JP2018148494A - Packet aggregation device and transmission processing program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a package aggregation device which excellently reduce a processing load of communication between a client and a server.SOLUTION: A coupling part 13 aggregate packets that a first reception part 10 receives up to a certain maximum packet size to generate aggregated packets. A first transmission part 14 transmits the aggregated packets that the aggregation part generates to a second communication device 5. A communication destination acquisition part 18 acquires processing load information as information representing a load on the second communication device based upon a response that a second reception part 15 receives. A communication path information acquisition part 19 acquires communication path information as information representing a load on a network based upon the response that the second reception part receives. An aggregation parameter determination part 20 determines at least one of a maximum packet size and an aggregation interval based upon at least one of the processing load information that the communication destination information acquisition part acquires and the communication path information that the communication path information acquisition part acquires.SELECTED DRAWING: Figure 2

Description

  Embodiments described herein relate generally to a packet aggregation device and a transmission processing program.

In a client-server type control system, if access from a large number of clients to a single server is concentrated at the same time, the network load increases and it may be difficult to perform real-time control. In particular, when a large amount of small-sized packets are processed, the processing load on the network interface of the server increases. For this reason, congestion may occur depending on the congestion status of the network and the load status of the communication destination.

Japanese Patent No. 3490000

The problem to be solved by the present invention is to provide a packet aggregation device and a transmission processing program that are excellent in reducing the communication processing load between a client and a server.

A packet aggregation device that aggregates a plurality of packets received from a first communication device and transmits the packets to a second communication device via a network, wherein the packet aggregation device includes a first reception unit, an aggregation unit, and a first transmission Unit, a second receiving unit, a communication destination information acquisition unit, a communication path information acquisition unit, and an aggregation parameter determination unit. The first receiving unit receives a packet from the first communication device. The aggregating unit aggregates the packets received by the first receiving unit at a certain aggregation interval until a certain maximum packet size is reached, and generates an aggregated packet. The first transmission unit transmits the aggregate packet generated by the aggregation unit to the second communication device. The second receiving unit receives a response to the aggregate packet transmitted from the first transmitting unit from the second communication device. The communication destination information acquisition unit acquires processing load information, which is information indicating the load of the second communication device, based on the response received by the second reception unit. The communication path information acquisition unit acquires communication path information, which is information indicating the load on the network, based on the response received by the second reception unit. The aggregate parameter determination unit
Based on at least one of the processing load information acquired by the communication destination information acquisition unit and the communication path information acquired by the communication path information acquisition unit, at least one of the maximum packet size and the aggregation interval. Either one is determined.

1 is a diagram illustrating an example of a device configuration of a transmission processing system according to a first embodiment. FIG. 3 is a diagram illustrating an example of a functional configuration of the packet aggregation device according to the first embodiment. The figure which shows an example of the flow of the aggregation process when the packet aggregation apparatus which concerns on Embodiment 1 determines a parameter. The figure which shows an example of the flow of the process which the packet aggregation apparatus which concerns on Embodiment 1 aggregates a packet. The figure which shows an example of the flow of a process when the packet aggregation apparatus which concerns on Embodiment 1 receives an aggregation packet from a 2nd communication apparatus. The figure which shows an example of the flow of a process when the packet aggregation apparatus 4 which concerns on Embodiment 1 receives a packet from the 1st communication apparatus 3. FIG. FIG. 6 is a diagram illustrating an example of a functional configuration of a packet aggregation device 4 according to the second embodiment. The figure which shows an example of the reception time calculation method of the multiplexed aggregation packet which concerns on Embodiment 2. FIG. The figure which shows an example of the flow of the band estimation and parameter determination process which the packet aggregation apparatus 4 which concerns on Embodiment 2 performs. FIG. 9 is a diagram illustrating an example of a device configuration of a transmission processing system according to a third embodiment. FIG. 9 is a diagram illustrating an example of a functional configuration of a packet aggregation device according to a third embodiment. The figure which shows an example of the flow of the band prediction which the packet aggregation apparatus which concerns on Embodiment 3 performs, and the calculation process of a parameter. The figure which shows an example of the apparatus structure which adapted the transmission processing system which concerns on Embodiment 4 to the station service system.

Hereinafter, a transmission processing system, a packet aggregation device, and a transmission processing program according to each embodiment will be described with reference to the drawings.

(Embodiment 1)
FIG. 1 is a diagram illustrating an example of a device configuration of a transmission processing system 100 according to the first embodiment.
The apparatus configuration of the transmission processing system 100 according to the first embodiment will be described with reference to FIG.

The transmission processing system 100 includes first communication devices 3a and 3b, a packet aggregation device 4, and a second communication device 5. For the sake of easy understanding, the first communication device 3 is abbreviated as the first communication device 3 when it is not necessary to distinguish between the first communication device 3a and the first communication device 3b.

The first communication device 3 and the packet aggregation device 4 are connected via the network 1, and the second communication device 5 and the packet aggregation device 4 are connected via the network 2, and can communicate with each other.

Although two first communication devices 3 are illustrated in FIG. 1, the present invention is not limited to this, and a plurality of first communication devices 3 may be connected via the network 1.

The network 1 and the network 2 may be local lines such as a LAN, for example, and may use a wide-area wired communication line such as the Internet or a dedicated line, or a wide-area wireless communication line such as 3G.

The packet aggregating device 4 combines a plurality of packets received from the first communication device 3 into one packet (hereinafter, a packet in which a plurality of packets are combined into one is called an aggregated packet),
It transmits to the 2nd communication apparatus 5, changing the communication amount per time flexibly. For example, the aggregate packet includes a header and individual packets, and the header includes information such as the number of packets and an identification number of the aggregate packet. In each packet, the identification number of the first communication device 3,
Contains information such as packet size and packet body.

Further, the packet aggregation device 4 divides the aggregated packet received from the second communication device 5 into individual packets and transmits them to the first communication device 3. For example, the packet aggregation device 4 removes the header from the aggregate packet and divides it into individual packets.

The second communication device 5 divides the aggregated packet received from the packet aggregating device 4 into individual packets, and performs necessary processing such as computation on the individual packets. Furthermore, the second communication device 5
As a response to the aggregated packet received from the packet aggregating apparatus 4, the processed aggregated packets are aggregated again and transmitted to the packet aggregating apparatus 4.

FIG. 2 is a diagram illustrating an example of a functional configuration of the packet aggregation device 4 according to the first embodiment. FIG.
Will be used to describe the functional configuration of the packet aggregation device 4 according to the first embodiment. The packet aggregating apparatus 4 includes a first receiving unit 10, an aggregation buffer 11, a cycle management unit 12, and an aggregation unit 1.
3, the first transmission unit 14, the second reception unit 15, the division unit 16, the second transmission unit 17, the communication destination information acquisition unit 18, the communication path information acquisition unit 19, and the aggregation parameter determination unit 20. And comprising.

The first receiving unit 10 receives a packet from the first communication device 3 and stores the received packet in the aggregation buffer 11.

The aggregation buffer 11 stores the packets in the order in which the first receiver 10 receives the packets. The aggregation buffer 11 may have a list structure in which data is sequentially recorded, or a ring buffer structure in which data is cyclically recorded.

The period management unit 12 extracts packets stored in the aggregation buffer 11 for each period (hereinafter referred to as an aggregation interval). The cycle interval at which the cycle management unit 12 extracts packets and the amount of packets are determined by an aggregation parameter determination unit described later.

The aggregating unit 13 aggregates the packets extracted by the cycle management unit 12 into one packet to generate an aggregated packet.

  The first transmission unit 14 transmits the aggregate packet generated by the aggregation unit 13 to the second communication device 5.

The second reception unit 15 receives from the second communication device 5 an aggregate packet that is a response to the aggregate packet transmitted by the first transmission unit.

  The dividing unit 16 divides the aggregated packet received by the second receiving unit 15 into individual packets.

The second transmission unit 17 transmits the individual packets divided by the division unit 16 to the corresponding first communication device 3.
Send to each.

The communication destination information acquisition unit 18 acquires processing load information that is information indicating the load of the second communication device 5 that is the communication destination. For example, the communication destination information acquisition unit 18 has a function of acquiring, as one of the processing load information, the number of reception buffer overflows that is an index of the processing load in the second communication device 5. The reception buffer overflow indicates a state in which packets exceeding the allowable upper limit are transmitted to the aggregation buffer of the second communication device 5 (not shown) and packets cannot be stored in the aggregation buffer of the second communication device 5.

As a method for obtaining the number of buffer overflows, for example, the second communication device 5 side counts the number of reception buffer overflows, adds information on the number of reception buffer overflows to the aggregated packet, and transmits it to the packet aggregating device 4 for second reception. The unit 15 may have a function of notifying the communication destination information acquisition unit 18.

Note that the information acquired by the communication destination information acquisition unit 18 is not necessarily the number of reception buffer overflows. For example, the CPU load state of the second communication device 5 (not shown), the remaining capacity of the reception buffer, and the round-trip time of communication ( Any value that varies depending on the processing load of the second communication device 5, such as a round trip time), may be used.

The communication path information acquisition unit 19 acquires communication path information that is information indicating the load on the network 2. For example, the communication path information acquisition unit 19 has a function of counting the number of packet drops on the communication path. When the number of packets that exceed the limit of the processing capability of a device such as a router on the network 2 arrives, the packets may be discarded (dropped). That is, the packet drop count is an index indicating that the load on the network 2 is large when the packet drop count is large.

The aggregation parameter determination unit 20 determines the maximum packet size and the aggregation interval at the time of packet aggregation from the information acquired by the communication destination information acquisition unit 18 and the communication path information acquisition unit 19.

For example, if the maximum packet size per aggregated packet is reduced, the amount of data transmitted per time is reduced. Further, if the aggregation interval is increased, the amount of packets transmitted per time is reduced. That is, the packet aggregating apparatus 4 flexibly changes the amount of data transmitted per time and the number of aggregated packets depending on the load status of the communication destination and information on the communication path.

For example, when the number of reception buffer overflows exceeds a threshold, the aggregation parameter determination unit 20 makes the maximum packet size smaller than the set value. On the other hand, if the number of reception buffer overflows is less than the threshold, the maximum packet size is set larger than the set value. Also,
The size of the maximum packet size changed by the aggregation parameter determination unit 20 may be a fixed value or 10% of the set value.

When the aggregation interval is constant, the amount of data arriving per hour can be changed according to the load status of the communication destination by the above method, and processing congestion can be reduced.

Further, the aggregation parameter determination unit 20 may control the aggregation interval instead of the maximum packet size from the information acquired by the communication destination information acquisition unit 18.

For example, the aggregation interval is lengthened if the number of reception buffer overflows is equal to or greater than the threshold, and the aggregation interval is shortened if the number of reception buffer overflows is equal to or less than the threshold. If the aggregation interval is lengthened, the number of packets included in the aggregate packet generally increases, so that the total number of packets flowing through the network 2 can be reduced. As a result, it is possible to reduce the load on the communication destination and the load on devices on the network such as routers and switching hubs.

Further, the aggregate parameter determination unit 20 may control each parameter by information obtained from the communication path information acquisition unit 19.

For example, the communication path information acquisition unit 19 is provided with a function of counting the number of packet drops on the communication path. For example, a method using a confirmation packet may be used to count the number of packet drops. The confirmation packet is a packet that the second communication device 5 transmits to the packet aggregation device 4 in order to confirm whether or not the aggregation packet transmitted by the packet aggregation device 4 has reached the second communication device 5. That is, it is possible to determine whether or not the packet has been dropped by using the confirmation packet, and the communication path information acquisition unit 19 may count this number of times.

A packet drop may occur when the number of packets that exceed the limit of the processing capability of a device such as a router on the network 2 arrives. Therefore, the aggregation parameter determination unit 20 increases the aggregation interval if the number of packet drops is equal to or greater than the threshold, and decreases the aggregation interval if it is equal to or less than the threshold.

As described above, if the aggregation interval is increased, the total number of packets flowing through the network 2 is reduced.
The load on network equipment can be reduced. Another factor of packet drop is packet data corruption when the quality of the communication path is low, that is, CRC (Cyclic R).
(edundancy Check) error is also considered.

Therefore, if the number of packet drops is equal to or greater than the threshold value, the maximum packet size may be set smaller than the set value, and if the number is equal to or greater than the threshold value, the maximum packet size may be set larger than the set value. In general, the smaller the packet size, the lower the packet drop rate. Therefore, by reducing the maximum packet size and lowering the packet drop rate, it is possible to avoid data retransmission and use the network efficiently. Further, the maximum packet size changed by the aggregation parameter determination unit 20 may be a fixed value, or 10 which is a set value.
It is good as%.

  Next, the flow of aggregation processing according to the first embodiment will be described with reference to FIGS.

FIG. 3 is a diagram illustrating an example of a flow of aggregation processing when the packet aggregation device 4 according to the first embodiment determines parameters. The flow of the aggregation process when determining parameters will be described with reference to FIG.

First, the communication destination information acquisition unit 18 acquires load information of the second communication device 5 such as the number of reception buffer overflows (S101).

Next, the aggregate parameter determination unit 20 determines whether or not the processing load exceeds the threshold based on the processing load information acquired by the communication destination information acquisition unit 18 (S102). When the load exceeds the threshold (S102, Yes), the aggregation parameter determination unit 20 decreases the upper limit of the packet size of the aggregation packet (S103).

On the other hand, when the processing load is equal to or less than the threshold (S102, No), the upper limit of the packet size of the aggregate packet is increased (S104).

Subsequently, the communication path information acquisition unit 19 acquires communication path information such as the number of packet drops (
S105).

The aggregation parameter determination unit 20 increases the aggregation interval (S107) when the communication path information acquired by the communication path information acquisition unit 19 exceeds the threshold (S106, Yes), and when the communication path information is equal to or less than the threshold (S
106, No), the aggregation interval is shortened (S108).

In FIG. 3, the maximum packet size is controlled from S101 to S105, and the aggregation interval is controlled in the processing flow after S105, but only one of them may be controlled.

In addition, in the flow of maximum packet size control in FIG. 3, the same threshold value is used, and if the load exceeds the threshold value, the maximum packet size is made smaller than the set value, and if it is less than the threshold value, the maximum packet size is set. Although it is larger than the set value, two threshold values are provided, and if the threshold value is 1 or less, the aggregation interval is made smaller than the set value, and if the threshold value is 2 or more, the aggregation interval is made larger than the set value. It's okay. Similarly, regarding the control of the aggregation interval, the same control may be performed by providing two threshold values.

In the flow of FIG. 3, the maximum packet size is controlled according to the load status of the second communication device 5, but the aggregation interval may be controlled according to the load status of the second communication device 5. In that case, when the load exceeds the threshold, the aggregation interval is made longer than the set value to increase the number of packets included in one aggregated packet, and when the load is less than the threshold, the aggregation interval is made shorter than the set value. , Improve real-time communication.

If the load exceeds the threshold, the maximum packet size is made smaller than the set value and the aggregation interval is made longer than the set value. If the load is less than the threshold, the maximum packet size is made larger than the set value. In addition, a plurality of parameters may be controlled simultaneously so that the aggregation interval is shorter than the set value. In addition, an upper limit and a lower limit are set for the change width of the maximum packet size. First, the maximum packet size is changed, and the maximum packet size is the upper limit.
Alternatively, the aggregation interval may be controlled when the lower limit is reached and cannot be changed any more.

The above parameter determination processing may be performed at regular time intervals, or may be performed when the number of packets that have arrived after the previous parameter determination exceeds a certain value, or the communication destination load or channel information value is You may carry out when it changes by a fixed value.

FIG. 4 is a diagram illustrating an example of a processing flow in which the packet aggregation device according to the first embodiment aggregates packets. The flow of packet aggregation will be described with reference to FIG.

First, the cycle management unit 12 determines whether or not the time corresponding to the aggregation interval determined by the flow of FIG. 3 has elapsed since the previous aggregation execution, and if not (S110, No), the process is performed. Return to S110. On the other hand, when the time corresponding to the aggregation interval has passed (S110, Yes), it is determined whether or not there is an untransmitted packet in the aggregation buffer 11.

When there is an untransmitted packet (S111, Yes), if one of the untransmitted packets is selected to obtain the size of the packet and the packets are aggregated into an aggregate packet, the packet size of the aggregate packet is shown. 3 determines whether or not the maximum packet size determined in the processing flow 3 is exceeded. If the maximum packet size is not exceeded (S114, No),
The aggregation unit 13 aggregates the packet into the aggregated packet (S114), and returns the process to S111.

As a packet to be extracted, a packet with the oldest reception date may be selected from among the packets in the aggregation buffer 11, or if there is packet priority information, a packet with a higher priority may be selected. You may choose from a thing with a small size, and you may choose so that the frequency | count of packet extraction between several clients may become fair.

On the other hand, when exceeding a maximum packet size (S114, Yes), the 1st transmission part 14 transmits an aggregation packet to the 2nd communication apparatus 5 (S115), and returns a process to S110.

If there are no untransmitted packets (No at S111), the aggregated packet is transmitted (S115), and the process returns to S110. Further, when one packet is not included in the aggregate packet, the packet need not be transmitted.

FIG. 5 is a diagram illustrating an example of a processing flow when the packet aggregation device 4 according to the first embodiment receives an aggregation packet from the second communication device 5. With reference to FIG. 5, the flow of processing when receiving an aggregate packet from the second communication device 5 will be described.

When the second receiving unit 15 receives the aggregated packet from the second communication device 5 (S120), the dividing unit 16 divides the received aggregated packet into individual packets (S121).

The second transmission unit 17 extracts one packet from among the divided packets that have not been transmitted to the first communication device 3 and transmits the packet to the corresponding first communication device 3 (S122).

Further, it is determined whether there is an untransmitted packet, and if there is an untransmitted packet (S
123, Yes), the process returns to S122. On the other hand, when there is no untransmitted packet (S1
23, No), the process is terminated.

FIG. 6 is a diagram illustrating an example of a processing flow when the packet aggregation device 4 according to the first embodiment receives a packet from the first communication device 3. The flow of processing when receiving a packet from the first communication device 3 will be described with reference to FIG.

When receiving a packet, the first receiving unit 10 determines whether or not there is an empty space in the aggregation buffer 11. If the aggregation buffer 11 is empty (S130, Yes), the packet is received (S131), the received packet is stored in the aggregation buffer 11, and the process returns to S130. On the other hand, when there is no space in the aggregation buffer 11 (No in S130), the process is performed again in S1.
Return to 30.

As described above, according to the first embodiment, congestion control can be performed by controlling parameters such as the packet aggregation interval and the maximum packet size of the aggregated packet. In particular,
The load on the device increases as the number of packets passing through the device in a router or the like and the number of received packets at a communication destination terminal increase. Therefore, by increasing the number of packets included in one aggregated packet, the number of packets passing through the router and the number of packets received at the communication destination can be reduced, leading to a reduction in processing load.

For example, assuming that the processing capacity of a router existing on the network 2 is 100 packets per second, if 200 packets per second arrive at equal intervals before being aggregated from the first communication device 3, the 200 packets are not changed as they are. In the case of transmission to the second communication device 5, the processing capacity of the router is exceeded, and a large communication delay or packet discard may occur.

On the other hand, when packet aggregation is performed with an aggregation interval of 100 ms, the actual number of packets to be transmitted is about 20, which does not exceed the limit of the processing capability of the router.

On the other hand, if the aggregation interval is increased, the real-time property of communication is impaired. For example, if the aggregation interval is 200 ms, the number of packets to be transmitted is 10, but the maximum communication delay time due to aggregation increases from 100 ms to 200 ms.

For this reason, in order to achieve both real-time performance and congestion avoidance, it is necessary to shorten the aggregation interval within a range that does not exceed the processing capability of the router or communication destination.

According to the first embodiment, the aggregation interval and the maximum packet size of the aggregate packet can be dynamically controlled within a range that does not exceed the processing limit with reference to the load information of the devices on the communication path and the communication destination. Similarly, processing congestion can be prevented by referring to the processing load information of the second communication apparatus 5 and dynamically controlling the aggregation interval and the maximum packet size of the aggregated packets.

In particular, it is possible to reduce the load on the network device and the communication destination by changing the number of packets included in the aggregate packet without reducing the number of packets before actual communication or before aggregation.

(Embodiment 2)
In the first embodiment, depending on the value of the aggregation interval and the maximum packet size, the packet aggregation device 4
May send data that exceeds the bandwidth capacity of the network 2. In that case,
There is a possibility that the line will be congested and the communication time will be greatly increased.

If the amount of data that can be sent per hour is known, the aggregation parameter is controlled based on the amount of data, communication delay can be suppressed, and real-time communication can be ensured.

According to the second embodiment, it is possible to predict the amount of data that can be transmitted per time, and to determine aggregation parameters such as the aggregation interval of packets, the maximum packet size of the aggregation packet, and the multiplicity of communication based on the data amount.

Here, the multiplicity will be described. In order to increase the reliability of communication or the success rate of communication, a technique is known in which several aggregate packets are duplicated and the duplicated aggregate packets are transmitted. Hereinafter, duplicating aggregated packets is referred to as multiplexing, and the number of aggregated packets to be replicated is referred to as multiplicity.

The second embodiment is premised on a configuration in which communication is highly reliable by multiplexing the same aggregate packet and continuously transmitting the same aggregate packet that has been multiplexed a plurality of times. For example, UDP (User Datagram Protocol), which is widely used as a communication protocol, does not have a mechanism for packet arrival confirmation. Therefore, UDP itself does not have a mechanism for improving communication reliability by retransmission or the like.

Therefore, by transmitting the same packet a plurality of times, the success rate of communication can be increased without relying on retransmission, and communication can be made highly reliable.

In the second embodiment, the amount of data that can be transmitted per hour is predicted using a communication multiplexing mechanism, and various parameters are adjusted. More specifically, a method for predicting the amount of data that can be transmitted per hour from the arrival interval of multiplexed aggregated packets will be described.

In order to make the explanation easy to understand, the same reference numerals are given to the same components as those in the first embodiment, and
The overlapping description will be omitted as appropriate.

FIG. 7 is a diagram illustrating an example of a functional configuration of the packet aggregation device 4 according to the second embodiment. FIG.
Will be used to describe the functional configuration of the packet aggregation device 4 according to the second embodiment.

The packet aggregation device 4 illustrated in FIG. 7 includes a multiplexing detection unit 200, a bandwidth determination unit 201, and a multiplexing unit 203 in addition to the functions illustrated in FIG.

The multiplexing unit 203 converts the aggregation packet generated by the aggregation unit 13 into the aggregation parameter determination unit 202.
Has a function of replicating (multiplexing) the aggregated packet according to the multiplicity determined by. For example, when the multiplicity is 2, the same aggregate packet is duplicated and passed to the first transmitter 14 twice, whereby the aggregate packet is transmitted twice.

The multiplexing detection unit 200 has a function of counting the number of arrivals of the same aggregate packet as the aggregate packet when the aggregate packet arrives from the second communication device 5. Further, the multiplexing detection unit 200 passes the aggregated packet to the dividing unit if the aggregated packet arrives for the first time, and discards the aggregated packet if it arrives for the second time or later.

The bandwidth determination unit 201 has a function of determining network bandwidth information, that is, the amount of data that can be transmitted per time. As a method of determining the data amount, there are a method of using a fixed value set in advance as a bandwidth, a method of receiving and using bandwidth information from the second communication device 5, and the like.

FIG. 8 shows a method for performing bandwidth prediction using reception time information of multiplexed aggregated packets.
Will be described with reference to FIG. FIG. 8 is a diagram illustrating an example of a reception time calculation method for multiplexed aggregate packets according to the second embodiment.

If the received aggregate packet is the first aggregate packet received, the bandwidth determination unit 201
The reception time (t1 in FIG. 8) is recorded. Also, if the same aggregate packet is received for the second time, the difference between the recorded first aggregate packet reception time (t1 in FIG. 8) and the second packet reception time (t2 in FIG. 8). (T = t2-t1 in FIG. 8) is obtained.

Furthermore, the size (X) of the packet is acquired, and the bandwidth B is
B = X / T (1)
Ask for. This is because the time T required to send a packet of packet size X on the bandwidth B line is
T = X / B (2)
It becomes.

In FIG. 8, there is a free time from the completion of reception of packet # 1 to the start of reception of packet # 2. When the next packet is transmitted immediately after the completion of the first packet transmission, and when all of the packets are transmitted through the network having the same bandwidth from the transmission terminal to the reception terminal, the packet # 2 is transmitted immediately after the reception of the packet # 1 is completed. Reception should begin. However, when there is a network with a narrow bandwidth in the network through which it passes, it takes time to transmit packet # 1.

For example, when 10 KB data is sent over a network with a bandwidth of 1 KB / second, it takes 10 seconds. However, when 10 KB data is sent over a network with a bandwidth of 10 KB / second, transmission is completed in 1 second. .

Here, let us consider a case where 10 KB data is continuously transmitted through a network having a bandwidth of 1 KB / second and a network having a bandwidth of 10 KB / second in this order. A store-and-forward network switch is sandwiched between the networks.

At this time, the packet receiving side is a 10 KB / second network, and the reception of the packet # 1 is completed in one second. However, a device such as a router at the boundary between a 1 KB / second network and a 10 KB / second network waits for the arrival of the entire packet and then performs a CRC check and transmits the packet.

Since it takes 10 seconds to wait for arrival of packet # 2 at this router, packet # 2 at the receiving side
Is started 9 seconds after the completion of reception of packet # 1. Reception of packet # 2 is completed in one second. When the arrival interval T = 9 seconds + 1 second = 10 seconds and the packet size X = 10 KB are applied to the equation (1), B = X / T = 10 KB / 10 seconds = 1 KB / second, and the bandwidth of the narrowest bandwidth A width is required.

In the case of other methods such as the network switch method such as cut-through method, there is almost no idle time on the receiving side in the first place, but it takes 10 seconds from the start of packet reception to the completion of reception through the 1 KB / second network. Therefore, the bandwidth can be obtained by the same equation (1).

The bandwidth was determined by using the time difference between the packet that arrived the first time and the packet that arrived the second time, but if the number of packets to be multiplexed is three, for example, the packet that arrived the first time and the packet that arrived the third time The bandwidth may be obtained using the arrival time difference between the received packets and the arrival time difference between the second arrival packet and the third arrival packet.

Further, when only a specific packet is multiplexed or when a packet drop occurs on the communication path, the same packet may arrive only once. In these cases, 1
Bandwidth calculation is not performed for packets that arrive only once, and bandwidth is calculated only when there is a second arrival packet.

The aggregation parameter determination unit 202 has a function of determining the packet aggregation interval, the maximum packet size of the aggregation packet, and the communication multiplicity using the bandwidth information determined by the bandwidth determination unit 201.

Here, the bandwidth is B, the aggregation interval is I, the maximum packet size is S, and the communication multiplicity is M. At this time, the number of packet aggregation per time is 1 / I. Therefore, the maximum packet transmission amount P per hour is
P = S × (1 / I) × M
It becomes. Since the amount of data that can be sent per hour is B, B ≧ P, that is,
B ≧ S × (1 / I) × M (3)
Need to be. To use up your bandwidth,
B = S × (1 / I) × M (4)
And it is sufficient.

The aggregate parameter determination unit 20 in the second embodiment has a function of determining each parameter of S, I, and M so as to satisfy (3). Hereinafter, some examples of functions of the parameter determination method of the aggregate parameter determination unit 20 will be described.

First, the simplest parameter determination method of the aggregate parameter determination unit 20 will be described. Here, equation (4) is transformed,
S = B × I / M (5)
And For example, if I and M are fixed values and B is a value determined by the band determining unit 201, the value of S is determined. Instead of setting I and M as fixed values, S and M may be fixed values to obtain I so as to satisfy the equation (4), or S and I may be fixed values and the equation (4) may be changed. The aggregate parameter determination unit 20 may be provided with a function of determining M so as to satisfy.

However, in this method, the remaining one parameter is determined by the network bandwidth and two fixed values. For example, when the bottleneck is the processing load of the second communication device 5, the congestion status changes when the maximum packet size S or the aggregation time I is controlled.

However, in equation (5), I and M are fixed values, and the maximum packet size S is the second communication device 5.
It does not depend on the processing load of the network and changes only with changes in the network bandwidth. At this time,
It is assumed that when the processing load of the second communication device 5 is increased for some reason, the network congestion is eliminated and the available bandwidth is widened. At this time, B is large in equation (5), and I and M are fixed values, so the maximum packet size S becomes large.

This means that the number of packets included in the arriving aggregate packet may increase, and the load on the second communication device 5 may further increase. That is, in order to avoid congestion of the processing load of the second communication device 5 and packet discard on the network, it is necessary to change parameters according to the load status of the second communication device 5 and the network.

Accordingly, the aggregation parameter determination unit 20 uses the communication destination information acquired by the communication destination information acquisition unit 18 and the communication path information acquired by the communication path information acquisition unit 19 to collect packets, the maximum packet size of the aggregate packet, and the multiplicity of communication. It may have a function of determining

Below, several systems are demonstrated about the function in which the aggregation parameter determination part 20 determines a parameter according to the load condition of the 2nd communication apparatus 5 or a network apparatus.

First, a method for introducing a variable W for controlling congestion will be described. When the equation (3) is transformed, B, I and M are all positive values.
S ≦ B × I / M (6)
It becomes.

here,
S = B × I / M × W (7)
When the variable W is introduced,
B × I / M × W ≦ B × I / M (8)
Therefore, if W ≦ 1, Equation (6) is satisfied.

By adjusting this value W according to the load status, that is, the communication destination information acquired by the communication destination information acquisition unit 18 and the communication path information acquired by the communication path information acquisition unit 19, the maximum packet size S
Will vary depending on the load conditions.

For example, AIMD (Additive increase / mult) used in TCP / IP congestion control
The aggregation parameter determination unit 20 may determine the variable W from the communication destination information and the communication path information using a generally well-known technique such as “iplicative decrease”.

For example, the buffer overflow occurrence count E (t
), When W is determined from the packet drop occurrence count D (t) on the communication path as the communication path information, A (A> 0), B (0 <B <1) are constants, and the value of W at time t + 1 W (t + 1)
Is a range satisfying W ≦ 1, using the value of W at time t, W (t),
W (t + 1) = W (t) + A (when E (t) <threshold 1 and D (t) <threshold 2)
W (t + 1) = W (t) × B (when E (t) ≧ threshold 1 or D (t) ≧ threshold 2)
And decide.

Further, the maximum packet size S is obtained by substituting this value into the equation (7). As a result, the maximum packet size S decreases as the load on the second communication device 5 or the network device increases, increases as the load decreases, and congestion can be avoided. Further, the aggregation interval I may be obtained by the same method with S as a fixed value. At this time, if equation (3) is transformed,
I ≧ S / B × M (9)
And
I = S / B × M × W (10)
Then, if W ≧ 1, Expression (10) satisfies Expression (9).

Here, in order to reduce the number of aggregated packets when the buffer overflow occurrence count E (t) per second time of the second communication apparatus and the packet drop occurrence count D (t) on the communication path as the communication path information are equal to or greater than a threshold value. It is necessary to lengthen the aggregation interval. That is, since it is sufficient to increase the value of W from Equation (9), constant F (F> 0) and constant G (G> 1)
W (t + 1) = W (t) −F (when E (t) <threshold 1 and D (t) <threshold 2)
W (t + 1) = W (t) × G (when E (t) ≧ threshold 1 or D (t) ≧ threshold 2)
And W are controlled, and I can be obtained from equation (9). The same method may be adopted when the degree of aggregation M is changed.

In (7) and (10), the parameter to be changed is in addition to S, I or M,
W only. However, depending on the congestion situation, it may be desired to control S, I, and M independently.

For example, it is useful to increase M when packet drops occur frequently due to network quality problems, and I may increase when buffer overflow occurs frequently due to the processing load of the second communication device. Useful.

In addition, it is useful to reduce M when there are too many packets passing through the network device and packet drops occur frequently because the processing is not in time.

Further, when packet drop and buffer overflow occur at the same time, it is desired to change M and I at the same time, but this is not possible with the equations (7) and (10). Therefore, the aggregation parameter determination unit 2
A method in which 0 controls each parameter independently will also be described.

First,
S = B × I / M (11)
far.

Here, I (t + 1) is calculated based on I (t) and the number of buffer overflows E (t) per time.
(> 0), J (> 1) as constants
I (t + 1) = I (t) −H (E (t) <threshold 1) (12)
I (t + 1) = I (t) × J (E (t) ≧ threshold 2) (13)
As a result, the aggregation interval becomes longer if the buffer overflow occurrence count E (t) per time is equal to or greater than the threshold value, and the aggregation interval becomes shorter if it is less than the threshold value.

On the other hand, P (t) is the number of packet drops due to garbled data due to the quality of the communication channel, and O (t) is the number of packet drops due to exceeding the number of packets that can be processed.
M (t + 1), which is a multiplicity M of 1, uses M (t),
M (t + 1) = M (t) +1 (P (t) ≧ threshold 3 and O (t) <threshold 4) (14)
M (t + 1) = M (t) −1 (O (t) ≧ threshold 4 and P (t) <threshold 3) (15)
M (t + 1) = M (t) (otherwise) (16)
And

As a result, when packet drops due to channel quality occur, the multiplicity increases to increase the communication success rate, and when packet drops occur due to increased device load on the channel, the load on the communication device In order to reduce the multiplicity, the multiplicity decreases.

However, in many cases, it is often impossible to distinguish the number of packet drops due to data corruption due to the quality of the communication path and the number of packet drops due to exceeding the number of packets that can be processed by the network device.

In that case, only the total value of P (t) and O (t) is known, and M (t) is controlled by that value. However, if only the total value information of the number of packet drops is determined, it cannot be determined whether the drop rate decreases if the value of M (t) increases or decreases if the value of M (t) decreases.

So, if the number of packet drops is greater than or equal to the threshold, first increase the value of M (t)
Thereafter, if the number of packet drops decreases, M (t) is changed to increase, and if the number of packet drops increases, M (t) may be changed to decrease.

Moreover, although I (t) is changed by the continuous value using AIMD, a table may be provided and a discrete value may be taken. For example, a table of discrete values such as 1 second, 2 seconds, and 4 seconds is prepared for possible values of I (t), and E (t) ≧ threshold when I (t) = 1 second. I (t + 1
) To 2 seconds.

According to the above method, a plurality of parameters such as I and M can be controlled independently, but there is often a limit to the range in which the parameters can be varied. For example, when the aggregation interval I is increased, the load on the second communication device 5 is reduced, but the real-time property of communication is lost.

Simply, an upper limit value and a lower limit value of a parameter may be provided and control may be performed within that range. However, there are cases where congestion cannot be avoided within an adjustable range.

For example, if congestion cannot be avoided even if the aggregation interval I is increased to the upper limit, in order to reduce the number of packets included in the aggregated packet, the maximum packet size S, which is a dependent variable in equation (11), is independent. Control as a variable. For this purpose, a variable W for controlling the maximum packet size S is introduced, and the maximum packet size S is indirectly controlled.
Here, as in equation (7),
S = B × I / M × W (17)
And

However, unlike equation (7), I and M are variables. Here, when I reaches the upper limit,
A method for controlling W will be described. W (0) = 1 (where W (t) has an upper limit of 1), and K (K> 0), L (L> 1), N (0 <N <1), and R (R> 0) As a constant
I (t + 1) = I (t) −K (W (t) = 1 and E (t) <threshold 1) (18)
I (t + 1) = I (t) × L (I (t) <upper limit value and E (t) ≧ threshold value 2) (19)
W (t + 1) = W (t) × N (I (t) = upper limit value and E (t) ≧ threshold value 6) (20
)
W (t + 1) = W (t) + R (W (t) <1 and E (t) <threshold 5) (21)
I (t + 1) and W (t + 1) are calculated as follows.

However, the upper limit value of W is 1. Further, M may be controlled using equations (14)-(16). As a result, when E (t) is equal to or greater than the threshold value, until I (t) reaches the upper limit value (19
) Formula is applied, the load of the second communication device 5 is reduced by extending the aggregation interval, and when the upper limit is reached, the formula (20) is applied, and the maximum packet size S is decreased to reduce the load per time. The load on the second communication device 5 can be reduced.

On the contrary, when the congestion is eliminated and E (t) becomes less than the threshold, first, the value of W (t) is returned to the original value by the equation (21) so as to approach the initial value 1, and then I ( t) is shortened (18
) Apply the formula.

FIG. 9 is a diagram illustrating an example of bandwidth prediction and parameter determination flow performed by the packet aggregation device 4 according to the second embodiment. The bandwidth prediction and parameter determination flow according to the second embodiment will be described with reference to FIG.

The second receiving unit 15 waits for reception of the aggregate packet (S210), and when receiving the aggregate packet, the multiplexing detection unit 200 determines whether or not the aggregate packet is received for the first time. For the first time (
In S211, Yes, the arrival time is recorded (S212), and the process returns to S210.

On the other hand, if the aggregate packet is not received for the first time (S211, No), it is determined whether the aggregate packet is received for the second time, and if it is not the second time (S213, No), the process is performed in S21.
Return to zero.

On the other hand, when the aggregate packet is received for the second time (S213, Yes), the time difference between the current time and the time when the first aggregate packet received recorded in S212 is calculated (S214).

Further, the bandwidth is calculated by dividing the aggregate packet size by the time difference calculated in S214 (
S215).

Also,
Bandwidth B ≧ maximum packet size S × (1 / aggregation interval I) × multiplicity M
Each parameter is calculated so that As the calculation method, the method already described may be used.

According to the second embodiment described above, the aggregation interval of aggregation packets, the multiplicity of communication, and the like can be flexibly changed according to the load status of the network device and the second communication device 5 within the network bandwidth.
The maximum packet size can be adjusted.

In particular, the normal congestion control is performed by adjusting the number of packets that can be sent per hour and the total amount of data. In the second embodiment, the total amount of packets and data included in the aggregated packets per hour is changed. Therefore, congestion control can be performed while flexibly controlling a plurality of aggregate parameters simultaneously.

(Embodiment 3)
In the second embodiment, the packet aggregation device 4 performs parameter determination by a single device. On the other hand, when there are a plurality of packet aggregating apparatuses 4 as in the configuration shown in FIG. 10, the bandwidth measurement result may vary. FIG. 10 is a diagram illustrating an example of a device configuration of a transmission processing system according to the third embodiment.

At this time, for example, when the maximum packet size S is calculated using the equation (5), the amount of data that can be transmitted per hour between the packet aggregation devices 4 changes, and an inequality state occurs. In order to ensure equality of communication between the packet aggregating apparatuses 4, it is necessary to perform arbitration so that the bandwidths calculated by the respective packet aggregating apparatuses are the same in a network having the same bandwidth.

Thus, in the third embodiment, a packet aggregation device that ensures communication equality will be described assuming an environment in which a plurality of packet aggregation devices are connected by a network as shown in FIG.

FIG. 11 is a diagram illustrating an example of a functional configuration of the packet aggregation device according to the third embodiment. FIG.
1 differs from FIG. 7 in that an arbitration unit 300 is provided. In order to make the description easy to understand, the same components as those in the first and second embodiments are denoted by the same reference numerals, and redundant descriptions are omitted as appropriate.

As in the second embodiment, the aggregate parameter determination unit 20 uses the bandwidth information determined by the bandwidth determination unit 201, the communication destination information acquired by the communication destination information acquisition unit 18, and the communication path information acquired from the communication path information acquisition unit 19. The parameters such as multiplicity M, maximum packet size S, and aggregation interval I are finally determined.

In the second embodiment, the bandwidth information uses the information determined by the bandwidth determination unit 201 as it is, but in the third embodiment, the bandwidth information obtained by the aggregate parameter determination unit 20 from the bandwidth determination unit 201 is used as the arbitration unit 300. To pass.

The arbitration unit 300 has a function of transmitting bandwidth information to another packet aggregation device via the network 320. The arbitrating unit 300 also has a function of acquiring bandwidth information calculated by each packet aggregation device from another packet aggregation device via the network 320.

The arbitration unit 300 receives the bandwidth information from other packet aggregation devices and the aggregation parameter determination unit 20.
Bandwidth information obtained from, and arbitrated bandwidth information is calculated from As a method of calculating the adjusted bandwidth information, an arithmetic average may be taken, a weighted average obtained by increasing the weight of the bandwidth information, or a median value may be taken.

The arbitrating unit 300 passes the arbitrated bandwidth information to the aggregate parameter determining unit 20. The aggregation parameter determination unit 20 has a function of determining parameters such as the multiplicity M, the maximum packet size S, and the aggregation interval I using not the bandwidth information obtained from the bandwidth determination unit 201 but the adjusted bandwidth information.

The calculation method uses the method described in the second embodiment, and the arbitrated bandwidth information is set as B ′.
S ≦ B ′ × I / M
It suffices to calculate so as to satisfy.

FIG. 12 is a diagram illustrating an example of a flow of bandwidth prediction and parameter calculation processing performed by the packet aggregation device according to the third embodiment. A bandwidth prediction and parameter calculation flow according to the third embodiment will be described with reference to FIG. FIG. 12 shows the flow of FIG. 9 described in the second embodiment in S310.
Is different from the point in which the process of S311 is changed. After the bandwidth determination unit 201 calculates the bandwidth in S215, the arbitration unit 300 uses the bandwidth information received from the other packet aggregation devices and the bandwidth information calculated by the bandwidth determination unit 201 to calculate the adjusted bandwidth information. Calculate (S310).

Furthermore, the aggregate parameter determination unit 20 determines each parameter using the arbitrated bandwidth instead of the bandwidth. Although not explicitly shown in FIG. 12, the arbitration unit 300 may transmit bandwidth information to other packet aggregation devices in the step of S310, or the arbitration unit 300 transmits bandwidth to other packet aggregation devices at regular intervals. The width information may be transmitted.

According to the third embodiment described above, when there are a plurality of packet aggregating apparatuses, an inequality situation where only a specific packet aggregating apparatus transmits a large number of packets is avoided, and communication equality is ensured. It becomes possible. Here, the method of taking the average value of the bandwidth information has been described. However, as variations of the aggregation parameter determination unit 20 and the arbitration unit 300, parameters such as multiplicity M, maximum packet size S, and aggregation interval I are used instead of bandwidth information. May be shared by a plurality of packet aggregating apparatuses and the aggregation parameters may be arbitrated directly.

(Embodiment 4)
Next, a fourth embodiment in which the transmission processing system according to the first, second, and third embodiments is adapted to a station service system used at a station such as a railway will be described with reference to FIG.

FIG. 13 is a diagram illustrating an example of a configuration in which the transmission processing system according to the fourth embodiment is adapted to a station service system.

In order to make the description easy to understand, the same components as those in Embodiments 1, 2, and 3 are denoted by the same reference numerals, and redundant descriptions are omitted as appropriate.

When the transmission processing system 100 is adapted to the station service system 101, as shown in FIG.
For example, the first communication device 3 corresponds to the ticket gate 50 and the second communication device 5 corresponds to the server 51.

The station service system 101 includes a ticket gate 50, a packet aggregation device 4, and a server 51.

The station service system 101 is used, for example, to perform entrance / exit processing into a station premises for a station user in a railway or the like.

The ticket gate 50 has a function of processing a utilization medium presented by the user. The use medium is, for example, a magnetic medium (entrance ticket, ordinary ticket, commuter pass, coupon ticket, prepaid card, etc.) on which record information such as a valid period and an available section is recorded, and record information is transmitted and received by wireless communication. A wireless medium (such as an IC card). In the following description, the use medium is IC.
It will be described as a card.

The IC card is an IC (Integrated Circuit) that can accept various processes by wirelessly transmitting and receiving data to and from various station service devices such as ticket vending machines and ticket gates.
it). Further, the IC card has a storage unit, for example, to use the IC card as a commuter pass, an SF (Stored Fare) card for fare settlement, or a post-payment card for fare settlement. The information is memorized. For example, each IC card stores a unique identification number, usage history information, balance information, commuter pass information, and the like.

The ticket gate 50 performs the ticket gate processing necessary for entry / exit such as the IC card identification number, entry / exit status information, remaining amount information, commuter pass information, etc. stored in the IC card from the IC card presented by the user. Receive information. The ticket gate 50, for example, at the time of entrance, an identification number, an entrance station,
Information such as admission time is received from the IC card as ticket gate processing information, and at the time of participation, information such as an identification number, a participating station and a participation time is received from the IC card as ticket gate processing information. Also,
The ticket gate 50 transmits a packet storing the ticket processing information to the packet aggregation device 4.

The packet aggregating apparatus 4 aggregates a plurality of packets received from the ticket checker 50, and a server 51
Send to.

The server 51 separates the aggregate packet received from the packet aggregation device 4 into individual packets. The server 51 performs arithmetic processing on individual packets, that is, ticket gate processing information.

For example, in the case of admission processing, the calculation processing is whether the remaining amount included in the ticket gate processing information satisfies the minimum fare or whether the identification number included in the ticket gate processing information is registered in the server's negative list. Processing for determining whether or not is performed is performed. In addition, regarding the entrance process, it may be configured such that only the ticket gate process information is received, and the process for allowing entrance is performed without performing the arithmetic process on the ticket gate process information.

In addition, for example, in the case of a participation process, the calculation process includes a calculation process for calculating a fare from information on an entry / exit station, a calculation process for removing a fare from an IC card, and the like. Furthermore, the server 51 aggregates a plurality of packets and transmits the results of the plurality of arithmetic processes to the packet aggregation device 4.

The packet aggregating apparatus 4 separates the aggregated packet received from the server 51 into individual packets, and transmits them to the corresponding ticket gate 50.

The ticket gate 50 transmits the ticket gate control information to the IC card based on the packet received from the packet aggregation device 4. For example, a gate opening / closing process of the ticket gate 50 and a process of writing a fare withdrawal process result from an IC card presented by the user are performed.

According to the station service system to which the transmission processing system according to the fourth embodiment described above is applied, parameters related to packet aggregation such as a packet aggregation interval and a maximum packet size are flexibly changed according to a network congestion state and a communication destination load state. This enables real-time control even during commuting rush hours when the number of packet communications increases rapidly.

The transmission processing system 100 according to the first, second, third, and fourth embodiments described above is not limited to the application to the station service system 101 described above, but various societies such as disaster prevention systems and water treatment systems. It can be applied to various industrial systems such as infrastructure systems and building controls. The first communication device 3 may be, for example, hardware or software of a social infrastructure system such as a plant, a disaster prevention system, or a traffic system, or may be hardware or software of an industrial system, a home appliance, a PC, an information terminal, or the like.

Note that the processing in the transmission processing system according to Embodiments 1, 2, 3, and 4 described above can be executed by some software. Therefore, the above processing can be easily realized only by installing and executing these programs on the packet aggregation device 4 through a computer-readable storage medium storing several programs for executing the above processing procedures. . For example, the packet aggregation device 4 can download the program via the network, store the downloaded program, and complete the installation of the program. Alternatively, the packet aggregation device 4 can read the program from the information storage medium, store the read program, and complete the installation of the program.

Although several embodiments of the present invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions can be made without departing from the spirit of the invention.
Can be replaced or changed. These embodiments and their modifications are included in the scope and gist of the invention, and are also included in the invention described in the claims and the equivalents thereof.

DESCRIPTION OF SYMBOLS 1 Network 2 Network 3 1st communication apparatus 3a 1st communication apparatus 3b 1st communication apparatus 4 Packet aggregation apparatus 5 2nd communication apparatus 10 1st receiving part 11 Aggregation buffer 12 Period management part 13 Aggregation part 14 1st transmission part 15 Second receiving unit 16 Dividing unit 17 Second transmitting unit 18 Communication destination information acquiring unit 19 Communication path information acquiring unit 20 Aggregated parameter determining unit 50 Ticket gate 51 Server 100 Transmission processing system 101 Station service system 200 Multiplexing detecting unit 201 Band determination Unit 202 aggregation parameter determination unit 203 multiplexing unit 300 arbitration unit 320 network

Claims (6)

A packet aggregation device that aggregates a plurality of packets received from a first communication device and transmits the packets to a second communication device via a network,
A first receiver for receiving packets from the first communication device;
Aggregating unit that aggregates the packets received by the first receiving unit at a certain aggregation interval and generates an aggregated packet until reaching a certain maximum packet size;
A first transmission unit for transmitting the aggregate packet generated by the aggregation unit to the second communication device;
A second receiver for receiving a response to the aggregate packet transmitted by the first transmitter from the second communication device;
Based on the response received by the second receiving unit, a communication destination information acquiring unit that acquires processing load information that is information indicating the load of the second communication device;
Based on the response received by the second receiving unit, a channel information acquisition unit that acquires channel information that is information indicating the load of the network;
Based on at least one of the processing load information acquired by the communication destination information acquisition unit and the communication path information acquired by the communication path information acquisition unit, at least one of the maximum packet size and the aggregation interval. An aggregation parameter determination unit that determines one of them,
A packet aggregation device comprising: The aggregate parameter determination unit further has a function of determining multiplicity based on at least one of processing load information of the second communication device and processing load information of the network,
A multiplexing unit that multiplexes aggregate packets according to the multiplicity determined by the aggregation parameter determination unit;
The packet aggregation device according to claim 1, further comprising: The multiplexed aggregate packet generated by the multiplexing unit further includes a bandwidth determination unit that determines a network bandwidth from a time difference received by the second communication device,
The packet aggregation device according to claim 2, wherein the aggregation parameter determination unit determines the aggregation interval, the maximum packet size, and the multiplicity based on the bandwidth determined by the bandwidth. Arbitration unit that transmits / receives one or more parameters of the aggregation interval, the maximum packet size, and the multiplicity to / from another packet aggregation device and determines a parameter based on the parameters of the other packet aggregation device When,
The packet aggregation device according to claim 2, further comprising: One or more parameters of the aggregation interval, the maximum packet size, the multiplicity, and the bandwidth are transmitted / received to / from another packet aggregation device, and the parameters are determined based on the parameters of the other packet aggregation device. An arbitration department to determine;
The packet aggregation device according to claim 3, further comprising: A transmission processing program provided in a packet aggregation device that aggregates a plurality of packets received from a first communication device and transmits the packets to a second communication device via a network,
Receiving a packet from the first communication device;
A procedure of aggregating packets received from the first communication device at a certain aggregation interval until a certain maximum packet size is reached, and generating an aggregate packet;
Transmitting the aggregate packet to the second communication device;
Receiving a response to the aggregate packet transmitted to the second communication device from the second communication device;
A procedure for acquiring processing load information, which is information indicating the load of the second communication device, based on the response;
A procedure for acquiring communication path information that is information indicating a load on the network based on the response;
A procedure for determining at least one of the maximum packet size and the aggregation interval based on at least one of the processing load information and the communication path information;
A transmission processing program that causes a computer to execute.

Images