Transmission device, method, and program
By calculating the predicted delay and preferring the path with large delay to send packets, the receiving-side forwarding delay problem caused by the rapidity of the small delay path is solved, and more efficient communication is achieved.
Patent Information
- Application Number
- CN202280093926.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2022-03-22
- Filing Date
- 2022-12-23
- Publication Date
- 2024-12-03
AI Technical Summary
When a packet is distributed by load by using multiple communication paths, the path with small delay is increased by a problem that the forwarding delay on the receiving side increases.
By calculating the predicted delay and selecting a path with a larger delay when selecting the transmission path, a packet with a packet with a later arrival order than the predicted delay is selected as a packet sent from a plurality of communication paths to reduce the forwarding delay on the receiving side.
It effectively suppresses the forwarding delay on the receiving side, reduces the frequency of Head Of Line blocking, and improves communication efficiency.
Smart Images

Figure CN119072911A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to a sending device, a sending method, and a sending program.
[0002] CROSS-REFERENCE TO RELATED APPLICATIONS
[0003] This application claims the benefit of priority based on Japanese Patent Application No. 2022-045816 filed on March 22, 2022, the entire contents of which are incorporated into this specification by reference. Background Art
[0004] Conventionally, there is a technology that provides a large-capacity communication path that is difficult to realize with a single path by distributing the communication load among communication paths.
[0005] However, although the bandwidth is expanded by multiplexing, there is a problem that the delay increases by multiplexing paths with different delays because the path with the largest delay among the multiple paths becomes the transmission rate limit.
[0006] In order to avoid the load distribution effect from being reduced due to the increase of the delay difference between such paths, for example, the technology described in Patent Document 1 is disclosed. In the technology disclosed in Patent Document 1, the predetermined transmission path of each packet to be sent is determined based on the estimated value of the delay of each path and the line speed, so that the order of the predicted arrival time to the receiving node is equal to the order of the arrival time at the sending node. Then, in the path 1 where the maximum delay allowed by each path is large, the packet that is not the head of the packet to be sent is sent. In this way, the delay difference between the paths can be suppressed and an appropriate load can be applied to each path, thereby optimizing the communication efficiency.
[0007] Patent Document 1: Japanese Patent No. 4488256
[0008] In the above-mentioned prior art, since the larger the allowable delay of path 1 is, the longer it takes for the last packet transmitted in path 1 to arrive, the more likely it is that the communication state of paths other than path 1 will change during this period. Therefore, when the line speed of paths other than path 1 is further increased, there is a problem that the forwarding delay of packets downstream of the receiving node occurs due to the waiting time for resolving the arrival order of packets at the receiving node. Summary of the invention
[0009] The present disclosure has been made in view of the above-mentioned point, and its object is to suppress the transfer delay on the receiving side even when the path with less delay is faster than predicted when packets are transmitted by load distribution using multiple communication paths.
[0010] The sending device involved in the present disclosure is a sending device that sends multiple packets via multiple communication paths by load distribution, and is composed of the following components: a calculation unit, which calculates the predicted delay based on the data amount of the packet maintained by the maintenance unit and the line speed of each of the multiple communication paths, wherein the maintenance unit maintains the multiple packets arriving from the upper layer in the order of arrival, and the predicted delay refers to the time required for the packet to reach the receiving device when the packet is sent using each of the multiple communication paths; a selection unit, which, when selecting a communication path for sending each of the multiple packets so that each of the multiple packets is received by the receiving device in the order of arrival, selects a packet whose arrival order is later than the packet selected based on the predicted delay as a packet sent by the communication path with the largest delay in the predicted delay when sending packets of the same data amount from the multiple communication paths; and a sending unit, which sends each of the multiple packets via each of the communication paths selected by the selection unit.
[0011] In addition, the sending method involved in the present disclosure is a method in which a computer executes processing including the following steps, in a sending device that sends multiple packets through multiple communication paths by load distribution, a calculation unit calculates a predicted delay based on the data amount of the packet maintained by a maintenance unit and the line speed of each of the multiple communication paths, wherein the maintenance unit maintains the multiple packets arriving from the upper layer in the order of arrival, and the predicted delay refers to the time required for the packet to reach the receiving device when the packet is sent using each of the multiple communication paths. When selecting a communication path for sending each of the multiple packets so that each of the multiple packets is received by the receiving device in the order of arrival, the selection unit selects a packet whose arrival order is later than the packet selected based on the predicted delay as a packet sent by a communication path with the largest delay having the largest predicted delay when sending packets of the same data amount from the multiple communication paths, and the sending unit sends each of the multiple packets via each of the communication paths selected by the selection unit.
[0012] In addition, the sending program involved in the present disclosure is a program for making a computer function as a sending device, which sends multiple packets through multiple communication paths by load distribution, and includes: a calculation unit, which calculates the predicted delay based on the data amount of the packet maintained by the maintenance unit and the line speed of each of the multiple communication paths, wherein the maintenance unit maintains the multiple packets arriving from the upper layer in the order of arrival, and the predicted delay refers to the time required for the packet to reach the receiving device when the packet is sent using each of the multiple communication paths; a selection unit, when selecting a communication path for sending each of the multiple packets so that each of the multiple packets is received by the receiving device in the order of arrival, selects a packet whose arrival order is later than the packet selected based on the predicted delay as the packet sent by the communication path with the largest delay when sending packets of the same data amount from the multiple communication paths; and a sending unit, which sends each of the multiple packets via each of the communication paths selected by the selection unit.
[0013] According to the transmission device, transmission method, and transmission program involved in the present disclosure, when packets are transmitted by load distribution using multiple communication paths, even if the path with less delay is faster than predicted, the transfer delay on the receiving side can be suppressed. BRIEF DESCRIPTION OF THE DRAWINGS
[0014] The above-mentioned object and other objects, features and advantages of the present disclosure will become more apparent through the following detailed description with reference to the accompanying drawings.
[0015] Figure 1 It is a block diagram showing a schematic structure of a communication system.
[0016] Figure 2 It is a block diagram showing the hardware structure of the transmitting device.
[0017] Figure 3 It is a functional block diagram of the sending device.
[0018] Figure 4 This is a diagram for explaining conventional transmission when the communication path is accelerated.
[0019] Figure 5 This is a diagram for explaining transmission in this embodiment when the communication path is accelerated.
[0020] Figure 6 This is a diagram for explaining an example of a method of setting a numerical value of a margin.
[0021] Figure 7 This is a diagram for explaining an example of a method of setting a numerical value of a margin when the margin is defined by quantity.
[0022] Figure 8 This is a diagram for explaining an example of a method of setting a numerical value of a margin when there are three or more communication paths.
[0023] Fig. 9 This is a timing chart showing an example of the timing of transmission and reception.
[0024] Fig.10 This is a flowchart showing an example of the transmission process. DETAILED DESCRIPTION
[0025] Hereinafter, an example of an embodiment of the present disclosure will be described with reference to the drawings.
[0026] like Figure 1 As shown, the communication system 100 involved in this embodiment includes a transmitting device 10 and a receiving device 20. The transmitting device 10 and the receiving device 20 are connected via a communication line 30 to transmit and receive data to each other. The communication line 30 is composed of multiplexing a plurality of communication paths. In addition, the number of each transmitting device 10 and receiving device 20 included in the communication system 100 is not limited to Figure 1 For example, the sending device 10 and the receiving device 20 may each include more than two devices.
[0027] exist Figure 2 The hardware structure of the sending device 10 is shown in FIG. Figure 2 As shown, the transmission device 10 includes a CPU (Central Processing Unit) 42, a memory 44, a storage device 46, an input device 48, an output device 50, a storage medium reader 52, and a communication I / F (Interface) 54. Each component is connected to each other via a bus 56 so as to be communicable.
[0028] The storage device 46 stores a transmission program for executing the transmission processing described later. The CPU 42 is a central processing unit that executes various programs or controls various structures. That is, the CPU 42 reads the program from the storage device 46 and executes the program using the memory 44 as a work area. The CPU 42 controls the above-mentioned structures and performs various calculations according to the program stored in the storage device 46.
[0029] The memory 44 is composed of a RAM (Random Access Memory) and temporarily stores programs and data as a work area. A part of the memory 44 functions as a transmission queue (holding unit). The storage device 46 is composed of a ROM (Read Only Memory), a HDD (Hard Disk Drive), an SSD (Solid State Drive), etc., and stores various programs including an operating system and various data.
[0030] The input device 48 is a device for performing various inputs, such as a keyboard and a mouse. The output device 50 is a device for outputting various information, such as a display and a printer. By using a touch panel display as the output device 50, it can also function as the input device 48.
[0031] The storage medium reading device 52 reads data stored in various storage media such as CD (Compact Disc)-ROM, DVD (Digital Versatile Disc)-ROM, Blu-ray Disc, USB (Universal Serial Bus) memory, etc., writes data to the storage medium, etc. The communication I / F 54 is an interface for communicating with other devices including the receiving device 20, for example, using standards such as Ethernet (registered trademark), FDDI or Wi-Fi (registered trademark).
[0032] The hardware structure of the receiving device 20 is similar to Figure 2 The hardware configuration of the transmitter 10 shown is substantially the same, and therefore the description thereof is omitted. In addition, the storage device 46 of the receiver 20 stores a reception program for executing each process of the receiver 20 .
[0033] Next, refer to Figure 3 , the functional structure of the transmitting device 10 involved in this embodiment is described. Figure 3 As shown in FIG. 1 , the transmitting device 10 includes a calculating unit 12, a selecting unit 14, and a transmitting unit 16. Each functional unit is composed of Figure 2 It is implemented by the CPU42 shown.
[0034] First, in the transmission queue of the transmission device 10, the data to be transmitted that arrives from the upper layer of the network layer is stored in each packet. The multiple packets are maintained in the transmission queue in the order of arrival. The packets maintained in the transmission queue are assigned sequence numbers (hereinafter, represented by "[seq]"). In addition, the transmission queue is an example of a holding unit disclosed in the present invention. A part of the memory 44 of the transmission device 10 functions as a transmission queue.
[0035] The calculation unit 12 calculates a predicted delay, which is the time required for the packet held in the transmission queue to be transmitted from the transmission device 10 and reach the reception device 20. The calculation unit 12 calculates the predicted delay based on the data amount of the packet held in the transmission queue and the line speed of each of the plurality of communication paths.
[0036] The selection unit 14 selects a communication path from among the multiple communication paths for transmitting each of the multiple packets so that each of the multiple packets held in the transmission queue is received by the receiving device 20 in the order of arrival. At this time, when a communication path having the largest value of the predicted delay calculated by the calculation unit 12 (hereinafter referred to as a "communication path with large delay") is selected, the selection unit 14 selects a packet having an arrival order later than the packet selected for transmission based on the predicted delay calculated by the calculation unit 12.
[0037] Here, use Figure 4 , a method of selecting packets in a comparative example compared with the present embodiment will be described. The comparative example is a method of selecting a communication path for transmitting each packet based only on a predicted delay so that the packets arrive at a receiving device in sequence number order.
[0038] Figure 4 An example of sequence numbers of packets selected to be sent in multiple paths based on predicted delay is shown. Here, t in the figure represents time, and v represents the line speed of the communication path. The figure shows that time t increases to the left and speed v increases upward. In addition, facing the paper, the left side is the sending side and the right side is the receiving side. The three-dimensional blocks represent packets that arrive at the sending queue and are arranged in sequence number order, and the two-dimensional blocks represent selected packets after being assigned to multiple paths. In addition, the horizontal width of the two-dimensional block represents the time required to send the packet corresponding to the block, and the vertical width of the block represents the speed of sending the packet corresponding to the block. The multiple planar blocks also indicate that the closer to the receiving side, the earlier the packet arrives. In addition, path 1 in the figure represents a communication path with a large delay, and in contrast, path 2 represents a communication path that is not a communication path with a large delay (hereinafter referred to as a "communication path with a small delay").
[0039] Reference Figure 4 , the selection and sending of the communication path for sending each packet in the comparative example are specifically described.
[0040] First, packets are stored in a sending queue in the order of sequence numbers. Next, a path for sending packets is selected from multiple paths. More specifically, Figure 4As shown in (a), when the line speed of path 1, which is a path with large delay, is v, and the line speed of path 2, which is a path with small delay, is 2v, if the data amount of each packet is the same, the predicted delay of one packet in path 1 is 2t, and the predicted delay of one packet in path 2 is t. Based on the predicted delay of each path, for example, it is selected to send the packet [seq1, seq2] from path 2 and send the packet [seq3] from path 1. Next, it is selected to send the packet [seq4, seq5] from path 2 and send the packet [seq6] from path 1. As a result, the packet is sent in a manner that the transmission and reception of [seq3] from path 1 is terminated at the timing when the packet [seq1, seq2] is sent from path 2 and the reception is terminated. Next, the packet is sent in a manner that the transmission and reception of [seq6] from path 1 is terminated at the timing when the packet [seq4, seq5] is sent from path 2 and the reception is terminated.
[0041] As described above, in the comparative example, a path for transmitting each packet is selected based on the calculated predicted delay so as to suppress waste of waiting time, that is, so that the packets arrive at the receiving device in sequence number order.
[0042] However, if Figure 4 As shown in (b), when actually sending packets, the line speed of path 2 may be higher than the time when the predicted delay is calculated. Figure 4 In (b), the example of increasing the line speed of path 2 from 2v to 4v is shown. That is, in path 2, the predicted delay of one packet is 1 / 2t. In this case, if Figure 4 (a) is sent in the order of the sequence numbers of the selected packets in the same manner as in (a), then the packet with a later sequence number will arrive in path 2 before the packet that has not arrived in path 1. Therefore, when forwarding packets to the downstream of the receiving device in the order of sequence numbers, the packet that arrives first in path 2 needs to wait for the arrival of the packet with a sequence number that is earlier than the packet. This situation is called Head Of Line blocking (hereinafter referred to as "HoLB"). As a result of the occurrence of HoLB, in the receiving device, a forwarding delay from the receiving device to the downstream is generated due to the sequence resolution waiting time.
[0043] exist Figure 4In the example (b), before the sending of the packet [seq3] in path 1 is completed, the packets [seq1, seq2, seq4, seq5] can be sent from path 2. However, before the sending of the packet [seq3] in path 1 is completed, the packets [seq4, seq5] sent in path 2 must wait for forwarding. In other words, HoLB occurs in the packets [seq4, seq5]. Similarly, HoLB also occurs in the packets after [seq7] in path 2. As described above, in the comparative example, there is a problem that due to the frequent occurrence of HoLB and the increase in the waiting time caused by HoLB, a forwarding delay to the downstream occurs when the line speed of path 2 is increased. That is, when only the arrival order based on the predicted delay is considered, a forwarding delay occurs when the communication path with a small delay is accelerated.
[0044] Therefore, when a communication path with a large delay is selected, the selection unit 14 of the present embodiment selects a packet whose arrival order is later than the packet selected for transmission based on the predicted delay calculated by the calculation unit 12. Specifically, Figure 5 As shown, in path 1 which is a communication path having a large delay, selection unit 14 selects a packet having a sequence number having a margin with respect to the sequence number of the packet selected based on the predicted delay calculated by calculation unit 12 .
[0045] For example, in Figure 4 In the same example as (a), as the packet sent from path 1, the selection unit 14 selects the packet [seq6] based on the predicted delay as follows Figure 5 As shown in (a) of FIG. 1 , there is a margin of +2. Thus, when the packet [seq6] is selected in the selection based on the predicted delay, the selection unit 14 selects the packet [seq8] as the packet to be transmitted from path 1. Similarly, the selection unit 14 selects [seq9] selected based on the predicted delay in Figure 5 (a) has a margin of +6, so the packet of [seq15] is selected as the packet sent from path 1.
[0046] In this way, the selection unit 14 selects the packet with the sequence number having the margin and transmits it from the path with the large delay. Figure 5 As shown in (b), although HoLB occurs initially, the impact on subsequent packets of path 2 becomes small. Figure 4 Compared with case (b), the downstream forwarding delay caused by HoLB is alleviated.
[0047] Next, a method of setting the margin in the selection unit 14 will be described.
[0048] Reference Figure 6, a specific example of the remainder when defined by natural numbers is described. Figure 6 In FIG. 1 , packets transmitted from path 1 are represented by P1, P2, and P3 in the order of transmission from earliest to latest.
[0049] First, as variables for setting the margin, for example, the following three variables are defined.
[0050] As the first variable, the predicted delay difference i is explained. The so-called predicted delay difference i is the time calculated by the difference between the predicted delay t1 of each packet sent from a communication path with a large delay and the predicted delay t2 of the final packet sent from a communication path with a small delay within the range of packets that can be sent within the allowable delay described later. Figure 6 In FIG. 1 , the blocks with slashes represent packets sent within the allowable delay of each communication path, and the blocks with dots represent packets sent after the packets in the blocks with slashes are sent. Figure 4 , Figure 5 as well as Figure 8 The same is true in . In addition, the timing of this transmission will be described later.
[0051] exist Figure 6 In the example of , the number of packets that can be sent within the allowed delay is three on both path 1 and path 2. In addition, the predicted delay of each packet is equal to the time from the time when the packet is sent from the transmitter 10 to the time when the packet arrives at the receiver 20, assuming that the delay is 0. For example, the predicted delay (P3) for packet P3 is equal to Figure 6 The width of the solid double arrow in . In this case, Figure 6 As shown by the dotted double arrow in , the predicted delay difference (P3) of packet P3 is the time obtained by subtracting the predicted delay of three packets on path 2 from the predicted delay (P3) of packet P3 on path 1. In addition, the predicted delay difference (P2) of packet P2 is also equivalent to Figure 6 The width of the dotted double arrow in . The predicted delay difference i is expressed by the following equation (1).
[0052] i[predicted delay difference] = t1[predicted delay of path 1] - t2[predicted delay of path 2] (1)
[0053] As a second variable, the offset j is described. The offset j is the amount of data obtained by multiplying the predicted delay difference i by the line rate v2 of the communication path with the smallest delay. The offset j is represented by the following equation (2).
[0054] j[offset] = i[predicted delay difference] × v2[line speed of path 2] (2)
[0055] As the third variable k, no-load skew k is described. No-load skew k is the amount of data obtained by multiplying the predicted delay t1 of a communication path with a large delay by the line speed v2 of a communication path with a small delay. No-load skew k is represented by the following equation (3).
[0056] k[offset at no load] = t1[predicted delay of path 1] × v2[line speed of path 2] (3)
[0057] Next, calculation of the margin using the above-mentioned variables i, j, and k will be described. Here, five calculation methods represented by the following equations (4) to (8) are exemplified.
[0058] In addition, the selection unit 14 obtains the variables i, j, and k by the above-mentioned equations (1) to (3) in equations (4) to (8), and sets an arbitrary value for the constant a. In addition, the selection unit 14 may set different values as the constant a in each equation. In addition, the selection unit 14 may dynamically set the constant a according to the number of packets held in the transmission queue so that the sequence number with the margin does not exceed the maximum value of the sequence numbers of the packets held in the transmission queue.
[0059] m1=v2[line speed of path 2]×a (4)
[0060] m2=k[offset when no load]×a (5)
[0061] m3=j[offset]×a (6)
[0062] m4 = t1 [predicted delay of path 1] × m2 (7)
[0063] m5=i[prediction delay difference]×m2 (8)
[0064] In addition, the selector 14 calculates the value of the margin m by using any one of the equations (4) to (8) when the predicted delay difference i is greater than a certain value, and may set the margin m to 0 when the predicted delay difference i is less than the certain value. When the deviation of the packet arrival order in the receiving device 20 is small enough to be ignored, priority is given to reducing the processing load and optimizing efficiency. The selector 14 converts the value of the margin m calculated by using any one of the equations (4) to (8) into the number of packets, and defines the margin as a natural number.
[0065] In addition, the effects of applying the margins calculated by equations (4) to (8) are as follows.
[0066] Equation (4): The delay reduction of the mitigation path 2 (e.g., Figure 6 As shown in (a1), the impact of the delay reduction from the time of delay 0 to the arrival of the first packet).
[0067] Formula (5): The bandwidth expansion of the mitigation path 2 (for example, Figure 6 As shown in (a2), the impact when the line speed of path 2 is increased compared to the predicted delay calculation.
[0068] Formula (6): Alleviates the influence of bandwidth expansion of path 2 after the final packet is received.
[0069] Formula (7): Alleviate the further expansion of the bandwidth of path 2 caused by the passage of time ( Figure 6 The impact of (a3)).
[0070] Formula (8): Alleviates the influence of further expansion of the bandwidth of path 2 due to the passage of time after the final packet is received.
[0071] Figure 5 The example is an example of adding the margin of the above formula (6) to the sequence numbers of the second and third packets sent from path 1. Specifically, between the difference between the predicted delay of each packet and the predicted delay of the final packet [seq4] sent from path 2, a value twice the number of packets expected to arrive from path 2 is set as the margin. This is equivalent to a setting that prevents HoL congestion even when the speed of path 2 is doubled after receiving the packet [seq4].
[0072] Next, we will explain the case where the margin is defined by the amount of data. The margin value itself can be calculated using any of the five calculation formulas above, but there are cases where the calculated value cannot be converted into the number of groups and the margin is defined using a natural number. For example, Figure 7 Therefore, the selection unit 14 defines the margin by the data amount and selects the packet of the sequence number based on the margin by the following method.
[0073] Specifically, the selection unit 14 calculates the total amount of data of each packet from the next packet selected as the packet for non-application of margin based on the predicted delay to the predetermined packet. Figure 7 In the example, the packet when the margin is not applied is [seq5], and the next packet is [seq6]. Next, the selection unit 14 selects a specified packet when the difference between the amount of data defined as the margin obtained by any of the above five calculation formulas and the above sum is within a specified range as the packet sent from the path with a large delay.
[0074] The selection unit 14 selects a predetermined group when the difference is within a predetermined range in any one of the following three modes.
[0075] 1. Select the largest sequence number for which the margin (data amount) is greater than the sum of the data amounts of each packet.
[0076] 2. Select the smallest sequence number for which the remainder (data amount) is ≤ the sum of the data amounts of each packet.
[0077] 3. Select the sequence number whose total data amount of each packet is closest to the margin (data amount).
[0078] For example, in Figure 7 In the case of mode 1, the selection unit 14 selects the group of [seq9], selects the group of [seq10] in the case of mode 2, and selects the group of [seq9] in the case of mode 3.
[0079] Next, a method of setting the margin when there are two or more communication paths with a small delay will be described.
[0080] When there are two or more communication paths with small delays, the selection unit 14 calculates the sum of the margins of each of the communication paths with small delays, and sets the sum as the final margin.
[0081] like Figure 8 As shown, for example, as packets transmitted by path 1, before applying the margin, the first packet [seq7], the second packet [seq14], and the third packet [seq21] are selected. Then, the margin of path 2 is calculated to be +2 for the second packet, and the margin of path 3 is calculated to be +6. In this case, the sum of the margins for the second packet of path 1 is 2+6, which is equal to 8, and the selection unit 14 selects the packet [seq22] obtained by adding 8 to [seq14] as the packet after the margin is applied.
[0082] The transmission section 16 transmits each of the plurality of packets held in the transmission queue via each of the communication paths selected by the selection section 14 .
[0083] The transmission unit 16 transmits packets within the set allowed delay. The allowed delay is the maximum value of the delay allowed in each path in order to achieve load balance among multiple communication paths. The larger the delay, the larger the allowed delay.
[0084] Specifically, when transmission starts, the transmission unit 16 of the transmission device 10 estimates Fig. 9 The no-load delay and line speed shown are Fig. 9 The transmission unit 16 then transmits packets of the window size selected to reflect the margin. The timing of the start of transmission is the same in both path 1 and path 2. Fig. 9 In (a), packets P1 to P5 are sent as packets of the window size. In addition, based on the sequence number selected by the selection unit 14, that is, the sequence number reflecting the surplus, the first packet sent from each path is represented as P1, the second packet sent is represented as P2, and so on. The sending unit 16 receives an ACK sent from the receiving device 20 at a specified timing, and the ACK is a positive response to the packet received by the receiving device 20. Fig. 9 In the path 1 shown in (a) of FIG. 1 , the example in which the receiving device 20 sends an ACK to the transmitting device 10 every time two packets are received is shown, but the timing of sending the ACK can be set appropriately. If the transmitting unit 16 receives the ACK, it detects that Fig. 9 In (a), the transmission and reception to P2 has been completed normally, and then the window size for sending is free for two packets (packets P6 to P7).
[0085] In addition, in the example of path 2 Fig. 9 Similarly, in (b) of FIG. 1 , the transmission unit 16 transmits packets at the timing of starting transmission and at the timing after receiving ACK. Fig. 9 In (b), the window size at the start of transmission is six packets, and the receiving device 20 sends an ACK to the transmitting device 10 every time it receives three packets. Therefore, since there is a free space of three packets at the subsequent transmission timing, the transmitting unit 16 sends three packets at a time at each subsequent transmission timing on the path 2. Fig. 9 In the figure, the asterisk indicates the sending timing of the packet.
[0086] The transmission unit 16 may select the packet pre-selected by the selection unit 14 from each path at each transmission timing described above, or may transmit the reselected packet at each transmission timing. In the latter case, specifically, each time the transmission unit 16 receives an ACK, it notifies the selection unit 14. The selection unit 14 reflects the margin based on the predicted delay at the time of receiving the notification, and reselects the packet to be transmitted on each path. In addition, the transmission unit 16 may determine the transmittable amount of the transmission timing after the ACK reception, instead of the free amount of the window, at each transmission timing by re-estimating the line speed at the time of the ACK reception.
[0087] When receiving a packet transmitted from the transmission device 10, the reception device 20 transmits ACK to the transmission device 10 at a predetermined timing. In addition, the reception device 20 transfers the packets received for each path to the downstream in the order of sequence numbers.
[0088] Next, the operation of the communication system 100 according to this embodiment will be described. Fig.10 The sending process shown. In addition, the sending process is an example of the sending method of the present disclosure.
[0089] Reference Fig.10 , the sending process is explained.
[0090] In step S10, when the calculation unit 12 detects that a packet has arrived at the transmission queue of the transmission device 10, it assigns consecutive sequence numbers to each packet in the order of arrival, starting from 1. Next, in step S12, the calculation unit 12 substitutes 1 into a variable n for specifying a sequence number.
[0091] Next, in step S14, the calculation unit 12 determines whether a path for transmitting a packet with sequence number n has not been selected. If the path has been selected (step S14: No), the process proceeds to step S16, and the variable n is incremented by 1, and the process returns to step S14. If the path has not been selected (step S14: Yes), the process proceeds to step S18.
[0092] In step S18 , the calculation unit 12 calculates the predicted delay for each path of the packet with sequence number n. The predicted delay is calculated based on the amount of data to be transmitted stored in the transmission queue and the estimated line speed of the communication line 30 .
[0093] Next, in step S20, the selection unit 14 determines whether the path with the earliest arrival prediction is a path with the largest delay among the multiple paths based on the predicted delay calculated by the calculation unit 12. If it is not a path with the largest delay (step S20: No), the process moves to step S26, and the selection unit 14 selects the path with the smallest delay as the path of the packet with sequence number n, and moves to step S28. On the other hand, if it is a path with the largest delay (step S20: Yes), the process moves to step S22, and the selection unit 14 sets the margin m. Next, in step S24, the selection unit 14 selects the path with the largest delay as the path of the packet with sequence number n+m, and moves to step S28.
[0094] In step S28, the selection unit 14 determines whether the selection of the paths for all packets in the transmission queue has been completed. If there are packets for which no path has been selected (step S28: No), the process returns to step S14. If the selection of the paths for all packets has been completed (step S28: Yes), the process moves to step S30, and the transmission unit 16 transmits each packet along the path selected by the selection unit 14, and the process returns to step S10.
[0095] Furthermore, in the above transmission processing, the case where packets are transmitted at the stage where the paths for all packets in the transmission queue are selected is described, but packets may be transmitted at the stage where the paths for a predetermined number of packets are selected.
[0096] As described above, according to the communication system involved in this embodiment, multiple packets are sent through multiple communication paths by load distribution. The sending device maintains multiple packets arriving from the upper layer in the order of arrival. Moreover, based on the data volume of the packet maintained by the sending queue and the line speed of each communication path in the multiple communication paths, the predicted delay is calculated, and the predicted delay is the time required for the packet to reach the receiving device when the packet is sent using each communication path in the multiple communication paths. In addition, the sending device selects the communication path for sending each of the multiple packets so that each of the multiple packets is received by the receiving device in the order of arrival at the sending queue. At this time, the sending device selects the packet whose arrival order is later than the packet selected based on the predicted delay as the packet sent by the communication path with the largest predicted delay when sending packets of the same amount of data from multiple communication paths. Moreover, the sending device sends each of the multiple packets via each communication path in the selected communication path. As a result, even if the line speed of the path other than the communication path with large delay is further increased, the frequency of Head Of Line blocking can be reduced, and the waiting time for resolving the arrival order of the packet to the receiving device can be reduced, so that the forwarding delay of the packet downstream of the receiving device can be suppressed.
[0097] The control unit and method described in the present disclosure may be implemented by a special-purpose computer, which constitutes a processor programmed to execute one or more functions embodied by a computer program. Alternatively, the device and method described in the present disclosure may be implemented by a special-purpose computer, which constitutes a processor composed of a special-purpose hardware logic circuit. Alternatively, the device and method described in the present disclosure may be implemented by one or more special-purpose computers, which constitute a combination of a processor that executes a computer program and one or more hardware logic circuits. In addition, the computer program may also be stored as an instruction executed by a computer in a computer-readable non-transient tangible recording medium.
[0098] The present disclosure is described based on the embodiments, but it is understood that the present disclosure is not limited to the embodiments and structures. The present disclosure also includes various modifications and modifications within the equivalent range. In addition, various combinations and methods, further including only one element, more than one, or less than one other combinations and methods are also included in the scope and thought range of the present disclosure.
[0099] The following are the additional items.
[0100] (Note 1)
[0101] A transmission device (10) for transmitting a plurality of packets via a plurality of communication paths by load distribution, comprising:
[0102] a calculation unit (12) for calculating a predicted delay based on the data volume of the packet held by the holding unit and the line speed of each of the plurality of communication paths, wherein the holding unit holds the plurality of packets arriving from the upper layer in the order of arrival, and the predicted delay is a time required for the packet to arrive at the receiving device when the packet is transmitted using each of the plurality of communication paths;
[0103] a selection unit (14) for selecting, when selecting a communication path for transmitting each of the plurality of packets so that each of the plurality of packets is received by the receiving device in the arrival order, selecting a packet whose arrival order is later than the packet selected based on the predicted delay as a packet to be transmitted through the communication path with the largest delay having the largest predicted delay when transmitting packets of the same amount of data from the plurality of communication paths; and
[0104] A transmitting unit (16) transmits each of the plurality of packets via each of the communication paths selected by the selecting unit.
[0105] (Note 2)
[0106] According to the sending device described in Appendix 1, wherein the above-mentioned selection unit uses the sequence numbers assigned to the packets maintained by the above-mentioned retention unit in the above-mentioned arrival order, and selects a packet with a sequence number that has a margin for the sequence number of the packet selected based on the above-mentioned predicted delay as the packet with a later arrival order.
[0107] (Note 3)
[0108] The transmission device according to Supplement 2, wherein the selection unit sets the margin based on the predicted delay of the packet that can be transmitted within the allowed delay, wherein the allowed delay is a value set to achieve a balance in load distribution among the plurality of communication paths.
[0109] (Note 4)
[0110] A sending device according to Note 2, wherein the selection unit sets the margin to a value corresponding to a predicted delay difference, wherein the predicted delay difference is a difference between a predicted delay of a packet sent from the communication path with large delay and a predicted delay of a packet sent from a communication path with small delay other than the communication path with large delay.
[0111] (Note 5)
[0112] The transmission device according to Supplementary Note 2, wherein the selection unit sets the margin to a value corresponding to a line speed of a communication path with small delay other than the communication path with large delay among the plurality of communication paths.
[0113] (Note 6)
[0114] The transmission device according to Supplementary Note 4, wherein the selection unit sets the margin to a value corresponding to an offset calculated by multiplying the predicted delay difference by a line speed of the communication path with the small delay.
[0115] (Note 7)
[0116] The transmission device according to any one of Supplementary Notes 2 to 6, wherein the selection unit defines the margin using a natural number.
[0117] (Note 8)
[0118] A sending device according to any one of Notes 2 to 6, wherein, when the data amount of each of the above-mentioned multiple groups is different, the above-mentioned selection unit defines the above-mentioned margin based on the data amount, and selects the above-mentioned specified group when the difference between the sum of the data amounts of each group from the next group of the group selected based on the above-mentioned predicted delay to the specified group and the data amount defined as the above-mentioned margin is within a specified range, as the above-mentioned group with a later arrival order.
[0119] (Note 9)
[0120] A sending device according to any one of Notes 2 to 8, wherein, when there are two or more communication paths with small delay other than the communication path with large delay, the selection unit sets the sum of the margins of each of the communication paths with small delay as the final margin.
[0121] (Note 10)
[0122] The transmission device according to any one of Supplementary Notes 1 to 9, wherein the transmission unit transmits the number of packets that can be transmitted within an allowed delay, wherein the allowed delay is a value set to achieve a balance in load distribution among the plurality of communication paths.
Claims
1. A transmission device, a transmission device (10) for transmitting a plurality of packets via a plurality of communication paths by load distribution, comprising: a calculation unit (12) for calculating a predicted delay based on the amount of data of the packet held by the holding unit and the line speed of each of the plurality of communication paths; in, The holding unit holds the plurality of packets arriving from the upper layer in order of arrival, the predicted delay being the time required for the packet to arrive at the receiving device when the packet is sent using each of the plurality of communication paths; a selection unit (14) for selecting, when selecting a communication path for transmitting each of the plurality of packets so that each of the plurality of packets is received by the receiving device in the arrival order, selecting a packet whose arrival order is later than the packet selected based on the predicted delay as a packet to be transmitted through the communication path with the largest delay having the largest predicted delay when transmitting packets of the same amount of data from the plurality of communication paths; and A transmitting unit (16) transmits each of the plurality of packets via each of the communication paths selected by the selecting unit.
2. The transmitting device according to claim 1, in, The selection unit selects a packet having a sequence number having a margin with respect to the sequence number of the packet selected based on the predicted delay as the packet with the later arrival order, using the sequence numbers assigned to the packets held by the holding unit in the arrival order.
3. The transmitting device according to claim 2, in, The selection unit sets the margin based on the predicted delay of the packet that can be transmitted within an allowable delay, wherein the allowable delay is a value set to achieve a balance in load distribution among the plurality of communication paths.
4. The transmitting device according to claim 2, in, The selection unit sets the margin to a value corresponding to a predicted delay difference between a predicted delay of a packet sent from the communication path with large delay and a predicted delay of a packet sent from a communication path with small delay other than the communication path with large delay.
5. The transmitting device according to claim 2, in, The selection unit sets the margin to a value corresponding to a line speed of a communication path with a small delay other than the communication path with a large delay among the plurality of communication paths.
6. The transmitting device according to claim 4, in, The selection unit sets the margin to a value corresponding to an offset calculated by multiplying the predicted delay difference by a line speed of the communication path with the smaller delay.
7. The transmitting device according to any one of claims 2 to 6, in, The selection unit defines the margin using a natural number.
8. The transmitting device according to any one of claims 2 to 6, in, When the data amount of each of the above-mentioned multiple groups is different, the above-mentioned selection unit defines the above-mentioned margin according to the data amount, and selects the above-mentioned specified group when the difference between the sum of the data amounts of each group from the next group of the group selected based on the above-mentioned predicted delay to the specified group and the data amount defined as the above-mentioned margin is within a specified range, as the group with the later arrival order.
9. The transmitting device according to any one of claims 2 to 6, in, When there are two or more communication paths with small delays other than the communication path with large delays, the selection unit sets the sum of margins for each of the communication paths with small delays as the final margin.
10. The transmitting device according to claim 1, in, The transmission unit transmits the number of packets that can be transmitted within an allowed delay, wherein the allowed delay is a value set to achieve a balance in load distribution among the plurality of communication paths.
11. A sending method, wherein a computer executes a process comprising the following steps: In a transmission device that transmits a plurality of packets by load distribution via a plurality of communication paths, the calculation unit calculates the predicted delay based on the data amount of the packet held by the holding unit and the line speed of each of the plurality of communication paths, in, The holding unit holds the plurality of packets arriving from the upper layer in order of arrival, the predicted delay being the time required for the packet to arrive at the receiving device when the packet is sent using each of the plurality of communication paths. The selection unit selects a communication path for sending each of the plurality of packets so that each of the plurality of packets is received by the receiving device in the arrival order, and selects a packet whose arrival order is later than the packet selected based on the predicted delay as a packet to be sent through the communication path with the largest delay having the largest predicted delay when packets of the same amount of data are sent from the plurality of communication paths. The transmission unit transmits each of the plurality of packets via each of the communication paths selected by the selection unit.
12. A sending program for causing a computer to function as a sending device. The sending device sends a plurality of packets via a plurality of communication paths by load distribution, comprising: a calculation unit that calculates a predicted delay based on the data volume of the packet held by the holding unit and a line speed of each of the plurality of communication paths, in, The holding unit holds the plurality of packets arriving from the upper layer in order of arrival, the predicted delay being the time required for the packet to arrive at the receiving device when the packet is sent using each of the plurality of communication paths; A selection unit, when selecting a communication path for transmitting each of the plurality of packets so that each of the plurality of packets is received by the receiving device in the arrival order, selects a packet whose arrival order is later than a packet selected based on the predicted delay as a packet to be transmitted through a communication path with a large delay having the largest predicted delay when packets of the same amount of data are transmitted from the plurality of communication paths; and The transmitting unit transmits each of the plurality of packets via each of the communication paths selected by the selecting unit.
Citation Information
Patent Citations
Photographing system
JP2022045816A
