JP2009212632A - Communication equipment and communication method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a communication method capable of suppressing actually used effective bands to be allocated bands or within it without discarding packets as unwanted when the actually used effective bands exceed the allocated bands. <P>SOLUTION: A packet processing part 205 of a relay node 201 calculates effective bands to be used by traffic from a packet size and a bit rate of the traffic, and the present communication rate, and a sets an aggregation condition in traffic with a small packet size so as to satisfy that the actually used effective bands is within the allocated effective bands. Aggregation processing parts 207, 209 and 211 perform aggregation to aggregate a plurality of packets into one packet according to the set aggregation condition to thereby suppress the effective bands to be used by the traffic. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a communication apparatus and a communication method, and more particularly to a communication apparatus and a communication method for controlling an effective band actually used in a flow to be equal to or less than an allocated band.

  As a wireless access method, a wireless LAN using a method defined by IEEE802.11 is widely used. In this type of wireless LAN, each time a data frame is transmitted, the sender receives an ACK frame that is a response confirmation from the communication partner. The sender determines whether the transmission is successful by receiving the ACK. Further, in the transmission of the data frame, in addition to the transmission time of the ACK, DIFS (Distributed Inter Frame Spacing) and SIFS (Short Inter Frame Spacing), which are waiting times in the frame transmission interval, occur.

Therefore, the substantial time required for the packet to be transferred in the wireless LAN is Tr (DATA), which is the time required to transmit the data frame, and Tr (ACK) is the time required to transmit the ACK frame.
DIFS + Tr (DATA) + SIFS + Tr (ACK)… (1)
Can be expressed as

  Tr (ACK), DIFS, and SIFS always occur during data communication, and thus affect the wireless communication throughput as overhead. In particular, since a substantially constant time is required regardless of the communication rate in the wireless section, the higher the wireless communication rate, the greater the influence of these overheads on the throughput. First, since ACK is transmitted at the lowest communication rate, even when the communication rate changes, the communication time Tr (ACK) due to ACK does not change. Similarly, since DIFS and SIFS are constant regardless of the communication rate, when the communication rate is high, the influence on overhead increases. In the transmission of the data frame itself, the preamble portion associated with the data frame is transmitted at the lowest communication rate lower than that of the data frame.

  Furthermore, since these overheads occur every time a frame is transmitted regardless of the packet size, the smaller packets are more susceptible to the overhead. Therefore, in the case of the same bit rate, the number of transmissions increases for flows with a small packet size, and as a result, the radio resources used per bit to be transmitted increase. In other words, the effective bandwidth actually used differs depending on the packet size even with the same bit rate traffic.

  In this way, in wireless LAN, the effective bandwidth actually used varies greatly depending on the communication rate and packet size, so bandwidth allocation and bandwidth control based on the bit rate performed in wired networks have different data sizes. It doesn't work effectively when there are multiple flows.

  An example of a bandwidth control technique that solves such a problem is described in Patent Document 1. In the technique described in Patent Document 1 (hereinafter referred to as related technique 1 of the present invention), when a new communication request is received, the traffic occupation time required for the new request is calculated in consideration of SIFS, ACK, and the like. Whether or not a new request can be accepted is determined based on the surplus bandwidth obtained from the occupied time allocated to another request that has already been accepted (admission control). When a new request is received, a bandwidth is allocated to the new request, and thereafter traffic is monitored. If the traffic exceeds the bandwidth allocated at the time of reception, a certain amount of packets are forcibly discarded to control the used bandwidth to be equal to or lower than the allocated bandwidth.

  On the other hand, as a technique for improving the throughput of the wireless LAN, there is an aggregation technique for transmitting a plurality of data packets in one packet. By transmitting a plurality of small-sized packets as one large packet by this aggregation technique, the number of transmissions is reduced, so overhead is suppressed and throughput is improved. However, as a problem of the aggregation technique, there is a side effect in which waiting time is delayed because a plurality of data is accumulated. There is also a problem that the larger the size of the packets collected by aggregation, the more likely the frame is broken due to an error in the wireless environment.

JP 2007-74210 A JP 2007-13520 A

  According to the related art 1 of the present invention, since the bandwidth control is performed from the traffic occupation time, the bandwidth control can be effectively performed even when there are a plurality of flows having packets of different sizes. However, when the bandwidth actually used exceeds the allocated bandwidth, there is a problem in terms of effective utilization of radio resources because there is only a method for discarding packets to reduce the bandwidth used.

  An object of the present invention is to reduce the effective bandwidth that is actually used to less than the allocated bandwidth without discarding packets as much as possible when the effective bandwidth that is actually used exceeds the allocated bandwidth. It is to provide a communication method.

  The communication apparatus according to the present invention measures an actual use band of the one or more flows, an aggregation processing means for aggregating a plurality of packets to be sent to one or more flows into one according to the set aggregation condition, And a packet processing unit that sets an aggregation condition in the aggregation processing unit in order to reduce a use band when the band exceeds the allocated band.

  According to the present invention, even when the effective bandwidth that is actually used exceeds the allocated bandwidth, the effective bandwidth that is actually used can be suppressed below the allocated bandwidth without discarding the packet as much as possible. . The reason is to suppress the used bandwidth to the allocated bandwidth or less by aggregation that aggregates a plurality of packets into one.

  Next, embodiments of the present invention will be described in detail with reference to the drawings.

“First Embodiment”
Referring to FIG. 1, an example of a network configuration to which the present invention is applied is as follows. Terminals 101, 102, 105, 106 such as personal computers (PCs) that require bandwidth allocation for their own traffic, and traffic It is composed of relay nodes 103 and 104 that perform bandwidth allocation and secure and manage bandwidth. Here, the terminal 101 is connected to the relay node 103 by wire, and the terminal 102 and the relay node 103 are connected wirelessly. Further, the relay node 103 and the relay node 104 are connected by radio, and the terminals 105 and 106 are connected to the relay node 104 by wires. The effect of the present invention is more conspicuous when the radio band is controlled as an effective band. However, the network to which the present invention is applied does not have to be connected to all communication apparatuses by radio, and the network illustrated in FIG. Thus, there may be a communication section connected by wire.

  Referring to FIG. 2, when the terminal 101 communicates with the terminal 106 and the terminal 102 communicates with the terminal 105 via the relay nodes 103 and 104, the terminal 101 and the terminal 102 are connected to each other before starting communication. For a traffic class consisting of one flow or a plurality of flows, the relay nodes 103 and 104 are requested to allocate a band. Specifically, bandwidth reservation is performed using RSVP (Resource Reservation Protocol) or the like (111, 113). Then, communication is performed (112, 114).

  The relay nodes 103 and 104 determine the flow rate for traffic from the terminals 101 and 102 based on various conditions such as the bit rate that is the bit amount per second of the flow, the packet size, etc. The effective bandwidth to be actually used is obtained and the bandwidth to be allocated is determined.

  Next, when the communication between the terminal 101 and the terminal 106 and the communication between the terminal 102 and the terminal 105 are started, the relay nodes 103 and 104 measure the communication status such as the average bit rate and the average packet size of the communication, The effective bandwidth actually used is obtained from the measurement result and compared with the allocated effective bandwidth. The effective bandwidth that is assigned to the effective bandwidth used by the flow or the entire traffic class due to changes in the traffic volume of the flow to which the bandwidth is assigned or the traffic class to which the bandwidth is assigned, or changes in the communication rate. In this case, the relay nodes 103 and 104 maintain the bit rate at the time of bandwidth allocation even by the aggregation by performing aggregation from the average packet size and average bit rate of flows or flows belonging to the traffic class. Find out if you can.

  If the effective bandwidth used by a flow or traffic class is less than or equal to the allocated bandwidth due to aggregation up to a certain data size, the relay nodes 103 and 104 are set to perform aggregation up to that data size. Bandwidth control by packet discard is not performed. On the other hand, the data size is already the maximum size allowed by the system, and there is no room to improve the bandwidth by aggregation, or even if there is room, the effective bandwidth that is actually used is allocated to the allocated bandwidth. If it cannot be suppressed below, the effective bandwidth used by the flow or traffic class is suppressed below the allocated bandwidth by using the bandwidth control by packet discard together.

  After that, the effective bandwidth used by the flow or the entire traffic class is changed to the effective bandwidth allocated to the flow or the entire traffic class due to the change in the traffic volume of the entire traffic class to which the bandwidth is allocated or the communication rate. When the bandwidth drops below a certain bandwidth that is a predetermined amount or a predetermined ratio below the bandwidth, the relay nodes 103 and 104 review the packet discard amount and the aggregation condition. The reason for reviewing is that there is a possibility that the used effective bandwidth can be kept below the allocated bandwidth even if the packet discard amount is reduced or even if it is zero. In addition, even if the packet size after aggregation is made smaller or the aggregation is canceled, the usable effective bandwidth may be able to be suppressed below the allocated bandwidth.

  Next, a configuration example of the relay nodes 103 and 104 will be described in detail.

  Referring to FIG. 3, an example of the relay node 201 to which the present invention is applied includes three communication units 202 to 204, an aggregation processing unit 207 and a bandwidth control unit 208 provided corresponding to the communication unit 202, a communication unit Aggregation processing unit 209 and band control unit 210 provided corresponding to 203, aggregation processing unit 211 and band control unit 212 provided corresponding to communication unit 204, these communication units 202 to 204, aggregation processing A packet processing unit 205 connected to the units 207, 209, and 211 and the band control units 208, 210, and 212; and a band management unit 206 connected to the packet processing unit 205. Each has the following functions.

  The communication units 202 to 204 perform wired or wireless communication between the own relay node 201 and other communication devices. For example, when the relay node 201 is used as the relay node 103 in FIG. 1, for example, the communication unit 204 communicates with the terminal 101, the communication unit 203 communicates with the terminal 102, and the communication unit 202 communicates with the relay node 104. The communication units 202 to 204 transmit packets received from other communication devices to the packet processing unit 205, and conversely, from the packet processing unit 205 through the bandwidth control units 208, 210, and 212 and the aggregation processing units 207, 209, and 211. The received packet is transmitted to another communication device. Although the relay node 201 in this example includes three communication units, the number of communication units may be two or less, or may be three or more.

  In addition to normal packet transfer processing such as header analysis and transfer destination determination, the packet processing unit 205 decomposes aggregated data, measures the amount of traffic such as measurement of the average bit rate and average packet size for each flow, The effective bandwidth is calculated based on the measurement result and compared with the allocated bandwidth, and the aggregation processing units 207, 209 and 211 and the bandwidth control units 208, 210 and 212 are controlled based on the comparison result.

  The aggregation processing units 207, 209, and 211 aggregate a plurality of packets transmitted from the packet processing unit 205 into one packet according to the aggregation condition set by the packet processing unit 205, and corresponding bandwidth control units 208, 210, Send to 212. In the present embodiment, one of the aggregation conditions is a waiting time. Aggregation processing units 207, 209, and 211 store packets sent from the packet processing unit 205 regarding the flow to be aggregated in the buffer for a set waiting time, and accumulate in the buffer when the waiting time elapses. Aggregate multiple packets into one packet. In addition, it is possible to make additional changes such as setting the number of packets as an aggregation condition, and collecting the set number of packets and collecting them into one packet.

  The bandwidth management unit 206 performs bandwidth allocation in units of flows or traffic classes composed of a plurality of flows. Specifically, the bandwidth management unit 206 obtains the effective bandwidth that is actually used from the packet size, the bit rate, and the communication rate (communication capability) of the self-relay node 201 at that time, manages it as the allocated bandwidth, The information is provided to the packet processing unit 205.

  Band control units 208, 210, and 212 transmit packets received from packet processing unit 205 and corresponding aggregation processing units 207, 209, and 211 to corresponding communication units 202 to 204. At that time, the bandwidth control units 208, 210, and 212 are set to the set amount when the packet discard amount is set from the packet processing unit 205 in units of traffic or in units of traffic classes including a plurality of flows. Process to discard the packet.

  The communication units 202 to 204, the packet processing unit 205, the bandwidth management unit 206, the aggregation processing units 207, 209, and 211, and the bandwidth control units 208, 210, and 212 described above are realized by, for example, a computer and a program that operates on the computer. be able to. The program is provided by being recorded on a computer-readable recording medium such as a magnetic disk, and is read by the computer when the computer is started up, etc., and by controlling the operation of the computer, the communication units 202 to 204, The packet processing unit 205, the bandwidth management unit 206, the aggregation processing units 207, 209, and 211 and the bandwidth control units 208, 210, and 212 are realized.

  Next, the operation of the relay node 201 will be described.

  The bandwidth management unit 206 of the relay node 201 performs bandwidth allocation in units of flows or traffic classes composed of a plurality of flows. The allocated bandwidth information is managed by the bandwidth management unit 206 and referred to by the packet processing unit 205. When the bandwidth is allocated, the bandwidth management unit 206 obtains an effective bandwidth that is actually used from the packet size and bit rate and the communication rate (communication capability) of the relay node 201 at that time, and manages this as the allocated bandwidth.

  Thereafter, when packet transmission / reception is started through the flow to which the bandwidth is allocated, the packet processing unit 205 of the relay node 201 executes the processing illustrated in FIG. 4 in units of flows or in units of traffic classes including a plurality of flows. .

  First, the packet processing unit 205 classifies the packets input from the communication units 202 to 204, measures the traffic amount such as the average packet size and the average bit rate for each flow, and from each measurement result, The effective bandwidth used by the traffic class is obtained (S101). Next, the packet processing unit 205 compares the obtained effective bandwidth with the allocated effective bandwidth of the flow or traffic class managed by the bandwidth management unit 206 (S102).

  Next, when the effective bandwidth actually used exceeds the allocated bandwidth (NO in S102), the packet processing unit 205 calculates an aggregation condition (S103), and performs aggregation of the packets of the flow. Set to 207, 209, and 211 (S104). Here, in the calculation step S103 of the aggregation condition, the goal is to calculate an aggregation condition that can suppress the actual effective bandwidth to be less than or equal to the allocated bandwidth. Calculate and set an aggregation condition that results in a packet length. Next, if the actual use effective bandwidth cannot be kept below the allocated bandwidth by aggregation alone, a packet discard amount is calculated according to the difference between the used effective bandwidth after the aggregation and the allocated bandwidth (S105). The bandwidth control units 208, 210, and 212 that perform bandwidth control are set (S106).

  On the other hand, if the effective bandwidth that is actually used does not exceed the allocated bandwidth (YES in S102), the packet processing unit 205 determines whether the effective bandwidth that is actually used is equal to or less than the condition review threshold ( S107). Here, the condition review threshold is set to a band that is a predetermined amount smaller than the allocated band or a band that is a predetermined ratio of the allocated band. If the effective bandwidth actually used exceeds the condition review threshold (NO in S107), or if it is not exceeded, aggregation is not being performed for the flow (NO in S108), the packet processing unit 205 proceeds to step S101. It returns and repeats the process mentioned above.

  If the effective bandwidth actually used is equal to or smaller than the condition review threshold value (YES in S107) and aggregation is being executed for the flow (YES in S108), the packet processing unit 205 sets the packet discard amount. Also, the aggregation conditions are reviewed (S109, S110). The reason is that, if a decrease in the traffic volume occurs, the effective bandwidth actually used may be suppressed below the allocated bandwidth without performing aggregation or packet discard.

  First, in the packet discard amount review process in step S109, if a packet discard amount larger than 0 is currently set, the packet processing unit 205 does not actually reduce all of the set packet discard amounts. Check whether the execution bandwidth used is less than the allocated bandwidth. If the current bandwidth is less than the allocated bandwidth even after reducing the total amount of packet discard, set the packet discard amount to 0 for the corresponding bandwidth control unit 208, 210, 212, and proceed to the aggregation condition review process. . If all the current packet discards are reduced and the allocated bandwidth is exceeded, the maximum reduction amount that is less than or equal to the allocated bandwidth is obtained, and the current packet discard amount is calculated for the corresponding bandwidth control units 208, 210, and 212. The remaining amount obtained by subtracting the maximum reduction amount is set as the packet discard amount. In this case, the aggregation condition review process is passed.

  In the aggregation condition review process in step S110, the packet processing unit 205 obtains the minimum packet size that can make the actual effective bandwidth equal to or less than the allocated bandwidth, and if the obtained minimum packet size is smaller than the average packet size before aggregation, Cancel execution of aggregation. If the minimum packet size is larger than the average packet size before aggregation, an aggregation condition is calculated and set so that the minimum packet size becomes the packet size after aggregation.

  Packets belonging to the flow subject to aggregation by the processing of the packet processing unit 205 are sent from the packet processing unit 205 to the corresponding aggregation processing units 207, 209, and 211, and a plurality of packets are aggregated into one packet according to the aggregation conditions. Are sent to the bandwidth control units 208, 210, and 212. On the other hand, packets belonging to flows that are not to be aggregated are sent from the packet processing unit 205 to the corresponding bandwidth control units 208, 210, and 212 as they are. The bandwidth control units 208, 210, and 212 transmit packets belonging to the flow to other communication devices through the corresponding communication units 202 to 204. If the packet discard amount is set from the packet processing unit 205, the bandwidth control units 208, 210, and 212 The discarded amount of packets is discarded from packets belonging to the flow.

  As a result of the above processing, even when the effective bandwidth actually used exceeds the allocated bandwidth, the effective bandwidth actually used is kept below the allocated bandwidth without discarding packets as much as possible. Can be suppressed.

  Next, a description will be given of a method for calculating the minimum aggregation condition necessary to keep the effective bandwidth actually used below the allocated bandwidth.

  Before the explanation, the premise idea is shown.

The time required for sending data with a packet size of P can be expressed as follows.
DIFS + DATA_RATE (P) + SIFS + Tr (ACK)… (2)
Here, DATA_RATE (P) represents the time required to send data of packet size P when the communication rate in the wireless section is RATE Mbps. Further, DIFS and SIFS represent the waiting time between frames, and Tr (ACK) represents the time required for transmitting the ACK frame.

As described above, the overhead α (P, P_Set) that compares the flow of the packet size P with the flow of the packet size P_Set when the communication rate in the wireless section is RATE Mbps can be obtained as follows.
α (P, P_Set) = (DIFS + DATA_RATE (P) + SIFS + Tr (ACK)) / (DIFS + DATA_RATE (P_Set) + SIFS + Tr (ACK)) * (P_Set / P)… (3)

  The overhead α (P, P_Set) above is the communication time per bit for a flow with a packet size of P, assuming that the communication time per bit when transferring a flow with a packet size of P_Set is 1. This is a value indicating whether or not this is required.

  Based on the above, the packet processing unit 205 performs the following operation. Here, in order to simplify the explanation, an example of two flows will be described. However, in practice, a large number of flows are targeted, and even in this case, the same procedure shown below is performed.

  Now, when the communication rate of the wireless section is RATE Mbps, the packet size is P_Set, the effective bandwidth for BW Mbps is assigned to a certain terminal (for example, terminal 102) as a traffic class, and the average packet size is averaged by P_A Assume that a flow A with a bit rate of B_A bps and a flow B with an average packet size of P_B and an average bit rate of B_B bps belong to the class. At this time, the effective bandwidth used in the flow A is B_A * α (P_A, P_Set), and the effective bandwidth used in the flow B is B_B * α (P_B, P_Set).

When the effective bandwidth used by the entire flows A and B becomes traffic exceeding the allocated bandwidth due to traffic fluctuations, the relationship is as follows.
BW <B_A * α (P_A, P_Set) + B_B * α (P_B, P_Set)… (4)

Here, when P_B << MTU (maximum permissible packet length), in order to maintain the predetermined bit rates B_A and B_B while keeping the effective bandwidth used by the traffic class in the aggregation below the allocated effective bandwidth Therefore, it is necessary to set the packet size P_B ′ after aggregation to a size satisfying the following expression.
α (P_B ', P_Set) ≤ (BW− B_A * α (P_A, P_Set)) / B_B… (5)

For simplicity,
OH = DIFS + SIFS + Tr (ACK)… (6)
DATA_RATE (P) = P / RATE… (7)
Α (P_B ', P_Set) is
α (P_B ', P_Set) = (OH + P_B' / RATE) / (OH + P_Set / RATE) * (P_Set / P_B ')… (8)
And solve for P_B '
P_B '= P_Set * OH / (OH * α (P_B', P_Set) + P_Set / RATE * (α (P_B ', P_Set) − 1))… (9)
It becomes. Since the upper limit value of α (P_B ′, P_Set) is obtained from the inequality (5), P_B ′ can be obtained by substituting the value of the upper limit value.

By performing the aggregation so that the obtained P_B ′ is obtained, the effective bandwidth used by the traffic class becomes equal to or less than the bandwidth allocated to the traffic class. Then, the waiting time for aggregation is obtained from P_B ′ and B_B by the following equation.
Wait time = P_B '/ B_B (10)

  If the obtained P_B ′ is equal to or greater than the MTU, the same is performed for the flow A using the MTU as the P_B ′. If it still exceeds, the bandwidth control unit 210 sets the packet discard amount so that the excess is discarded.

  As described above, it is possible to realize bandwidth control of the effective bandwidth in consideration of the packet size and dynamic aggregation according to the status of the effective bandwidth used by the traffic.

  When performing the above processing, overhead α (P, P_Set) needs to be calculated everywhere. Although this may be obtained by calculation every time, data shown in a graph as shown in FIG. 5 is stored in advance in a storage device, and the calculation can be omitted by referring to this data. This method can reduce the processing load by using a relay node having a small number of computing resources. The graph of FIG. 5 shows the communication time per bit (that is, overhead α) at each packet size when the communication time per bit when transferring the flow when the packet size is 1450 bytes is 1. The communication rates in the wireless section are obtained and graphed for 54 Mbps, 36 Mbps, and 11 Mbps. Incidentally, since the influence of the overhead on the throughput increases as the wireless communication rate increases, the value of the overhead α increases as the communication rate increases even with the same packet size.

  In contrast to the above method, when a value is actually entered into the formula, the result is as follows.

  Assume that in a wireless LAN communicating at a communication rate of 54 Mbps in a wireless section, an effective bandwidth of 10 Mbps (packet size converted to 1450 bytes) is allocated to a certain traffic class.

  Here, if the flow with an average packet size of 1450 bytes is 7 Mbps and the flow with an average packet size of 200 bytes is 1 Mbps, the bandwidth used in the packet processing unit 205 is considered in consideration of the packet size. The total bandwidth is 11 Mbps.

  The packet processing unit 205 checks that the effective size of a flow with an average packet size of 200 bytes is 3 Mbps or less and that the effective size is 2 Mbps if the packet size is at least 400 bytes. Finally, from 400 * 8/10 ^ 6 = 0.0032, the aggregation waiting time in the aggregation processing unit is set to 3.2 msec.

  In the above description, in the situation where the communication rate in the communication section does not change, the application example of the present invention in the case where the amount of traffic becomes equal to or greater than the allocated bandwidth, but as a result of the change in the communication rate in the communication section, traffic The present invention can also be applied as described below even when the amount of data exceeds the allocated bandwidth.

  In many cases, the communication rate of the wireless section varies depending on the communication quality of the wireless section. In such a case, in the bandwidth allocation using the effective bandwidth, the bit rate that can be realized even with the same allocated bandwidth varies according to the variation of the communication rate in the wireless section. Even when the communication rate changes, by applying the present invention, it is possible to maintain the bit rate that has been communicated up to now within the allocated effective bandwidth.

  Now, when the communication rate of the wireless section is RATE Mbps, the BW Mbps bandwidth (packet size as P_Set) is assigned to a traffic class, the average packet size is P_B and the average bit rate is Assume that the flow of B_B bps belongs.

Here, the communication rate of the wireless section changes from RATE Mbps to RATE 'Mbps due to changes in the wireless environment, and the bit rate that can be realized with the effective bandwidth allocated accordingly changes from BW Mbps to BW' Mbps, When the bit rate is equal to or higher than BW ′, it is necessary to perform aggregation so that the packet size becomes P_B ′ that satisfies the following expression.
BW '/ BW = (OH + P_B' / RATE ') / (OH + P_Set / RATE) * (P_Set / P_B')… (11)

So, solving for P_B ',
P_B '= P_Set * OH / (BW' / BW (OH + P_Set / RATE)-P_Set / RATE ')… (12)
Thus, the obtained packet size after aggregation is necessary to maintain the bit rate within the effective band to which the obtained P_B ′ is allocated. The waiting time in the aggregation processing unit is obtained from the above equation (10) from P_B ′ and B_B.

  As a result, even when the communication rate changes, it is possible to maintain the bit rate at the time of bandwidth allocation within the effective bandwidth range originally secured by aggregation.

  Here, it is an example when the communication rate of the relay node's wireless section changes, but even if the effective bandwidth that can be secured by the relay node fluctuates due to other competing wireless devices, based on the secured effective bandwidth The same processing as shown in the previous example can be applied.

  Next, the effect of this embodiment will be described.

  The first effect is that the effective bandwidth actually used by the traffic can be grasped. Therefore, even when flows of different packet sizes are mixed, which is not possible with the bandwidth control based on the bit rate, the actually used bandwidth is grasped. Be able to. Thereby, bandwidth control based on the effective bandwidth to be used can be performed.

  The second effect is that even when the band to be used is equal to or greater than the allocated band, the effective bit band to be used can be maintained within the allocated band by performing aggregation to maintain the maximum bit rate.

  A third effect is that even when the throughput that can be realized within a band assigned to a certain traffic changes due to a change in the communication rate of the communication section, it is possible to perform bandwidth control that maintains the throughput according to the situation. The reason is that aggregation is performed according to the bandwidth that can be used for traffic.

  The fourth effect is that each relay node operates according to the situation where each relay node is placed, and each relay node performs autonomously. Therefore, the total throughput is achieved by aggregation according to the situation without exchanging information such as signaling. It is possible to improve.

"Other embodiments"
In the relay node 201 illustrated in FIG. 3, a set of bandwidth control unit and aggregation processing unit is provided for each of the communication units 202 to 204, but the relay node includes a communication unit that does not have the bandwidth control unit and the aggregation processing unit. However, the present invention is also applicable. The configuration in that case is shown in FIG. The relay node 221 in this example has a configuration in which the aggregation processing unit 211 and the bandwidth control unit 212 are removed from the relay node 201 in FIG. In the relay node 221, data received from the plurality of communication units 202 to 204 is processed by the packet processing unit 205, and the transfer destinations sent through the communication units 202 and 203 are aggregated by the aggregation processing units 207 and 209. Band control is performed by the band control units 208 and 210. On the other hand, data transferred via the communication unit 204 having no aggregation processing unit and bandwidth control unit is transferred as it is.

  In the first embodiment, the mechanism shown in FIG. 3 for controlling the actual effective bandwidth to be equal to or lower than the allocated bandwidth using aggregation is incorporated in the relay nodes 103 and 104 in the network shown in FIG. It can also be incorporated into a terminal. In this case, regarding the traffic generated in the terminal, the terminal itself assigns a band to the band management unit and is processed according to the procedure of the flowchart shown in FIG.

  In the first embodiment, the present invention is applied to a network having two relay nodes 103 and 104 as shown in FIG. 1, but the number of relay nodes and terminals included in the network is arbitrary. For example, as shown in FIG. 7, the present invention is also applicable to a network in which communication is performed between terminals through three or more relay nodes 103, 104, and 107.

  As a method of allocating a bandwidth to a flow or a traffic class required in the present invention, in addition to a reservation method using RSVP (Resource Reservation Protocol), a communication apparatus and a relay node that implement the present invention in advance. This can also be realized by a method in which allocation information is set for.

It is a block diagram which shows the network structural example to which this invention is applied. It is an operation | movement sequence diagram of the network to which this invention is applied. It is a block diagram which shows an example of the relay node to which this invention is applied. It is a flowchart which shows the process example of the packet process part in the relay node to which this invention is applied. It is a figure which shows the overhead of each packet size when the communication time per bit concerning the transfer of the flow whose packet size is 1450 bytes is set to 1. It is a block diagram which shows the other example of the relay node to which this invention is applied. It is a block diagram which shows another network structural example to which this invention is applied.

Explanation of symbols

101,102,105,106 ... terminal
103, 104, 107, 201, 221 ... Relay node
202 ~ 204 ... Communication Department
205: Packet processor
206… Bandwidth management unit
207, 209, 211 ... Aggregation processor
208, 210, 212 ... Band control unit

Claims (20)

Aggregation processing means for aggregating a plurality of packets flowing in one or more flows into one according to the set aggregation condition;
A packet processing unit configured to measure an actual bandwidth used for the one or more flows and to set an aggregation condition in the aggregation processing unit in order to reduce the bandwidth used when the bandwidth used exceeds the allocated bandwidth. A communication device characterized by the above.   The packet processing means calculates a minimum amount of bandwidth reduction required to keep the used bandwidth of the one or more flows below the allocated bandwidth, and maintains the average bit rate of the one or more flows while maintaining the current state. The packet size necessary for reducing the used bandwidth by aggregation by an amount equal to the calculated bandwidth reduction amount is calculated, and the aggregation condition is calculated from the calculated packet size and set. Communication equipment. Bandwidth control means for performing bandwidth control of the one or more flows by discarding packets;
When the packet processing means cannot reduce the used bandwidth to less than the allocated bandwidth by merely reducing the used bandwidth by aggregation, the packet processing means may cause the bandwidth control means to discard a certain amount or a certain ratio of packets in order to reduce the used bandwidth. The communication apparatus according to claim 1 or 2, characterized in that:   The packet processing means performs a review of a packet discard amount by the bandwidth control means and an aggregation condition of the aggregation processing means when a use bandwidth of the one or more flows decreases below the allocated bandwidth. The communication apparatus according to any one of claims 1 to 3.   In the communication section of the communication rate RATE, the communication time per bit when transferring the flow of the packet size P when the communication time per bit when transferring the flow of the packet size P_Set is 1 is the overhead α (P, P_Set), and when the BW effective bandwidth is assigned to the one or more flows as the communication rate RATE and the packet size P_Set, the packet processing means By measuring the average bit rate B, calculating the effective bandwidth actually used in the one or more flows from the measurement result and the overhead α (P, P_Set), and comparing with the allocated bandwidth BW The communication apparatus according to claim 1, wherein it is determined whether or not a used band exceeds an allocated band.   Storage means for storing the calculation result of the overhead α (P, P_Set) at each packet size when the communication time per bit when transferring the flow of the packet size P_Set is 1, and the packet processing 6. The communication apparatus according to claim 5, wherein means reads out and uses the overhead α (P, P_Set) necessary for calculating the effective bandwidth from the storage means.   The communication apparatus according to claim 1, which has a function of transferring packets transmitted and received between a plurality of other communication apparatuses.   The communication apparatus according to claim 7, wherein the communication apparatus is wirelessly connected to the other communication apparatus. a) the packet processing means measuring the actual bandwidth used for one or more flows, and if the bandwidth used exceeds the allocated bandwidth, setting an aggregation condition in the aggregation processing means to reduce the bandwidth used; ,
b) The aggregation processing means includes a step of aggregating a plurality of packets flowing in the one or more flows into one in accordance with a set aggregation condition.   The packet processing means calculates a minimum amount of bandwidth reduction required to keep the used bandwidth of the one or more flows below the allocated bandwidth, and maintains the average bit rate of the one or more flows while maintaining the current state. 10. The packet size necessary for reducing the bandwidth used by aggregation by an amount equal to the calculated bandwidth reduction amount is calculated, and an aggregation condition is calculated from the calculated packet size and set. Communication method.   c) The bandwidth control means includes a step of discarding a certain amount or a certain percentage of packets in order to reduce the used bandwidth when the bandwidth cannot be suppressed below the allocated bandwidth only by reducing the used bandwidth by the aggregation. The communication method according to claim 9 or 10.   d) The step of the packet processing means revising the discard amount of packets by the bandwidth control means and the aggregation conditions of the aggregation processing means when the use bandwidth of the one or more flows decreases below the allocated bandwidth. The communication method according to claim 9, wherein the communication method is included.   In the communication section of the communication rate RATE, the communication time per bit when transferring the flow of the packet size P when the communication time per bit when transferring the flow of the packet size P_Set is 1 is the overhead α (P, P_Set), and when the BW effective bandwidth is assigned to the one or more flows as the communication rate RATE and the packet size P_Set, the packet processing means By measuring the average bit rate B, calculating the effective bandwidth actually used in the one or more flows from the measurement result and the overhead α (P, P_Set), and comparing with the allocated bandwidth BW The communication method according to claim 9, wherein it is determined whether or not the used band exceeds the allocated band.   Storage means for storing the calculation result of the overhead α (P, P_Set) at each packet size when the communication time per bit when transferring the flow of the packet size P_Set is 1, and the packet processing 14. The communication method according to claim 13, wherein means reads out the overhead α (P, P_Set) necessary for calculating the effective bandwidth from the storage means and uses it. Computer
Aggregation processing means for aggregating a plurality of packets flowing in one or more flows into one according to the set aggregation condition;
In order to measure the actual bandwidth used for the one or more flows and, when the bandwidth used exceeds the allocated bandwidth, to cause the aggregation processing means to function as a packet processing means for setting an aggregation condition in order to reduce the bandwidth used Program.   The packet processing means calculates a minimum amount of bandwidth reduction required to keep the used bandwidth of the one or more flows below the allocated bandwidth, and maintains the average bit rate of the one or more flows while maintaining the current state. 16. The packet size required for reducing the bandwidth used by aggregation by an amount equal to the calculated bandwidth reduction amount is calculated, and an aggregation condition is calculated and set from the calculated packet size. Program. Further causing the computer to function as bandwidth control means for performing bandwidth control of the one or more flows by discarding packets; and
When the packet processing means cannot reduce the used bandwidth to less than the allocated bandwidth by merely reducing the used bandwidth by aggregation, the packet processing means may cause the bandwidth control means to discard a certain amount or a certain ratio of packets in order to reduce the used bandwidth. The program according to claim 15 or 16, characterized in that   The packet processing means performs a review of a packet discard amount by the bandwidth control means and an aggregation condition of the aggregation processing means when a use bandwidth of the one or more flows decreases below the allocated bandwidth. The program according to any one of claims 15 to 17.   In the communication section of the communication rate RATE, the communication time per bit when transferring the flow of the packet size P when the communication time per bit when transferring the flow of the packet size P_Set is 1 is the overhead α (P, P_Set), and when the BW effective bandwidth is assigned to the one or more flows as the communication rate RATE and the packet size P_Set, the packet processing means By measuring the average bit rate B, calculating the effective bandwidth actually used in the one or more flows from the measurement result and the overhead α (P, P_Set), and comparing with the allocated bandwidth BW The program according to any one of claims 15 to 18, wherein it is determined whether or not a used band exceeds an allocated band.   Storage means for storing the calculation result of the overhead α (P, P_Set) at each packet size when the communication time per bit when transferring the flow of the packet size P_Set is 1, and the packet processing The program according to claim 19, wherein means reads out and uses the overhead α (P, P_Set) necessary for calculating the effective bandwidth from the storage means.

Images