3, the shaded packets indicate lost packets, that is, packets in which an error was detected by the receiving device 20, or packets that did not arrive at the receiving device 20. In stream #1-1, packet #1 is a lost packet, and in stream #1-2, packet #3 is a lost packet. However, packet #1 is received normally in stream #1-2, and packet #3 is received normally in stream #1-1. Therefore, the receiving device 20 notifies the transmitting device 10 that there are no lost packets in stream #1, that is, sends an ACK.

On the other hand, packet #6 is a lost packet in both streams #2-1 and #2-2. Therefore, in stream #2, receiving device 20 notifies transmitting device 10 of a NACK indicating the loss of packet #6. Note that the ACK/NACK for a stream is transmitted in all substreams of that stream.

Figure 4 is a sequence diagram showing the process when the transmitting device 10 starts transmitting data to the receiving device 20. In S1, the transmitting device 10 determines the number of streams N based on the target speed S. The target speed S is notified from the application program that transmits the data. For example, the transmitting device 10 can determine the number of streams N by rounding up the decimal point of the value obtained by dividing the target speed S by the average speed or peak speed of each stream. In this case, if the target speed S is 10 Gbps and the average speed of each stream is 2 Gbps, the number of streams N is 5. In addition, in order to allow for a margin, this value can be multiplied by a predetermined number. For example, if the predetermined number is 2, the number of streams N is 10. Note that the average speed or peak speed of each stream can use a value measured in past communication. If there is no value measured in past communication, a predetermined initial value can be used. In addition, the transmitting device 10 can set the number of interfaces to the number of streams. For example, if the number of interfaces for accessing the network 30 is 3, the transmitting device 10 can determine the number of streams to be 3.

In S2, the transmitting device 10 determines the number of substreams M for each stream based on the target loss rate L of the stream. The target loss rate L of a stream is a target value for the ratio of the number of lost packets of the stream to the number of packets transmitted in the stream. The target loss rate L of the stream is set in advance in the transmitting device 10. For example, the transmitting device 10 determines the number of substreams M of a certain stream as follows:
M = ceil(log x L) (1)
Here, x is the loss rate of the substream, and ceil() is a ceiling function. At the start of communication, a predetermined initial value is used for x. For example, when x=0.001 and L=0.000001, the number of substreams M is 2.

The transmitting device 10 determines an identifier for each stream and a sender identifier for each substream. For example, the sender identifier for a substream is a combination of the sender IP address of the packets sent in that substream and the sender port number of the TCP packets. The stream identifier identifies the stream and can be any value. In S3, the transmitting device 10 notifies the receiving device 20 of the sender identifier of the substream and the identifier of the stream to which the substream belongs for each substream. Note that the interfaces for transmitting multiple substreams belonging to one stream may be the same or different.

In S3, the receiving device 20 receives from the transmitting device 10 the sender identifier of the substream and the identifier of the stream to which the substream belongs, thereby recognizing the number of streams and the number of substreams for each stream. The receiving device 20 then determines the receiver identifier corresponding to the sender identifier of each substream. The receiver identifier of a substream is a combination of the destination IP address of the packet received in that substream and the receiver port number of the TCP packet. Note that the interface for receiving multiple substreams belonging to one stream may be the same or different. In S4, the receiving device 20 notifies the transmitting device 10 of the correspondence between the sender identifier and receiver identifier of the substream. From this notification, the transmitting device 10 recognizes the destination IP address and destination port number used in each substream.

Then, the transmitting device 10 transmits the data as described with reference to Figures 2 and 3. The receiving device 20 also determines whether a packet is lost for each stream, and notifies the transmitting device 10 if a packet is lost.

Figure 5 is a sequence diagram of the substream change process executed during data transmission. The receiving device 20 counts the number of lost packets for each substream. For example, when the receiving device 20 receives packets as shown in Figure 3, the number of lost packets for each substream is 1. In S10, the receiving device 20 notifies the transmitting device 10 of the number of lost packets for each substream at a predetermined period. When the transmitting device 10 receives the number of lost packets for each substream, it determines the loss rate, which is the ratio of lost packets occurring in each substream, based on the number of packets transmitted in each substream. Next, the transmitting device 10 calculates the average value of the loss rates for each substream in the same stream, and in S11, the transmitting device 10 calculates the number of substreams for the stream by using this average value as x in equation (1).

If the number of substreams determined in S11 is the same as the current number of substreams, the transmitting device 10 maintains the current number of substreams. On the other hand, if the number of substreams determined in S11 is greater than the current number of substreams, the transmitting device 10 adds a substream so that the number of substreams becomes the number of substreams determined in S11. For this reason, the transmitting device 10 notifies the receiving device 20 of the transmitting side identifier of the substream to be added in S12. Note that this notification is performed using the current substream of the same stream, so there is no need to notify the identifier of the stream. The receiving device 20 determines the receiving side identifier of the substream to be added, and notifies the transmitting device 10 of the correspondence between the transmitting side identifier and the receiving side identifier of the substream to be added in S13. As a result, a new substream is added to the stream. Also, if the number of substreams determined in S11 is less than the current number of substreams, the transmitting device 10 deletes a substream so that the number of substreams becomes the number of substreams determined in S11. For this reason, in S12, the transmitting device 10 notifies the receiving device 20 of the transmitting side identifier of the substream to be deleted. Note that if the number of substreams determined in S11 is less than the current number of substreams, the configuration may be such that no substreams are deleted and the current number of substreams is maintained.

Figure 6 is a configuration diagram of the transmitting device 10 according to this embodiment. As described above, the stream determination unit 101 determines the number of streams to be set based on the target speed S, and notifies the substream determination unit 102 and the communication unit 103 of the determined number of streams. As described above, the substream determination unit 102 determines the number of substreams for each stream based on the target loss rate L, and notifies the communication unit 103 of the number of substreams for each stream.

The congestion control unit 105 of the communication unit 103 determines the identifier of each stream and the sender identifier of each substream, and notifies the receiving device 20 of the sender identifier of each substream and which stream's identifier each substream belongs to. Furthermore, when the congestion control unit 105 receives the correspondence between the sender identifier and the receiver identifier of the substream from the receiving device 20, it starts transmitting packets in each substream. The allocation unit 104 packetizes the data received from the application program, and allocates each packet to one of the multiple streams.

During data transmission, the congestion control unit 105 manages the window size for each stream. The congestion control unit 105 also receives reception status information indicating the reception status of packets from the receiving device 20. The reception status information includes information indicating the reception status at the substream level and information indicating the reception status at the stream level. The reception status at the substream level indicates the number of lost packets (first lost packets) that have occurred in each substream. The congestion control unit 105 obtains a loss rate for each substream based on the number of first lost packets for each substream notified from the receiving device 20 and notifies the substream determination unit 102. The substream determination unit 102 determines the number of substreams for each stream as described above based on the loss rate of each substream and notifies the congestion control unit 105. The congestion control unit 105 determines whether it is necessary to add/remove a substream based on the number of substreams for each stream notified from the substream determination unit 102, and if necessary, performs processing for adding/removing a substream as described above.

On the other hand, the stream-level reception status indicates lost packets (second lost packets) that occurred in each stream. The second lost packet is a packet that became the first lost packet in all substreams of the stream. When the congestion control unit 105 is notified of a second lost packet, it retransmits the second lost packet in all substreams that transmitted the second lost packet. Note that the target loss rate L in the above formula (1) is the target value for the occurrence rate of second lost packets. The target loss rate L may differ for each stream.

Figure 7 is a configuration diagram of the receiving device 20. When the congestion control unit 202 of the communication unit 201 receives the identifier of each stream and the sender identifier of each substream belonging to each stream from the transmitting device 10, it determines the receiver identifier that pairs with the sender identifier of each substream and notifies the transmitting device 10 of the correspondence between the sender identifier and the receiver identifier of each substream.

While receiving data, the congestion control unit 202 performs congestion control on a stream-by-stream basis. In other words, when a packet is received in any of the substreams within a stream, an ACK is sent as an indication that the packet was received normally, even if the packet is a lost packet in the other substreams. On the other hand, for packets that were not received normally in any of the substreams (second lost packets), a NACK is sent to prompt the transmitting device 10 to retransmit the packet. In this way, the congestion control unit 202 performs congestion control on a stream-by-stream basis, and outputs the data of the received packets to the application program.

Furthermore, the congestion control unit 202 counts the number of first lost packets for each substream for each predetermined period and notifies the transmitting device 10 of the number of first lost packets as reception status information. When the transmitting device 10 notifies the congestion control unit 202 of the addition/deletion of a substream according to the notified reception status information, the congestion control unit 202 adds/deletes the substream.

As described above, in this embodiment, a stream is composed of one or more substreams, and the same packets are redundantly transmitted in one or more substreams of the same stream. This configuration can substantially reduce the loss rate of one stream. Therefore, the number of streams required can be smaller than the number of streams required in conventional MPTCP for the same loss rate, and therefore the processing load on the communication device can be reduced.

In this embodiment, the number of substreams for each stream is calculated using formula (1), but the present invention is not limited to calculating the number of substreams for each stream using formula (1). Specifically, any other method can be used as long as it calculates the number of substreams required to make the occurrence rate of second lost packets in a stream (loss rate) equal to or less than the target value for that stream (L in formula (1)).

The communication device according to the present invention, i.e., the above-mentioned transmitting device 10 and receiving device 20, can be realized by a program that, when executed by one or more processors of a device, causes the device to operate as the transmitting device 10 and/or the receiving device 20. These programs can be stored in a device-readable storage medium or distributed via a network.

The above configuration enables high-speed data transfer while reducing processing load. This will contribute to achieving Goal 9 of the United Nations-led Sustainable Development Goals (SDGs), which is to "build resilient infrastructure, promote sustainable industrialization and foster innovation."

The invention is not limited to the above-described embodiment, and various modifications and variations are possible within the scope of the invention.

104: Allocation unit, 105: Congestion control unit

Claims (9)

A communication device that transmits a packet including data to a receiving device via a network, comprising:
A plurality of streams, and a setting means for setting one or more sub-streams for each of the plurality of streams;
an allocation means for allocating the packet to any one of the plurality of streams;
a transmitting means for transmitting the packets allocated to the stream on each of the one or more sub-streams of the stream;
A communication device comprising: The communication device according to claim 1, characterized in that the transmission means manages the window size in units of streams. The communication device according to claim 1 or 2, characterized in that the transmission means acquires from the receiving device reception status information indicating the number of first lost packets that the receiving device was unable to receive normally for each of the one or more substreams of the multiple streams, and determines whether or not to change the number of the one or more substreams for each of the multiple streams based on the number of first lost packets indicated by the reception status information. The communication device according to claim 3, characterized in that, when a second lost packet that could not be normally received is notified from the receiving device, the transmitting means retransmits the second lost packet in each of the one or more substreams that transmitted the second lost packet. The communication device according to claim 4, characterized in that the second lost packet is a packet that has become the first lost packet in all of the one or more substreams of the stream. The communication device according to claim 4 or 5, characterized in that the transmission means holds a target value for the occurrence rate of the second lost packets for each of the multiple streams, and determines whether or not to change the number of the one or more substreams for each of the multiple streams by calculating the number of substreams required to make the occurrence rate of the second lost packets equal to or lower than the target value for each of the multiple streams based on the number of the first lost packets for each of the one or more substreams indicated by the reception status information. A communication device that receives a packet including data from a transmitting device via a network,
A plurality of streams, and a setting means for setting one or more sub-streams for each of the plurality of streams;
a first determination means for determining a first lost packet that is not normally received for each of the one or more sub-streams of the plurality of streams;
second determining means for determining, for each of the plurality of streams, a second lost packet that has become the first lost packet in all of the one or more substreams of the stream;
a notification means for notifying the transmitting device of information indicating the second lost packet in order to retransmit the second lost packet;
A communication device comprising: The communication device according to claim 7, characterized in that the notification means notifies the transmitting device of reception status information indicating the number of the first lost packets for each of the one or more substreams of the multiple streams. A program that, when executed by one or more processors of a device having one or more processors, causes the device to function as a communication device according to any one of claims 1 to 8.

Images