JP2001197063A - Flow control method and communication element implementing same - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a flow control method which can guarantee a band of a minimum cell rate (MCR) with simple constitution by reducing a burst increase in the traffic in an ATM communication network and a communication element which actualizes it. SOLUTION: An MCR decision part 1109 and a cell buffer decision part 1105 of the communication element 1100 decides the MCR of the VC of arriving cells and sends the cells to one of cell buffers 1101-1 to 1101-n according to the MCR. Queue length management part 1102-1 to 1102-n, congestion decision parts 1103-1 to 1103-n, and cell discarding parts 1104-1 to 1104-n perform congestion control over the cell buffers. A transmission order control part 1106, a transmission timing determination part 1107, and an active buffer management part 1108 sends cells from the cell buffers so that a band wider that the band predetermined for the respective buffers is allocated.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a flow control method and a communication element for executing the flow control method,
ATM (Asy) for transmitting multimedia information in a unified manner
GFR (Guaranteed Frame) in an nchronous Transfer Mode (NFC) communication network
e Rate) A flow control method performed on a VC of a service category and a communication element for executing the method.

[0002]

2. Description of the Related Art In recent years, as a service category in an ATM communication network, a minimum cell rate (MCR: Minim) has been proposed.
um Cell Rate) is defined, and a GFR service category is defined so that network resources can be effectively used. For the definition of the GFR service category, see, for example, The AT
M Forum, "ATM Forum / BTD-TM
-01.03: Traffic Management
Baseline Text Document ".

In this GFR service category, ATM
Flow control is performed in a plurality of communication elements (for example, an exchange and a multiplexing device) constituting a communication network. Bandwidth is guaranteed so that the cell input speed of each VC for which flow control is performed is always equal to or higher than MCR. When there is a vacancy in the transmission band, the vacant band is distributed to each VC. Such a flow control can be realized by, for example, a first or second conventional technique described below.

(First Prior Art) First, a first prior art will be described. The first prior art is, for example, "AT
M Forum 97-0310: TCP over
GFR Implementation with
Different Service Discipl
ines: A Simulation Study "
Are disclosed.

In the first prior art, each VC
The magnitude of the input cell rate input to the network from the (virtual channel) is observed, and tagging is performed on cells exceeding the minimum guaranteed bandwidth. Each cell element is provided with one cell buffer in the multiplexing part. If congestion has occurred in this cell buffer, the cell being tagged is preferentially discarded. Hereinafter, the configuration and operation of the first related art will be described in detail.

FIG. 8 is a block diagram showing a configuration example of a communication network using the first prior art. In FIG. 8, the communication network includes communication elements 100 and 110 and a transmitting terminal 12.
0, 130, and 140, and a packet policer CLP bit setting unit 150. The communication element 100 includes a cell buffer 101, a queue length management unit 102, a congestion determination unit 103, a CLP bit observation unit 104, and a cell discard unit 105. The communication elements 100 and 110 are connected to each other by a communication path. Also, the transmitting terminal 12
0, 130 and 140 are communication elements 1
00 is connected.

In this communication network, transmitting terminals 120 and 13
0 and 140 are transmitting information via the communication element 100 to a receiving terminal (not shown) provided in the direction of the communication element 110. The information is transferred in the form of a cell. These cells are multiplexed by a cell buffer 101 provided in the communication element 100 and output to the communication element 110.

Although the present communication network is based on the premise that there is only one multiplexing form, the same configuration is required for portions where congestion occurs in which cells are multiplexed. Also, packet policer CLP bit setting section 150
Is a UNI (User Network Interf.)
ace).

The packet policer CLP bit setting unit 15
0 indicates whether or not the input rate when cells are input from each VC to the network exceeds the MCR. In this observation, generally, F-GCRA (Fr
ame based Generic Control
An algorithm called Rate Algorithm is used. For more information on F-GCRA,
"ATM Forum / BTD-TM-01.03: T
Rafic Management Baselin
e Text Document ”, p13,
It is disclosed in detail.

Cells determined to exceed the MCR by such an algorithm are identified by the packet policer C
The CLP bit is set to 1 by LP bit setting section 150. The CLP bit setting operation by such a network is called tagging.

Next, the operation of the communication element 100 will be described. The cell discard unit 105 manages a cell discard determination flag. When detecting the arrival of the cell, the cell discarding unit 105 transmits a cell arrival signal to the CLP bit observing unit 104. Upon receiving the cell arrival signal, CLP bit observing section 104 observes the value of the CLP bit of the cell and transmits the value to cell discarding section 105.

[0012] The cell discarding unit 105 includes a congestion determining unit 1.
When the congestion determination signal is received from the control unit 03, the value of the cell discard determination flag is set to 1. When receiving the non-congestion determination signal, cell discarding section 105 sets the value of the cell discard determination flag to 0. Note that the value of the cell discard determination flag is set to 0 when the communication element operation starts.

When receiving the value of the CLP bit from CLP bit observing section 104, cell discarding section 105 takes the logical product of the value and the value of the cell discard determination flag. When the result of the logical product is 1, the cell discarding unit 105 discards the cell. When the result of the logical product is 0, the cell discarding unit 105 transmits the cell to the cell buffer 101.

The cell buffer 101 includes a cell discarding unit 105
When a cell arrives from the terminal, the cell is stored and a cell arrival signal is transmitted to the queue length management unit 102. The cell buffer 101 transmits the stored cells to the communication path at a cell transmission timing with a fixed time interval.

Here, the transmission of the stored cells is performed in the same order as the arrival order of the cells in the cell buffer 101. Therefore, the cell buffer 101 has a FIFO
(First In First Out) method is used. At the cell transmission timing, the cell buffer 1
01, there is no cell in the cell buffer 10
1 waits until the next cell transmission timing without performing cell transmission. When transmitting a cell, the cell buffer 101 transmits a cell transmission signal to the queue length management unit 102.

The queue length management unit 102 manages the queue length. The queue length indicates the data length stored in the corresponding cell buffer. Queue length management unit 102
Increases the queue length by one upon receiving a cell arrival signal. After that, the queue length management unit 102 outputs the queue length to the congestion determination unit 103. Also, upon receiving the cell transmission signal, the queue length management unit 102 sets the queue length to 1
Only decrease. After that, the queue length management unit 102 outputs the queue length to the congestion determination unit 103. The queue length is set to 0 when the operation of the communication element starts.

The congestion judging section 103 stores a congestion judgment threshold value. The congestion determination unit 103 is a queue length management unit 1
When the queue length is input from 02, the queue length is compared with the congestion determination threshold value. As a result of the comparison, if the queue length is larger, the congestion determination unit 103 determines that congestion is occurring, and transmits a congestion determination signal to the cell discard unit 105. As a result of the comparison, when the queue length is smaller, the congestion determination unit 103 determines that congestion is not occurring, and transmits a non-congestion determination signal to the cell discard unit 105.

(Second Prior Art) Next, a second prior art will be described. Regarding the second prior art, see, for example, “ATM Forum 97-1035: AS
immulation Study of GFR Im
prementation ".

In the second prior art, a cell buffer is provided for each VC in a multiplex portion of a communication element, and a transmission order when cells are transmitted from each cell buffer is controlled. By such control, the output rate value of each cell buffer is set to be equal to or more than the MCR value of the VC. In addition, when congestion occurs in a certain cell buffer, a part of cells arriving at the cell buffer is discarded. Hereinafter, the configuration and operation of the second conventional technique will be described in detail.

FIG. 9 is a block diagram showing an example of a communication network using the second prior art. In FIG. 9, the communication network includes communication elements 200 and 210 and a transmission terminal 22.
0, 230 and 240. The communication elements 200 and 210 are connected to each other via a communication path.
Transmission terminals 220, 230 and 240 are connected to communication element 200 via a communication path.

The communication element 200 includes n cell buffers 2
01-1 to 201-n and n queue length management units 202
−1 to 202-n and n congestion determination units 203-1 to 203-2
03-n and n cell discard units 204-1 to 204-n
, VC determination section 205, transmission order control section 206, transmission timing determination section 207, active VC management section 2
08.

In this communication network, transmission terminals 220, 2
30 and 240 are transmitting information via the communication element 200 to a receiving terminal (not shown) provided in the direction of the communication element 210. The information is transferred in the form of a cell. Each of these cells is a communication element 20 depending on the VC.
It is stored in each cell buffer provided in 0. Thereafter, the stored cells are multiplexed on the communication path and transmitted to the communication element 210.

The communication network shown in FIG. 9 is based on the assumption that only one multiplexing form is used, as in the communication network shown in FIG. 8. However, the communication network shown in FIG. Configuration is required respectively.

Next, the operation of the communication element 200 will be described. When a cell arrives, the VC determination unit 205 determines the VC number from the value of the VCI field and the value of the VPI field of the cell. According to the determined VC number, V
C determining section 205 transmits the cell from cell discarding section 204-1 to any of cell discarding sections 204-n.

Here, in the communication network of FIG. 9, the total number of VCs is n or less. Therefore, for each VC, there are n sets of one cell discarding unit, one cell buffer, one queue length managing unit, and one congestion judging unit, for a total of n
Each is prepared. Here, each cell discard unit, each cell buffer, each queue length management unit and each congestion determination unit,
The same operation is performed for each VC. Therefore, only the operation for one VC is described below.
Instead of all explanations.

The cell discard unit 204-1 manages a cell discard flag. When a cell arrives, the cell discard unit 204-1 reads the value of the cell discard flag. When the value of the cell discard flag is 1, the cell discard unit 204-1
Discard the cell. When the value of the cell discard flag is 0, the cell discard unit 204-1 transmits the cell to the cell buffer 201-1. Note that the value of the cell discard determination flag is set to 0 when the communication element operation starts.

When receiving the congestion judgment signal from congestion judgment section 203-1, cell discarding section 204-1 sets the value of the cell discarding judgment flag to 1. Also, the cell discard unit 2
When receiving the non-congestion determination signal from the congestion determination unit 203-1, the 04-1 sets the value of the cell discard determination flag to 0.

When a cell arrives, the cell buffer 201-1 stores the cell and transmits a cell arrival signal to the queue length management unit 202-1. Also, the cell buffer 201
-1, upon receiving the cell transmission command signal, transmits the cell to the communication path. Cell transmission is performed in the same order as the arrival order to the cell buffer 201-1. Therefore, as in the case of FIG. 8, the FIFO system is used for the cell buffer 201-1 in FIG. Also, cell buffer 2
01-1 transmits a cell transmission signal to the queue length management unit 202-1 when transmitting a cell.

The transmission timing determining section 207 generates a transmission timing at a fixed time interval inside. Then, the transmission timing determination unit 207 outputs a transmission timing signal to the transmission order control unit 206 in accordance with the timing. Specifically, the transmission timing signal is output to a transmission VC number management unit 2068 described later provided inside transmission order control unit 206.

The transmission order control unit 206 manages transmission VC numbers. When receiving the transmission timing signal, transmission order control section 206 transmits a cell transmission command signal to the cell buffer corresponding to the transmission VC number. However, when the transmission VC number is set to 0, transmission order control section 206 does not transmit a cell transmission command signal. The configuration of the transmission order control unit 206, the operation of determining the transmission VC number, and the like will be described later.

The queue length management unit 202-1 manages the queue length. Upon receiving the cell arrival signal, the queue length management unit 202-1 increases the queue length by one. After that, the queue length management unit 202-1 outputs the queue length to the congestion determination unit 203-1 and outputs the active signal and the corresponding VC number to the active VC management unit 208.

When receiving the cell transmission signal, the queue length management unit 202-1 reduces the queue length by one.
Thereafter, the queue length management unit 202-1 outputs the queue length to the congestion determination unit 203, and when the queue length is 0,
An inactive signal and a corresponding VC number are output to active VC management section 208. Note that the queue length is set to 0 at the start of operation of the communication element.

The active VC management unit 208 manages whether each VC is an active VC by using an active flag. Here, the active VC means a VC in which a cell is input from a transmitting terminal and cells are stored in a cell buffer.

When receiving the active signal and the VC number, the active VC management unit 208 sets the value of the active flag corresponding to the corresponding VC to 1. Further, when receiving the inactive signal and the VC number, the active VC management unit 208 sets the value of the active flag corresponding to the corresponding VC to 0.

The congestion judgment unit 203-1 stores a congestion judgment threshold value. When the queue length is input, the congestion determination unit 203-1 compares the value of the queue length with the magnitude of the congestion determination threshold. As a result of the comparison, if the queue length is larger, the congestion determination unit 203-1 determines that congestion is occurring, and sends a congestion determination signal to the cell discard unit 2.
Send to 04-1. As a result of the comparison, when the queue length is smaller, the congestion determination unit 203-1 determines that the state is not congested, and transmits a non-congestion determination signal to the cell discard unit 204-1. .

The transmission order determining algorithm used in the transmission order control unit 206 is a weighted round robin (WRR: Weighted Round Robi).
A method called n) is typical. By using this WRR, each VC is weighted in proportion to the corresponding MCR. Therefore, the transmission band of the communication path is distributed among the active VCs in proportion to the respective weights. Note that, as described above, the active VC here also means a VC in which a cell is input from the transmitting terminal and cells are stored in the cell buffer.

For the WRR, for example, “Weigh
ted Round-RobinCell Multi
plexing in a General-Purp
once ATM Switch Chip, "IEEE
Journal on Selected Areas
in Communications, Vol.
9, No. 8, October 1991.

Next, a configuration example and an operation example of the transmission order control unit 206 for executing the above-described algorithm will be described with reference to FIG.
Shown in In FIG. 10, the transmission order control unit 206
C weight management table 2061, VC weight division element table 2062, weight counter unit 2063, weight determination unit 2064, transmission reservation VC number management unit 206
7 and a transmission VC number management unit 2068.

The VC weight management table 2061 stores each VC
Manages the weight of These weights are integer values, and values are selected so as to be proportional to the MCR. The VC weight division element table 2062 manages the weight division elements of each VC.

Here, the weight division element of each VC is determined by each VC managed by the VC weight management table 2061.
Are calculated as follows from the weights of, and stored in the VC weight division element table 2062.

Assuming that the weight of a certain VC is N, the weight division element is calculated as indicating each term on the right side when N is represented by the sum of integers of powers of two. That is, a certain V
Assuming that the weight of C is N, N = 2 ** (i1) + 2 **
Each term on the right side when expressed by (i2) +2 ** (i3) +... +2 ** (in) is defined as a weight division element. Here, A ** B represents A to the Bth power, and i1 to in
Represents an integer of 0 or more.

For example, consider a method of calculating a weight division element when a certain VC has a weight of 5. In this case, since N = 5, from the above equation, 5 = 2 ** 0 + 2 *
* 2 = 1 + 4. Therefore, if the weight of a certain VC is 5, the weight dividing element is 1
And 4 are calculated.

The reason for dividing the weights in this way is that the weighting counter unit 2063 performs the scheduling using the power of 2 weight. Hereinafter, the operation of the weighting counter unit 2063 will be described.
This will be described in detail.

The weight counter 2063 manages the counter increment and the counter value. The counter value is used to calculate a weight value, and is 1 to 2 ** m−
In the range up to 1, the count is repeated. The counter increment indicates how much the counter value has been counted up since the counter increment reset signal was received. At the start of operation of the communication element, the counter value is set to 1 and the counter increment is set to 0.

First, the weighting counter unit 2063 determines whether the VC weight division element table 206
2 or a count-up signal from the transmission reservation VC number management unit 2067. Upon receiving the count-up signal, the weighting counter unit 2063 refers to the counter increase number. Here, when the value of the referred counter increment is equal to 2 ** m-1, the counter increment is 2 *
* It means that it continued to increase without being reset until it reached m-1. This means that the counter value is 1 to 2 *
* It means that the range up to m-1 has been counted up for exactly one lap. Weight counter unit 2063
Detects this state, resets the counter increment number to 0, and then transmits the transmission V number to the transmission VC number management unit 2068.
Outputs C number reset signal.

Also, the number of increments of the referred counter is 2 ** m
If it is smaller than -1, the weight counter unit 2063
Increments the counter increment and the counter value by one. However, the counter value before the increase is 2 ** m-1
, The weight counter unit 2063 returns the counter value to 1. After that, the weight counter unit 2063
Outputs the counter value to the weight determination unit 2064.

Further, the weight counter 2063
In a predetermined case described later, a counter increment count reset signal is received from the active VC management unit 208 or the transmission VC number management unit 2068. Upon receiving the counter increment count reset signal, the weighting counter unit 2063 resets the counter increment count to zero.

Upon receiving the counter value from the weight counter unit 2063, the weight determination unit 2064 converts the value into a binary number. Next, the weight determination unit 2064
The lowest bit number whose value is 1 (hereinafter, referred to as a non-zero least significant bit) is calculated for each digit of the binary value.

For example, it is assumed that the counter value received by weight determining section 2064 is 5. This value is 101 when converted to a binary number. Therefore, the non-zero least significant bit is 1. Considering another example, it is assumed that, for example, the counter value received by the weight determination unit 2064 is 12. This value is 1100 when converted to a binary number. Therefore, the non-zero least significant bit is 3.

Incidentally, the maximum value of the counter value is 2 **
m-1. If this value is converted to a binary number, it can be represented by m bits. Therefore, the maximum value of the non-zero least significant bit is m. Using this binary number represented by m bits, scheduling using a power of 2 weight is performed.

After calculating the non-zero least significant bit value, the weight determination unit 2064 calculates the non-zero least significant bit reverse value. The non-zero least significant bit reverse value is
(M + 1-non-zero least significant bit value). The reason why the non-zero least significant bit reverse value is calculated in this way is to use the fact that the magnitude of the non-zero least significant bit reverse value is inverted with the magnitude of the non-zero least significant bit reverse value. Of course, the minimum value of the non-zero least significant bit reverse value is 1, and the maximum value is m.

Further, the weight determination section 2064 calculates a weight determination value from the non-zero least significant bit reverse value calculated as described above. The weight determination value is (2
** (Non-zero least significant bit reverse-1)). Here, the counter value is 1 to 2 ** m-1
While counting up to
1, 2, 2 ** 2, ..., 2 ** (m-1). This weight determination value matches the number of times the counter value is counted up. Therefore, by using the number of times that the counter value is counted up, cell transmission can be scheduled according to the weight of each VC in the following manner.

The weight determination section 2064 calculates the weight determination value as described above. The calculated weight determination value is transmitted from the weight determination unit 2064 to the VC weight division element table 2062.

The VC weight division element table 2062 receives the weight judgment value, and checks whether or not a weight division element having the same value as the received weight judgment value exists in the weight division elements managed by the VC weight division element table 2062. I do.

As a result of the comparison, if there is a weight division element having the same value as the weight determination value, the VC weight division element table 2062 determines all VC numbers of VCs including the weight division element, Reservation VC with the VC number
Output to the number management unit 2067. When there is no weight division element having the same value as the weight determination value, the VC weight division element table 2062 outputs a count-up signal to the weight counter unit 2063 in order to obtain the next weight determination value. .

The transmission reservation VC number management unit 2067
A transmission reservation flag corresponding to C is managed. Upon receiving the VC number from the VC weight division element table 2062,
The transmission reservation VC number management unit 2067 sets the transmission reservation flag of the VC corresponding to the received VC number to 1.

Further, the transmission reservation VC number management unit 2067
Determines the VC having the smallest VC number among the VCs for which the transmission reservation flag is set to 1. Then, the transmission reservation VC number management unit 2067 outputs the VC number and the active determination command signal to the active VC management unit 208. After outputting them, the transmission reservation VC number management unit 2067 sets the transmission reservation flag of the output VC number to 0.
Set to. At the start of operation of the communication element, the transmission reservation flags are all set to 0.

When receiving the active determination command signal and the VC number from transmission reserved VC number managing section 2067, active VC managing section 208 refers to the value of the active flag of the VC corresponding to the received VC number. Note that the active VC management unit 208 determines that a certain VC
As described above, whether or not C is managed by the value of the active flag.

If the value of the referenced active flag is 1, the active VC management unit 208
The number is output to transmission VC number management section 2068, and a counter increase count reset signal is transmitted to weight counter section 2063. When the value of the referenced active flag is 0, the active VC management unit 208
Sends VC recalculation command signal to transmission reservation VC number management unit 2067
Output to

Receiving the VC recalculation command signal from active VC management section 208, transmission reservation VC number management section 2067 checks again the VC whose transmission reservation flag is 1.
Further, the transmission reservation VC number management unit 2067 determines the VC having the smallest VC number among the VCs whose transmission reservation flag is 1, and, as described above, determines the VC number and the active determination command signal as the active VC management signal. Output to the unit 208. After outputting them, the transmission reservation VC number management unit 2067 sets the transmission reservation flag of the output VC number to 0.
Set to. The above operation is performed when the transmission reservation VC number management unit 2067 receives the VC recalculation command signal.
This operation is repeated until there is no longer any VC for which the transmission reservation flag is 1.

If a VC whose transmission reservation flag is 1 does not exist when the VC recalculation command signal is received, transmission reservation VC number management section 2067 outputs a count-up signal to weight counter section 2063. I do.

The transmission VC number management unit 2068 transmits the transmission VC number.
Manage numbers. The transmission VC number management unit 2068
When a VC number is received from the active VC management unit 208, the transmission VC number is updated so as to match the received VC number. When receiving the transmission VC number reset signal, the transmission VC number management unit 2068 resets the transmission VC number to 0. Further, upon receiving the transmission timing signal, transmission VC number management section 2068 transmits a cell transmission command signal to the cell buffer corresponding to the transmission VC number, as described above.

Thereafter, the transmission VC number management unit 2068
A counter increase number reset signal is transmitted to the weight counter unit 2063, and the transmission reservation VC number management unit 206
7, a VC recalculation command signal is transmitted. However,
When the transmission VC number is set to 0, the transmission VC number
The number management unit 2068 does not transmit the cell transmission command signal,
Only the transmission of the VC recalculation command signal to the transmission reservation VC number management unit 2067 is performed.

Next, the operation of the WRR algorithm in the second prior art will be described using a specific example. Here, as a specific example, a case where VC1 to VC9 are multiplexed in a communication element using the second related art is considered.

In the above case, M of VC1 to VC8
The CR is assumed to be 10 Mbps, and the MCR of the VC 9 is assumed to be 80 Mbps. It is also assumed that the cell input rate to each VC is sufficiently high and cells are always stored in the buffer of each VC. Further, it is assumed that the value of the weight counter changes in the range of 1 to 15. On the premise of the above, a case where the weight of VC1 to VC8 is 1 and the weight of VC9 is 8 is considered. In this case, when the weight division elements of these VCs are calculated as described above, the weight division elements of VC1 to VC8 are 1
And the weight division element of VC9 is 8.

FIG. 11 shows that the weighting counter is 1 to 15
6 is a table showing changes in each of the non-zero least significant bit, the non-zero least significant bit reverse value, and the weight determination value when changing in order, and the VC number at which cell transmission is performed at each time. The method of calculating these values is as described above.

In FIG. 11, when the value of the weight counter is 8, the non-zero least significant bit is calculated as 4,
The non-zero least significant bit reverse value is calculated as 1, and the weight determination value is calculated as 1. At this point, VC
The cells are sequentially transmitted one by one from the cell buffers 1 to VC8.

When the value of the weighting counter is 1, 3,
In the case of 5, 7, 9, 11, 13 and 15, the non-zero least significant bit is calculated as 1, the non-zero least significant bit reverse value is calculated as 4, and the weight determination value is calculated as 8. At these times, one cell is transmitted from the cell buffer of VC9.

FIG. 12 shows VC1 to VC9 in the communication path.
6 is a table showing the cell transmission times of FIG. In FIG.
At time 32, a change in the VC number at which cell transmission is performed is shown. However, at subsequent times, the change of the VC number is repeated in the same order. Specifically, the change of the VC number is repeated in the order from time 5 to time 20.

Referring to FIG. 12, time 13 to time 20
In, cells of VC9 are continuously transmitted. This is the same as after time 29. As described above, in the second related art, when VCs having different MCRs are multiplexed, it can be said that traffic burstiness increases.

[0071]

The above first and second prior arts have the following problems, respectively. The following describes the problems in the first prior art and the second problem.
The problems in the prior art will be described separately.

(Issues in the First Conventional Technique) First, there are three issues with the first conventional art. Hereinafter, those problems will be described.

The first problem is that according to the first prior art,
When the transmitting terminal sets the CLP bit to indicate the priority of the cell, the MCR cannot be guaranteed for each VC.

Specifically, according to the first prior art, M
CLP bits are set in cells exceeding the CR. As a result, when the network is congested, the cell in which the CLP bit is set is discarded, and the cell in which the CLP bit is not set is transmitted preferentially. that way,
MCR can be guaranteed. However, when the transmitting terminal sets the CLP bit to indicate the priority of the cell, in the multiplexing part of the communication element, the cell that exceeds the MCR and the cell that does not exceed the MCR may be distinguished by the CLP bit. Can not. Therefore, if the transmitting terminal sets the CLP bit to indicate the priority of the cell, MCR cannot be guaranteed.

The second problem is that according to the first prior art,
TCP (Transmission Control P)
protocol) and UDP (User
VC using Datagram Protocol)
When TCP and TCP coexist, VC using TCP protocol
Cannot be guaranteed for MCR.

Here, for the sake of explanation, it is assumed that VC1 using TCP and VC2 using UDP are multiplexed in a communication element using the first conventional technique using the GFR service category. I do. It is assumed that the transmission speed of the communication path is 100 Mbps.

FIG. 13 is a graph showing an example of a change in the input rate of each VC in the above case. Where VC
1 is input with a cell at a rate allowed in the TCP rate control, and the VC (Peak Ceak
It is assumed that cells are input at a rate equal to (ll Rate). It is also assumed that the PCR of VC2 is 80 Mbps.

As shown in FIG. 13, the input rate of VC1 oscillates within a range that takes a small value. If congestion occurs in the communication element, the tagged cells are discarded. For this reason, the input rate is greatly reduced by TCP flow control. Then, when the congestion is released, the input rate of VC1 increases, and congestion occurs again. By repeating such increase and decrease, VC1
The input rate oscillates at a small value. On the other hand, UDP does not perform flow control. Therefore, the input rate of VC2 is PCR, and is always constant.

FIG. 14 is a graph showing the change over time of the transmission rate in the communication path for each of VC1 and VC2. The transmission rate of each VC on the communication path is
The input rate of each VC is obtained by subtracting the decrease caused by cell discard during congestion. Therefore, the transmission rate of each VC on the communication path changes as shown in FIG.

As described above, VC and U using TCP
If the VC using the DP is mixed, the VC using the UDP monopolizes the transmission band of the communication path. Therefore, according to the first prior art, V using TCP
MCR cannot be guaranteed for C.

The third problem is that in the first prior art,
There are multiple VCs that use TCP, and those VCs
If the values of MCR corresponding to
MCR cannot be guaranteed for C.

Here, for the sake of explanation, it is assumed that there are VC1 to VC5 each using TCP, which are multiplexed in a communication element executing the first prior art. In this case, the MC of VC1 to VC4
R is 10 Mbps, and MCR of VC5 is 3
It is assumed that the transmission speed is 0 Mbps.

FIG. 15 is a graph showing an example of a change in the input rate of each VC in the above case. Here, it is assumed that all VCs input cells at an input rate allowed in TCP rate control.

As shown in FIG. 15, VC1 to VC5
The input rate fluctuates. If congestion occurs in the communication element, the tagged cells are discarded. For this reason, the input rate is greatly reduced by TCP flow control. When the congestion is released, the input rate of each VC increases, and congestion occurs again when the input rate exceeds 20 Mbps. By repeating such increase and decrease,
The input rate of each VC will oscillate within a range of about 20 Mbps or less.

FIG. 16 is a diagram showing the change over time of the transmission rate in the communication path of each VC. Each V in the communication path
The transmission rate of C is obtained by subtracting the decrease due to cell discard at the time of congestion from the input rate of each VC. Therefore, the transmission rate of each VC on the communication path changes as shown in FIG. Here, the MCR of VC5 is 30
It can be seen that the MCR of VC5 is not guaranteed as shown in FIG.

As described above, according to the first prior art,
There are multiple VCs that use TCP, and those VCs
If the MCR values corresponding to
MCR cannot be guaranteed for a VC with a large R.

(Problem in Second Conventional Technique) Next, the second conventional technique has two problems. Hereinafter, those problems will be described.

The first problem is that according to the second prior art,
Since it is necessary to provide a cell buffer for each VC and control the queue length of each buffer, the discarding of cells, and the transmission order between buffers, respectively, when the number of VCs is large, the device configuration is complicated. It is to become.

The second problem is that according to the second prior art,
When VCs having different weights are multiplexed, cell transmission is continuously performed for a certain VC, and as a result, traffic burstiness increases. When the traffic burstiness is large, congestion in the network is likely to occur, and problems such as an increase in the cell discard rate occur. When VCs having different weights are multiplexed, cell transmission is continuously performed for a certain VC, and the burstiness of traffic increases.
As described above.

Accordingly, a first object of the present invention is to provide a flow control method capable of guaranteeing MCR for each VC even when a transmitting terminal sets a CLP bit to indicate the priority of a cell, and to realize the flow control method. To provide a communication element.

Further, a second object of the present invention is to provide a flow which can guarantee MCR for a VC using a TCP protocol even when a VC using a TCP and a VC using a UDP are mixed. An object of the present invention is to provide a control method and a communication element for implementing the control method.

Further, a third object of the present invention is to enable MCR guarantee for each VC even when there are a plurality of VCs using TCP and the MCR values corresponding to the VCs are different from each other. To provide a simple flow control method and a communication element for realizing the same.

A fourth object of the present invention is to provide a cell buffer for each VC even when the number of VCs is large, and to determine the queue length of each buffer, the discard of cells, and the transmission order between buffers. An object of the present invention is to provide a flow control method which does not need to be controlled individually and whose apparatus configuration is simple, and a communication element for realizing the flow control method.

Finally, a fifth object of the present invention is to provide a communication system in which, when VCs having different weights are multiplexed, cell transmission is not performed continuously for a certain VC so that traffic burstiness does not increase. An object of the present invention is to provide a flow control method and a communication element for realizing the same.

[0095]

SUMMARY OF THE INVENTION A first aspect of the present invention is directed to a plurality of VCs that transfer information in the form of fixed-length cells using GFR (Guaranteed Frame Rate) service categories. A flow control method used by a communication element that multiplexes a plurality of input VC cells in an ATM (Asynchronous Transfer Mode) communication network accommodating a virtual channel). An MCR determining step of determining an MCR (minimum cell rate), a cell accumulating step of distributing and accumulating cells to different cell buffers provided in the communication element in accordance with the determined MCR size, and a cell buffer. Multiplexing the cells stored in the ATM communication network at a speed higher than a predetermined output rate to an ATM communication network and outputting the multiplexed cells.

According to the first aspect, each VC
Even if there are multiple VCs using the TCP protocol to accommodate different cell buffers depending on the size of the MCR, and the MCR value differs among them,
MCR can be guaranteed for each VC.

[0097] The second invention is the flow control method according to the first invention, wherein the predetermined output rate in the cell output step is determined based on the M of a plurality of VCs stored in the cell buffer.
It is characterized in that it is a value obtained by summing CR.

According to the second aspect of the present invention, each VC
Even if there are multiple VCs using the TCP protocol to accommodate different cell buffers depending on the size of the MCR, and the MCR value differs among them,
MCR can be guaranteed for each VC, and the output rate is guaranteed to be the sum of the MCRs of a plurality of VCs stored in the cell buffer.

A third aspect of the present invention is the flow control method according to the first aspect, wherein the cell storing step determines a cell input to the communication element according to the determined MCR size.
Distributing to a different cell buffer provided in the communication element, and, if the number of cells in the cell buffer is greater than a preset congestion determination threshold, discarding the cell input to the communication element, Storing the cells that have not been discarded in the cell buffer.

According to the third aspect of the present invention, each VC
Even if there are multiple VCs using the TCP protocol to accommodate different cell buffers depending on the size of the MCR, and the MCR value differs among them,
MCR can be guaranteed for each VC.
When congestion occurs, predetermined cells can be discarded to eliminate the congestion state.

A fourth aspect of the present invention is an ATM (Asynchronous Channel) accommodating a plurality of VCs (virtual channels) for transferring information in a fixed-length cell format using a GFR (Guaranteed Frame Rate) service category. Transfer mode) In a communication network, a plurality of input VCs
A flow control method used by a communication element that multiplexes the cells of the above, wherein a protocol determination step of determining a type of an upper layer communication protocol in a cell of the VC input to the communication element; A cell storing step of distributing and storing cells to different cell buffers provided in a communication element; and a cell for multiplexing and outputting cells stored in the cell buffer to an ATM communication network at a speed higher than a predetermined output rate. And an output step.

According to the fourth aspect of the present invention, the TCP
Even in the case where VCs using higher transmission protocols such as the protocol and the UDP protocol are mixed, they are accommodated in different cell buffers for each protocol. Therefore, the MCR of each VC can be guaranteed.

A fifth aspect of the present invention is the flow control method according to the fourth aspect, wherein the predetermined output rate in the cell output step is determined based on the M of a plurality of VCs stored in the cell buffer.
It is characterized in that it is a value obtained by summing CR.

According to the fifth aspect as described above, even when VCs using the upper transmission protocol are mixed, the VCs are accommodated in different cell buffers for each protocol, so that the MCR of each VC is guaranteed. The output rate is guaranteed to be the sum of the MCRs of a plurality of VCs stored in the cell buffer.

A sixth aspect of the present invention is the flow control method according to the fourth aspect of the present invention, wherein the cell storing step determines a cell input to the communication element according to the determined MCR size.
Distributing to a different cell buffer provided in the communication element, and, if the number of cells in the cell buffer is greater than a preset congestion determination threshold, discarding the cell input to the communication element, Storing the cells that have not been discarded in the cell buffer.

According to the sixth aspect described above, even when VCs using the upper transmission protocol are mixed, the VCs are accommodated in different cell buffers for each protocol, so that the MCR of each VC is guaranteed. When congestion occurs, predetermined cells can be discarded to eliminate the congestion state.

The seventh invention is the flow control method according to the fourth invention, wherein the upper layer communication protocol is TCP
(Transmission Control Protocol) and UDP (User Datagram Protocol).

According to the seventh aspect of the present invention, the TCP
Even when VCs using the protocol and VCs using the UDP protocol coexist, they are accommodated in different cell buffers for each protocol. Therefore, M of each VC
CR guarantee can be performed.

According to an eighth aspect of the present invention, a transmitting terminal for transmitting information, a receiving terminal for receiving information, a plurality of communication paths for transmitting information, and information arriving from a plurality of communication paths are multiplexed into one communication path. A flow control method used by a communication element in a communication network including a communication element having a plurality of queues. A transmission order management step of managing the order in which information is transmitted from the transmission path to the communication path by a transmission order table; and determining whether information is accumulated in each queue in the order managed by the transmission order table. A transmission step of multiplexing the information to a communication channel at a predetermined timing and transmitting the information. When a new entry indicating the transmission order is to be inserted into the transmission order table, the existing entry number M, which is the number of entries already registered in the transmission order table, is replaced by a new entry, which is the number of entries to be newly inserted. Perform the operation of dividing by the number N in the range of integers,
A division step for calculating the quotient Q and the remainder R, and a plurality of entry insertion intervals (I) (where I is a natural number equal to or less than N-1) are expressed by the following equation (A): Entry insertion interval (I) = Q + 1 (where 1 ≦ I ≦ R) Entry insertion interval (I) = Q (where R + 1 ≦ I ≦ N−1) (A) An entry insertion interval calculation step calculated by the above equation (A), and an entry insertion interval of the entry inserted in the transmission order table Using each entry insertion position (I) as an integer value and an entry insertion start position determined by a predetermined method, the following expression (B): entry insertion position (L) = Mod {entry insertion start position +} (1 ≦ k ≦ L−1) Entry insertion interval (k), total number of entries｝... (B) (where L is a natural number equal to or less than N, and if L−1> 0, Σ (1 ≦ k ≦ L-1) Entry insertion interval (k)
Is from the entry insertion interval (1) to the entry insertion interval (L
-1), and represents 0 when L-1 = 0. ) An entry insertion position calculation step of calculating by the above equation (B), and a table creation step of newly inserting a corresponding entry into each entry insertion position (L) in the transmission order table.

According to the eighth aspect described above, each VC
Are adjusted so that they are as close as possible in time. Therefore, even when VCs having different weights are multiplexed, the burstiness of traffic does not increase.

A ninth invention is the flow control method according to the eighth invention, wherein the entry insertion position calculating step comprises:
The entry insertion start position is determined by a method of selecting from randomly generated integer values of 1 or more and M or less.

According to the ninth aspect, each VC
The cell transmission order is adjusted so as to be as close as possible in time, so that even when VCs with different weights are multiplexed, the burstiness of traffic does not increase and the entry insertion start position is generated randomly. Since the selection is made from the set integer values, it can be realized using a simple configuration.

A tenth invention is the flow control method according to the eighth invention, wherein the entry insertion position calculation step includes transmitting an insertion start position counter taking a value from 1 to a predetermined integer K, and transmitting the new entry. A step of incrementing by one from 1 to K each time an insertion is attempted to be made in the order table, and managing the insertion start position counter to return to 1 at the next increment when the insertion start counter reaches K; C) Number of existing entries × Insertion start position counter / K (C) A real number as a calculation result is obtained by the above equation (C), and the maximum integer value that is equal to or smaller than the real value is calculated. Determining an integer value as the entry insertion start position.

According to the tenth aspect, the entry insertion start position is determined so as not to approach.
The cell transmission order of each VC is adjusted so as not to be close in time. Therefore, even when VCs having different weights are multiplexed, the burstiness of traffic does not increase.

The eleventh invention is the flow control method according to the eighth to tenth inventions, wherein the communication network is an ATM accommodating a plurality of VCs (virtual channels) for transferring information in the form of fixed-length cells. (Asynchronous transfer mode) This is a communication network, and a predetermined criterion in the information storage step is determined according to the MCR (minimum cell rate) of the VC cell.

According to the eleventh aspect, each V
Since the cell transmission order of C is adjusted so as to be as close as possible in time, even when VCs having different weights are multiplexed, the burstiness of traffic does not increase, and the MCR guarantee of each VC is simultaneously ensured. It can be carried out.

The twelfth invention is an ATM which accommodates one or a plurality of VCs (virtual channels) for transferring information in a fixed-length cell format using a GFR (guaranteed frame rate) service category. (Asynchronous transfer mode) A flow control method used in a communication network, comprising: an observation step of observing an input rate of a cell of a VC at a UNI (user network interface); A tagging step of marking a predetermined bit of a GFC (Generic Flow Control) area in a cell of a portion exceeding the MCR when the MCR exceeds the minimum cell rate;
If congestion occurs in the ATM communication network, GFC
A cell discarding step of discarding cells in which bits in the area are marked.

According to the twelfth aspect, assuming a point-to-point terminal configuration, there is no need to perform contention control. Therefore, paying particular attention to the fact that the GFC bit is not used, For cells that exceed the MCR,
Tagging is performed using a bit different from the CLP bit. Therefore, even when the transmitting terminal sets the CLP bit to indicate the priority of the cell, it is possible to guarantee the MCR for each VC.

The thirteenth invention is an ATM which accommodates one or a plurality of VCs (virtual channels) for transferring information in the form of fixed-length cells using a GFR (guaranteed frame rate) service category. (Asynchronous transfer mode) A flow control method used in a communication network, comprising: an observation step of observing an input rate of a cell of a VC at a UNI (user network interface); If it is larger than the minimum cell rate, the VPI (virtual path identifier) region, the VCI (virtual channel identifier) region, or the PTI (payload type identifier) in the cells that exceed the MCR Any of the fire) areas If the tagging step of marking, the congestion in the ATM network has occurred for a predetermined bit in the is provided with a cell discard step of discarding cells marking bit is performed in the region.

According to the thirteenth aspect as described above, CL
When any bit of VPI, VCI, or PTI different from the P bit is not used, tagging is performed on a cell exceeding the MCR using the bit.
To this end, the transmitting terminal needs to transmit the CL to indicate the cell priority.
Even if the P bit is set, the M for each VC
CR can be guaranteed.

A fourteenth aspect of the present invention is an ATM (Asynchronous Network) accommodating a plurality of VCs (virtual channels) for transferring information in a fixed-length cell format using a GFR (Granted Frame Rate) service category. Transfer mode) In a communication network, a plurality of input V
A communication element for multiplexing cells of C, wherein the input VC
MCR (minimum cell rate) of a cell of a plurality of cells, a plurality of cell buffers for accumulating cells of the input VC, and a cell to a plurality of cell buffers corresponding to the determined size of the MCR. A cell buffer determining unit for allocating the cells and a cell buffer multiplexing unit in a predetermined order with respect to the plurality of cell buffers so that the cells stored in the plurality of cell buffers are output to the ATM communication network at a speed higher than a predetermined output rate. And a transmission order control unit for controlling the transmission order.

According to the fourteenth aspect, each V
In order to accommodate different VCs in different cell buffers depending on the MCR size of C, there are a plurality of VCs using the TCP protocol, and even if the MCR values are different among them, the MCR is Can be guaranteed.

A fifteenth aspect of the present invention is directed to an ATM (Asynchronous Channel) accommodating a plurality of VCs (Virtual Channels) for transferring information in a fixed-length cell format using a GFR (Guaranteed Frame Rate) service category. Transfer mode) In a communication network, a plurality of input V
A communication element for multiplexing cells of C, wherein the input VC
A protocol determining unit for determining the type of the upper layer communication protocol in the cell, a plurality of cell buffers for storing input VC cells, and a plurality of cell buffers corresponding to the determined type of the upper layer communication protocol. A cell buffer judging unit for distributing cells to the ATM communication network so that the cells stored in the cell buffers are output to the ATM communication network at a speed higher than a predetermined output rate. And a transmission order control unit for controlling the transmission order to be multiplexed and output.

According to the fifteenth aspect, TC
Even in the case where VCs using upper layer communication protocols such as the P protocol and the UDP protocol are mixed, they are accommodated in different cell buffers for each protocol. Therefore, the MCR of each VC can be guaranteed.

A sixteenth invention is directed to a communication network including a transmitting terminal for transmitting information, a receiving terminal for receiving information, and a plurality of communication paths for transmitting information. A plurality of queues for storing input information, a plurality of queues for storing input information, and a determining unit for distributing and storing input information in a plurality of queues according to a predetermined type, respectively. A transmission order control unit that controls information to be multiplexed and output in a predetermined order to a plurality of queues, so that the information stored in the queue is output to the communication network at a speed higher than a predetermined output rate. The transmission order control unit includes: a transmission order table that stores an order in which information is output; and a transmission order table when a new entry indicating the transmission order of a queue is to be inserted into the transmission order table. An operation of dividing the number of existing entries, which is the number of entries already registered, by the number of new entries N, which is the number of entries to be newly inserted, is executed in the range of integers to calculate the quotient Q and the remainder R, A plurality of entry insertion intervals (I) (where I is a natural number of N-1 or less)
From the following equation (D), the entry insertion interval (I) = Q + 1 (1 ≦ I ≦ R), the entry insertion interval (I) = Q (where R + 1 ≦ I ≦ N−1) (D) The entry insertion interval calculation unit to be calculated and each entry insertion position (I) calculated by the entry insertion interval calculation unit are expressed by the following equation (E) using an entry insertion start position which is an integer value and determined by a predetermined method. Entry insertion position (L) = Mod {entry insertion start position + {(1 ≦ k ≦ L−1) entry insertion interval (k), total number of entries}... (E) (where L is a natural number equal to or less than N) Therefore, if L-1> 0, then {(1 ≦ k ≦ L−1) entry insertion interval (k)
Is from the entry insertion interval (1) to the entry insertion interval (L
-1), and represents 0 when L-1 = 0. ) Includes an entry insertion position calculation unit that calculates by the above equation (E), and an entry insertion unit that newly inserts a corresponding entry at each entry insertion position (L) in the transmission order table.

According to the sixteenth aspect, each V
The cell transmission order of C is adjusted so as to be as close as possible in time. Therefore, even when VCs having different weights are multiplexed, the burstiness of traffic does not increase.

A seventeenth invention is directed to an ATM which accommodates one or a plurality of VCs (virtual channels) for transferring information in a fixed-length cell format using a GFR (guaranteed frame rate) service category. (Asynchronous transfer mode) In a UNI (user network interface) of a communication network, when an input rate of a cell of a VC is higher than an MCR (minimum cell rate), a G in a cell in a portion exceeding the MCR is used.
A communication element that communicates with a transmitting / receiving terminal via a packet policer GFC bit setting unit for marking a predetermined bit in an FC (Generic Flow Control) area, and a bit in the GFC area of an input cell. GFC bit observation unit that observes
A congestion determination unit that determines that congestion has occurred in the communication network, and when it is determined that congestion has occurred by the congestion determination unit, as a result of observation by the GFC bit observation unit, if the bit is marked, A cell discarding unit for discarding cells on which marking has been performed.

According to the seventeenth aspect, assuming a point-to-point terminal configuration, there is no need to perform contention control. Therefore, paying particular attention to the fact that the GFC bit is not used, For cells that exceed the MCR,
Tagging is performed using a bit different from the CLP bit. Therefore, even when the transmitting terminal sets the CLP bit to indicate the priority of the cell, it is possible to guarantee the MCR for each VC.

[0129]

DESCRIPTION OF THE PREFERRED EMBODIMENTS (First Embodiment) FIG. 1 is a block diagram showing a configuration of a communication network including communication elements for realizing a flow control method according to a first embodiment of the present invention.
In this communication network, MC included in communication element 1100
The configuration and operation of the R determination unit 1109 are characterized. Hereinafter, the configuration and operation of the communication network according to the present embodiment will be described.

In FIG. 1, the communication network includes a communication element 11
00 and 1110, and transmission terminals 1120, 1130 and 1140. The communication element 1100 includes n cell buffers 1101-1 to 1101-n, n queue length management units 1102-1 to 1102-n, n congestion determination units 1103-1 to 1103-n, n cell discard units 1104-1 to 1104-n, cell buffer determination unit 1105, transmission order control unit 1106, transmission timing determination unit 1107, and active buffer management unit 11
08 and an MCR determination unit 1109.

The communication elements 1110 and 1100 are connected to each other via a communication path. Also, the transmitting terminal 11
20, 1130 and 1140 are connected to the communication element 1100 via a communication path. In the present communication network, communication elements 1120, 1130 and 1140 send communication element 1
Information is transmitted via 100 to a receiving terminal (not shown) provided in the direction of the communication element 1110. The information is transferred in the form of a cell.

Each cell is stored in each cell buffer corresponding to each VC, and then multiplexed on a communication path and transmitted to communication element 1110. However, in this communication network, cells of a plurality of VCs are stored in one cell buffer according to the size of the MCR. FIG. 1 shows only a case where there is only one multiplexing form. However, in the part where congestion occurs such that cells are multiplexed,
Similar configurations are required respectively.

Next, the operation of communication element 1100 will be described. The MCR determination unit 1109 determines whether a preset MC
The R threshold is managed. A total of (n-1) MCR thresholds are provided. These MCR thresholds are referred to as MCR threshold 1, MCR threshold 2,.
When expressed as MCR threshold (n-1), MCR threshold 1 <MCR threshold 2 <... <MCR threshold (n-
It is assumed that the relationship of 1) is satisfied.

When a cell arrives, MCR determining section 1109 analyzes the cell header of the cell and determines whether or not the cell is a signaling cell. If it is determined that the arriving cell is a signaling cell, the MCR determination unit 1109
Is the MC of the VC from the contents of the signaling cell.
Read the value of R.

The MCR determination unit 1109 determines that the read M
The magnitude of the CR value is sequentially compared with the magnitude of the MCR threshold value, and the minimum MCR threshold value j (where j is a natural number) that is equal to or greater than the read MCR value is selected. . Here, the value of j is defined as an MCR determination value.
However, when the value of the read MCR is larger than all the MCR thresholds, the MCR determination value is n.

Thereafter, MCR determining section 1109 outputs the VC number and MCR determination value of the cell to cell buffer determining section 1105. Further, the MCR determination unit 1109
Transmits the cell to the cell buffer determination unit 1105. If it is determined that the arriving cell is not a signaling cell, MCR determining section 1109 immediately transmits the cell to cell buffer determining section 1105.

The cell buffer judgment section 1105 manages a cell buffer judgment table. The cell buffer determination table is a table showing the correspondence between the VC number and the MCR determination value. The cell buffer determination unit 1105
When the number and the MCR determination value are received from the MCR determination unit 1109, they are added to the cell buffer determination table and stored in association with each other.

When a cell arrives, cell buffer determining section 1105 determines the VCI field and VP of the arriving cell.
The VC number is determined from the value of the I field. Then, the cell buffer determination unit 1105 determines the MCR determination value corresponding to the determined VC number with reference to the VC number determination table. After determining the MCR determination value, the cell buffer determination unit 1105 transmits the cell to the cell discard unit 1104-m (where m is a natural number) corresponding to the MCR determination value. Here, m represents the above-mentioned MCR determination value.

Next, each of n cell discarding units, cell buffers, queue length managing units and congestion judging units is provided, and the operation is the same for all n units. Therefore, only the operation in one of these will be described below.

The operations of the cell discarding unit 1104-1, the cell buffer 1101-1, the transmission timing determining unit 1107, and the congestion judging unit 1103-1 are the same as those of the cell discarding unit 204-1, the cell buffer in the second prior art. 2
01-1, the operation of the transmission timing determination unit 207 and the operation of the congestion determination unit 203-1 are the same as those of the first embodiment, and a description thereof will be omitted.

The queue length management section 1102-1 manages the queue length. When receiving the cell arrival signal from the cell buffer 1011-1, the queue length management unit 1102-1 increases the queue length by one. Thereafter, the queue length management unit 1102-1 updates the queue length with the congestion determination unit 11
03-1 and outputs the active signal and the corresponding cell buffer number to the active buffer management unit 1108.

When receiving a cell transmission signal from cell buffer 1011-1, queue length management section 1102-1 decreases the queue length by one. Thereafter, the queue length management unit 1102-1 updates the queue length with the congestion determination unit 1
If the queue length is 0, the inactive signal and the corresponding cell buffer number are output to the active buffer management unit 1108. The queue length is set to 0 when the operation of the communication element starts.

The active buffer management unit 1108 manages whether each cell buffer is in an active state by using an active flag. Here, the case where the cell buffer is in the active state refers to the case where cells are accumulated in the cell buffer.

When the active buffer management unit 1108 receives the active signal and the cell buffer number from the queue length management unit 1102-1, the active buffer management unit 1108 sets the value of the active flag in the cell buffer corresponding to the received number to 1
Set to. When receiving the inactive signal and the cell buffer number, the active buffer management unit 1108 sets the value of the active flag in the cell buffer corresponding to the received number to 0.

The transmission order control section 1106 manages transmission cell buffer numbers. The transmission order control unit 1106
Upon receiving the transmission timing signal from transmission timing determining section 1107, it transmits a cell transmission command signal to the cell buffer corresponding to the transmission cell buffer number. However,
If the transmission cell buffer number is set to 0,
Transmission order control section 1106 does not transmit a cell transmission command signal.

The configuration and operation of transmission order control section 1106 are substantially the same as the configuration and operation of transmission order control section 206 in the second prior art. However, the following points are different. First, in this embodiment, a transmission cell buffer number management unit (not shown) is provided instead of the transmission VC number management unit 2068 in the second related art. Also, the transmission VC number management unit 2068 in the second related art manages transmission VC numbers, whereas the transmission cell buffer number management unit in the present embodiment manages transmission cell buffer numbers.

Next, instead of the VC weight management table 2061 in the second prior art,
A VC weight management table 2061 is provided. The VC weight management table 2061 according to the second related art manages the weight of each VC, but the cell buffer weight management table according to the present embodiment manages the weight of each cell buffer. Here, the weight of each cell buffer is an integer value, and the MC of the VC accommodated in the cell buffer is
Each value is selected so as to be proportional to the total value of R.

Further, in this embodiment, a cell buffer weight division element table (not shown) is provided in place of the VC weight division element table 2062 in the second prior art. The VC weight division element table 2062 according to the second conventional technique manages the weight division elements of each VC, but the cell buffer weight division element table according to the present embodiment manages the weight division elements of each cell buffer. .

Finally, instead of the transmission reservation VC number management section 2067 in the second prior art, a transmission reservation VC number management section 2067 is provided in the present embodiment.
Further, the transmission reservation VC number management unit 2067 in the second related art manages the transmission reservation flag of each VC, but the transmission reservation cell buffer management unit in the present embodiment manages the transmission reservation flag of each cell buffer.

As described above, in the present embodiment, a plurality of VCs using the TCP protocol exist in order to accommodate each VC in a different cell buffer according to the size of the MCR, and the MCR of the MCR is interposed between them. Even if the values are different, the MCR can be guaranteed for each VC.

(Second Embodiment) FIG. 2 shows a second embodiment of the present invention.
FIG. 14 is a block diagram illustrating a configuration of a communication network including a communication element that implements the flow control method according to the embodiment. This communication network is characterized by the configuration and operation of the protocol determination unit 41105 included in the communication element 41100. Hereinafter, the configuration and operation of the communication network according to the present embodiment will be described.

The communication network comprises communication elements 41100 and 4100
1110, transmission terminals 41120 to 41170, and routers 41181 to 41183. Communication element 411
00 indicates cell buffers 41101-1 and 41101.
-2, and queue length management units 411022-1 and 4110
2-2, congestion determination units 41103-1 and 41103
-2, and cell discarding units 41104-1 and 41104-
2, a cell buffer determination unit 41105, a transmission order control unit 41106, a transmission timing determination unit 41107,
An active buffer management unit 41108 and a protocol determination unit 41109 are included.

Communication elements 41100 and 41110 are:
The transmission terminal 411 is connected to each other via a communication path.
20 to 41170 are connected to the communication element 41100 via a communication path.

In this communication network, transmitting terminals 41120 to 41120 to
41170 transmits information to a receiving terminal (not shown) provided in the direction of the communication element 41110 via each of the routers 41181 to 41183 and the communication element 41100. The information is transferred in the form of a cell.

Each cell is stored in each cell buffer according to the VC, and then multiplexed and transmitted on the communication path to the communication element 41110. However, in the present embodiment, a VC using TCP and a VC using UDP are stored in different cell buffers. Although the communication network of FIG. 2 shows only one multiplexing mode, a similar configuration is required in a portion where congestion occurs in which cells are multiplexed.

The routers 41181 to 41183 are:
It has an ATM interface and transmits incoming data in the form of cells. At this time, routers 41181-4
1183 indicates different Vs on the communication path depending on whether the protocol of the arriving data is TCP or UDP.
Send to C

Next, the operation of communication element 41100 will be described. Cell buffers 41101-1 and 41101 provided in communication element 41100 according to the present embodiment
-2, and queue length management units 411022-1 and 4110
2-2, congestion determination units 41103-1 and 41103
-2, and cell discarding units 41104-1 and 41104-
2, a cell buffer determination unit 41105, a transmission order control unit 41106, a transmission timing determination unit 41107,
The operation of each unit in the active buffer management unit 41108 is the same as that of the cell buffers 1101-1 to 1101- provided in the communication element 1100 according to the first embodiment.
n, queue length management units 1102-1 to 1102-n,
Congestion determining units 1103-1 to 1103-n, cell discarding units 1104-1 to 1104-n, and cell buffer determining unit 1
105, a transmission order control unit 1106, a transmission timing determination unit 1107, and an active buffer management unit 1108
This is the same as the operation of each unit. Therefore, the description of the operation of these components will be omitted.

[0158] The protocol determination section 41105 manages a protocol determination table. The protocol determination table is a table indicating a correspondence between the VC number and the protocol type. Here, it is assumed that the protocol type is either TCP or UDP protocol.

When a cell arrives, the protocol determining section 41105 determines the VC number from the values of the VCI field and the VPI field of the cell. The protocol determining unit 41105 determines a protocol type corresponding to the VC number from the determined VC number. If the determined protocol is TCP, the protocol determination unit 4
1105 transmits the cell to the cell discarding unit 41104-1. When the protocol is UDP, the protocol determining unit 41105 discards the cell in the cell discarding unit 4.
Send to 1104-2.

As described above, according to the present embodiment, TC
Even when VCs using the P protocol and VCs using the UDP protocol are mixed, they are accommodated in different cell buffers for each protocol. Therefore, the MCR of each VC can be guaranteed.

In this embodiment, it is assumed that VCs using the TCP protocol and VCs using the UDP protocol coexist. However, if the upper layer communication protocol is used, only these VCs coexist. However, the present invention is not limited to the case where other communication protocols are mixed.

Further, even when VCs using three or more upper layer communication protocols are mixed, three or more different cell buffers are provided for each protocol, and each cell is stored correspondingly. According to the present embodiment, the MCR guarantee of each VC can be performed if configured to accommodate.

(Third Embodiment) FIG. 3 shows a third embodiment of the present invention.
FIG. 14 is a block diagram illustrating a configuration of a communication network including a communication element that implements the flow control method according to the embodiment. The configuration and operation of this communication network are almost similar to those of the communication network according to the first embodiment. Hereinafter, the configuration and operation of the communication network according to the present embodiment will be described.

The communication network comprises communication elements 1200 and 12
10 and transmission terminals 1220 to 1240. The communication element 1200 includes n cell buffers 1201-1 to 1201-1.
202-n and n queue length management units 1202-1 to 120-1
202-n and n congestion determination units 1203-1 to 120-120
3-n and n cell discard units 1204-1 to 1204-
n, a VC determination unit 1205, and a transmission order control unit 1206
, A transmission timing determining unit 1207, and an active VC
It includes a management unit 1208 and an MCR determination unit 1209.

The communication elements 1200 and 1210 are connected via a communication path. Sending terminals 1220-124
0 is connected to the communication element 1200 via a communication path. In this communication network, transmitting terminals 1220 to 1240
Is transmitting information via the communication element 1200 to a receiving terminal (not shown) provided in the direction of the communication element 1210. The information is transferred in the form of a cell.

Each cell is stored in each cell buffer according to the VC, and then multiplexed and transmitted on the communication path to the communication element 1210. Although the communication network shown in FIG. 3 shows only one multiplexing form, a similar configuration is required in a portion where congestion occurs in which cells are multiplexed.

In this communication network, the total number of VCs is n or less. In this communication network, a cell discarding unit, a cell buffer, a queue length managing unit, and a congestion determining unit are provided for each of the VCs.

Next, the operation of communication element 1200 will be described. The cell buffers 1201-1 to 1201-n provided in the communication element 1200 according to the present embodiment, the queue length management units 1202-1 to 1202-n, and the congestion determination unit 12
03-1 to 1203-n and a cell discarding unit 1204-1 to 1203-n
The operations of the respective units in the communication element 1204-n are the same as those of the cell buffers 1101-1 to 1101-n provided in the communication element 1100 according to the first embodiment and the queue length management unit 1
102-1 to 1102-n and a congestion determination unit 1103-1
To 1103-n and cell discarding units 1104-1 to 1104
-N is the same as the operation of each unit. Therefore,
The description of the operations related to these components will be omitted.

The operation of the transmission order control unit 1206 provided in the communication element 1200 according to this embodiment is different from that of the first embodiment, and will be described below. The transmission timing determining unit 1207 generates a transmission timing at intervals of a predetermined time, and transmits a transmission timing signal at the time to a transmission VC provided inside the transmission order control unit 1206.
Output to the number management unit 12064.

The transmission VC number management section 12064 transmits the transmission V
The C number is managed. Upon receiving the transmission timing signal, the transmission VC number management unit 12064 transmits the received transmission V
The cell transmission command signal is transmitted to the cell buffer corresponding to the C number. However, when the transmission VC number is set to 0, transmission VC number management section 12064 does not transmit a cell transmission command signal. Transmission order control unit 1206
And the configuration of the transmission VC number management unit 12064 and the transmission V
The method of determining the C number will be described later.

The transmission order determination algorithm used in the transmission order control section 1206 performs weighting in proportion to MCR for each VC, in a manner similar to the method described above with reference to the second conventional technique, so as to be in proportion to each weight. The distribution band of the communication path is distributed among the active VCs. As described above, the active VC is a VC having a cell input from the transmitting terminal and having cells stored in the cell buffer. However, in the present embodiment, the method of determining the transmission order is different from the second conventional technique. This will be described later.

FIG. 4 is a block diagram showing a detailed configuration of transmission order control section 1206. In FIG. 4, the transmission order control unit 1206 includes a VC weight management table 12061
, A transmission order table 12062, a transmission order pointer unit 12063, a transmission VC number management unit 12064, an entry insertion start position determination unit 12065, an entry insertion interval calculation unit 1266, and an entry insertion unit 12067.

When a cell is input, MCR determining section 1209 determines the type of the input cell. If the determined cell type is a signaling cell for performing VC setting, the MCR determination unit 1209 determines whether the unused VC
A number is newly set for the VC. Here, the VC number is a natural number equal to or less than n, and is equal to one of the cell buffer, the queue length management unit, the congestion determination unit, and the cell discard unit.
It corresponds to one to one. Also, the MCR determination unit 1209
The value of the MCR is determined from the contents of the signaling cell.

Next, the MCR determination unit 1209 outputs the VC number and the value of the MCR to the VC weight management table 12061. After that, the MCR determination unit 1209 outputs the VC number and the cell to the VC determination unit 1205 at the same time.

When a cell arrives, the VC determination section 1205 determines the type of the cell. If the determined cell type is V
If the cell is a signaling cell for setting C, the VC number that is simultaneously input from the MCR determination unit 1209 is stored. Further, the values of the VCI field and VPI field used for the information cell are stored in association with the VC number. Thereafter, the VC determination unit 1205 determines that the cell is a cell discarding unit 1204-1 to 120-1 corresponding to the VC number.
204-n.

If the determined cell type is the information cell, the VC determination section 1205 determines the VC number from the values of the VCI field and VPI field of the cell. Thereafter, the VC determination unit 1205 determines that the cell is a cell discarding unit 1204-1 to 1204 corresponding to the VC number.
-N.

The VC weight management table 12061 indicates that each V
The weight of C is managed. In the present embodiment, the weight of each VC is associated with each VC and the transmission order table 12
062 is the number of entries to be newly inserted. Hereinafter, the number of entries is defined as the number of inserted entries.

The VC weight management table 12061 contains the MC
When the values of the VC number and the MCR are received from the R determining unit 1209, the weight is calculated by the following equation (1). Where m
ax_int {A} represents the largest integer that is equal to or smaller than A. The unit band is a preset constant. Weight = max_int {MCR / unit band} (1)

The weight calculated by the above equation (1) corresponds to the transmission order table 120 corresponding to the newly set VC.
62 is the number of inserted entries newly stored. Also,
By calculating the weights in this manner, the weights are calculated as MCR
Can be defined as an integer value that is approximately proportional to

The weight is associated with the VC number,
It is temporarily stored in the VC weight management table 12061. Thereafter, the VC weight management table 12061 stores the weight (that is, the number of inserted entries) in the transmission order pointer unit 12.
063, an entry insertion start position determination unit 12065 and an entry insertion interval calculation unit 1266 are output. Further, the VC weight management table 12061 outputs the VC number to the transmission order table 12062.

Entry insertion start position determining section 12065
Manages the insertion start position counter and the total number of entries. The total number of entries is stored in the transmission order table 120.
62 is the total number of entries. The insertion start position counter is a natural number equal to or smaller than N.

The insertion start position counter is determined from the VC weight management table 12061 by the entry insertion start position determining unit 12.
Each time a weight is input to 065, it is incremented by one. However, after the insertion start position counter is incremented from 1 to N, it is further incremented by one.
And then incremented to N repeatedly. Specifically, the value of the insertion start position counter is 1, 2,..., N-1, N, 1, 2,.
N-1, N, 1,... Note that N is a preset constant. The initial value of the insertion start position counter is set to 1.

Entry insertion start position determining section 12065
Receives the weight from the VC weight management table 12061, calculates the entry insertion start position by the following equation (2). Entry insertion start position = max_int (total number of entries × insertion start position counter / N) (2)

The entry insertion position determining section 12065 outputs the value calculated by the above equation (2) to the entry inserting section 12067. After that, the entry insertion start position determination unit 12065 increases the total number of entries by the received weight value (that is, the number of inserted entries). Furthermore, the entry insertion start position determination unit 12065 increments the value of the insertion start position counter. The order of the increment is as described above.

In the present embodiment, the entry insertion start position is calculated by the above-mentioned equation (2). However, the above equation (2) is used to adjust the cell transmission order of each VC so as to be as close as possible in time. Therefore, the entry insertion start position can be configured to be selected from, for example, an integer value randomly generated within a range of not less than 1 and not more than the existing total number of entries. With such a configuration, each V
The cell transmission order of C can be adjusted so as to be as close as possible in time.

The entry insertion interval calculation unit 12066 manages the total number of entries. The total number of entries is the total number of entries included in the transmission order table 12062, and the initial value is set to 0 as described above. Upon receiving the weight, the entry insertion interval calculation unit 12066 calculates the quotient Q and the remainder R by performing the division expressed by the following equation (3) in the range of integers. Total number of entries / weight ... (3)

Further, using the quotient Q and the remainder R calculated from the above equation (3), the entry insertion interval calculation section 1206
6 calculates a plurality of entry insertion intervals by the following equation (4). However, assuming that the received weight value is W, there are (W-1) entries to be calculated, and these are calculated as entry insertion intervals (1), entry insertion intervals (2),. It shall be represented as an entry insertion interval (W-1). Entry insertion interval (I) = Q (where 1 ≦ I ≦ WR−1) Entry insertion interval (I) = Q + 1 (where WR ≦ I ≦ W−1) (4)

The values and weights of the plurality of entry insertion intervals calculated by the above equation (4) are output to the entry insertion unit 12067 by the entry insertion interval calculation unit 1266. After that, the entry insertion interval calculation unit 12066 increases the total number of entries by the received weight value (that is, the number of inserted entries).

[0189] The entry insertion unit 12067 manages the total number of entries. The entry insertion unit 12067 receives the entry insertion start position from the entry insertion start position determination unit 12065. Further, an entry insertion unit 1206
7 receives the entry insertion intervals (1),..., Entry insertion intervals (W-1) from the entry insertion interval calculation unit 12066, and calculates a plurality of entry insertion positions according to the following equation (5) in response to these. I do. However, the plurality of entry insertion positions are represented as entry insertion positions (L), and L is a natural number equal to or less than (W-1). Entry insertion position (L) = Mod {entry insertion start position + {(1 ≦ k ≦ L−1) entry insertion interval (k), total number of entries} (5)

Here, in the above equation (5), L-1> 0
In the case of Σ, the (1 ≦ k ≦ L−1) entry insertion interval (k) represents the sum total from the entry insertion interval (1) to the entry insertion interval (L−1). In this case, it represents 0. Specifically, the entry insertion interval (1) + the entry insertion interval (2) +,.
1) represents the calculated value. Also, Mod @ A,
B｝ represents a remainder when A ÷ B is executed in an integer range.

The process of calculating a plurality of entry insertion positions in order by the above equation (5) can be specifically shown as follows. First, the entry insertion position (1) is M
od {entry insertion start position, total number of entries}. The entry insertion position (2) is calculated by Mod {entry insertion start position + entry insertion interval (1), total number of entries}. Then, the calculation after the entry insertion position (3) is similarly performed. Finally, the entry insertion position (W) is Mod {entry insertion start position + {
(1.ltoreq.k.ltoreq.W-1) Calculated from the entry insertion interval (k), the total number of entries. Therefore, W entry insertion positions are calculated, and their values are represented as entry insertion position (1), entry insertion position (2),..., Entry insertion position (W).

The entry insertion unit 12067 outputs the plurality of entry insertion position values calculated as described above to the transmission order table 12062. Further, the entry inserting unit 12067 increases the total number of entries by the received weight value (that is, the number of inserted entries).

FIG. 5 is a diagram showing an example of the contents of the transmission order table 12062. As shown in FIG. 5, the transmission order table 12062 manages the transmission order of each VC. Referring to FIG. 5, the transmission order table 12062
Are assigned one VC number. In FIG. 5, the weights of VC1 to VC8 are each one.

The transmission order table 12062 receives the VC number from the VC weight management table 12061. The transmission order table 12062 indicates that the entry insertion unit 1
From entry 2067, each entry insertion position represented by entry insertion position (1), entry insertion position (2),..., Entry insertion position (W) is received. Thereafter, the transmission order table 12062 inserts the entry having the received VC number immediately after each received entry insertion position. However, if the received entry insertion position is 0, the transmission order table 12062 inserts the entry having the received VC number at the head of the table.

Next, the operation for inserting a new entry into the transmission order table 12062 will be specifically described. Here, as an example, a case where a VC 9 with a weight of 8 is inserted into the transmission order table 12062 shown in FIG.

Entry insertion start position determining section 12065
Determines the entry insertion start position as follows.
Now, assuming that N = 8 and the insertion start position counter is 1, the entry insertion start position becomes max_int (8 ×）) = 1 from the above equation (2).

The entry insertion interval calculation unit 12066 calculates the entry insertion interval as follows. Since the total number of entries is 8, the value of quotient Q = 1 and remainder R = 0 can be obtained by calculating 8 ÷ 8 from the above equation (3). Further, all entry insertion intervals are 1 ≦ I ≦
7. Therefore, from the above equation (4), all the entry insertion intervals are equal to the entry insertion interval (I) = 1.
Is calculated.

The entry insertion section 12067 calculates the entry insertion position as follows. From the above equation (5),
Each entry insertion position is the entry insertion position (1) = Mo
d {1, 8} = 1, entry insertion position (2) = Mod
{1 + 1,8} = 2,..., Entry insertion position (8) = M
od {1 + 7,8} = 0. Each entry insertion position calculated in this manner is stored in the entry insertion unit 120.
67 to the transmission order table 12062.

In the transmission order table 12062, each entry is inserted at the position where each entry is received. FIG.
Transmission order table 12 after inserting each entry
It is the figure which showed the content of 062. Hereinafter, when the total number of entries included in the transmission order table 12062 is M, each entry in the transmission order table 12062 is represented as E (1), E (2),. A (1), A (2),..., A
(M).

For example, referring to FIG. 5, E (1) = V
C1, E (2) = VC2,..., E (8) = VC8. Referring to FIG. 6, E (1) = VC9, E
(2) = VC1, E (3) = VC9, E (4) = VC
2,..., E (16) = VC8.

However, even after a new entry is inserted, the address of an existing entry does not change. For example,
A (1) in FIG. 5 is equal to A (2) in FIG. That is, even if an entry is inserted, the transmission order pointer value in transmission order pointer section 12063 does not change. The configuration and operation of the transmission order pointer unit 12063 will be described later.

The transmission VC number management unit 12064 transmits the transmission V
The C number is managed. Transmission VC number management unit 1206
When the VC number is received from the active VC management unit 1208, the transmission VC is set to the received VC number.
Update the number. On the other hand, when the transmission VC number reset signal is received, the transmission VC number management unit 12064 resets the transmission VC number to 0.

Upon receiving the transmission timing signal, transmission VC number management section 12064 transmits transmission VC signal as described above.
The cell transmission command signal is transmitted to the cell buffer corresponding to the number. Thereafter, a VC recalculation command signal is transmitted to the transmission order pointer unit 12063. However, transmission V
When the C number is set to 0, the transmission VC number management unit 12064 does not transmit the cell transmission command signal, but only transmits the VC recalculation command signal to the transmission order pointer unit 12063.

The transmission order pointer section 12063 manages the total number of entries, a pointer increase counter, and a transmission order pointer. The total number of entries is in the transmission order table 1
2062 indicates the total number of entries included. The total number of entries and the pointer increment counter are set to 0 at the start of operation of the communication element.

The transmission order pointer is the transmission order table 1
2062 is to designate a certain entry inside. Specifically, the transmission order pointer is stored in the transmission order table 12.
062 is the address A (I) of a certain entry.
When the operation of the communication element starts, the value of the transmission order pointer is
The address of a predetermined entry in the transmission order table 12062 is set. Here, the transmission order table 12062 creates the first entry at a predetermined address when the communication element operation starts.

Upon receiving the weight from the VC weight management table 12061, the transmission order pointer unit 12063 increases the value of the total number of entries by the value of the received weight (ie, the number of inserted entries). The transmission order pointer unit 1
2063, when a reset signal is received from an active VC management unit 1208 described below, the counter value is set to 0.
Reset to.

When receiving the VC recalculation command, the transmission order pointer unit 12063 compares the value of the counter with the total number of entries. As a result of the comparison, if the total number of entries is larger, the value of the transmission order pointer is set to A (I)
To A (I + 1).

Next, the transmission order pointer unit 12063
With reference to the VC number stored in the entry E (I + 1), the VC number and the active determination command signal are output to the active VC management unit 1208. Thereafter, the transmission order pointer unit 12063 increases the value of the counter by one. It should be noted that I is a natural number, and A (I + 1) represents A (1) only when I is equal to the total number of entries.

On the other hand, as a result of comparing the value of the counter with the total number of entries, if the total number of entries is equal to the counter value, transmission order pointer section 12063 transmits to transmission VC number management section 12064 A VC number reset signal is output, and the value of the counter is reset to 0.

When receiving the active determination command signal and the VC number, active VC management section 1208 refers to the value of the active flag of the VC corresponding to the received VC number. If the value of the referenced active flag is 1, the active VC management unit 1208 outputs the VC number to the transmission VC number management unit 12064 and sends a reset signal to the transmission order pointer unit 12063.

On the other hand, if the value of the referenced active flag is 0, the active VC management unit 1208
The VC recalculation command signal is output to the transmission order pointer unit 12063.

According to the present embodiment as described above, as shown in FIG. 6, the cell transmission order of each VC is adjusted so as to be as close as possible in time. Therefore, even when VCs having different weights are multiplexed, the burstiness of traffic does not increase.

This embodiment is based on the premise that communication elements in an ATM communication network multiplex VCs having different weights. However, the present embodiment is not limited to such a case. For example, the present embodiment can be applied to a case where an IP switch or an IP router in an IP communication network multiplexes a plurality of packets.

Specifically, an IP switch or an IP router that performs packet exchange or routing is an FTP or TELN
Information stored in a queue provided for each application such as ET or for each address may be transmitted according to the weight based on the method used in the present embodiment. In this case, the weight is determined from the amount of data communication per unit time, the amount of information obtained by multiplying the packet size by the number of packets, and the like. Even in such a case, the transmission order from each queue is adjusted so as to be as close as possible in time, so that the burstiness of traffic does not increase.

(Fourth Embodiment) FIG. 7 shows a fourth embodiment of the present invention.
FIG. 14 is a block diagram illustrating a configuration of a communication network including a communication element that implements the flow control method according to the embodiment. This communication network is characterized by the configuration and operation of the GFC bit observation unit 1004 included in the communication element 1000 and the configuration and operation of the packet policer GFC bit setting unit 1050. Hereinafter, the configuration and operation of the communication network according to the present embodiment will be described.

This communication network comprises communication elements 1000 and 10
10, a transmitting terminal 1020 to 1040, and a packet policer GFC bit setting unit 1050. The communication element 1000 includes a cell buffer 1001, a queue length management unit 1002, a congestion determination unit 1003, a GFC bit observation unit 1004, and a cell discard unit 1005.

Communication elements 1000 and 1010 are connected via a communication path. In addition, the sending terminal 1020
Reference numeral 1040 is connected to the communication element 1000 via a communication path. In this communication network, transmitting terminals 1020 to 1020
40 is a communication element 1010 via the communication element 1000
To a receiving terminal (not shown) provided in the direction of
Submit information. The information is transferred in the form of a cell. Each of these cells is a cell buffer 1 of the communication element 1000.
001 and are output to the communication element 1010.

Although only one multiplexing form is shown in the communication network of FIG. 7, a similar configuration is required in a portion where congestion occurs in which cells are multiplexed. The packet policer GFC bit setting unit 10
50 is User Network Interface
e (UNI).

Packet Policer GFC Bit Setting Unit 10
50 observes whether the input rate does not exceed the MCR when a cell is input from each VC to the network. This observation is based on F-GCRA (Frame based Generator)
ic Control Rate Algorithm
This is performed using an algorithm called m). As a result of this observation, cells that exceed the MCR are identified by the packet policer GFC bit setting unit 1050 as GFC (Gen
1 is set to a predetermined bit in an eric flow control area. For the algorithm of F-GCRA, see "ATM Forum / BTD-
TM-01.03: Traffic Management
nt Baseline Text Documen
t ", p13. The operation of setting bits such as CLP bits by a network is called tagging.

Next, the operation of communication element 1000 will be described. The cell discard unit 1005 manages a cell discard flag. When a cell arrives, the cell discard unit 1005
A cell arrival signal is transmitted to the GFC bit observation unit 1004.

Upon receiving a cell arrival signal from cell buffer 1001, GFC bit observation section 1004 observes a predetermined bit value in the GFC area of the cell. Then, GFC bit observing section 1004 transmits the value of the observed GFC bit to cell discarding section 1005.

Upon receiving the value of the GFC bit, cell discarding section 1005 calculates the logical product of the value and the value of the cell discard flag. The cell discard unit 1005 determines that the result of the logical product is 1
If so, discard the cell. When the result of the logical product is 0, the cell discarding unit 105 transmits the cell to the cell buffer 1001. Note that the value of the cell discard determination flag is set to 0 when the communication element operation starts.

The cell discarding unit 1005
When the congestion determination signal is received from No. 3, the value of the cell discard determination flag is set to 1. When the non-congestion determination signal is received, the cell discard unit 1005 sets the value of the cell discard determination flag to 0.

The cell buffer 1001 provided in the communication element 1000 according to the present embodiment and the queue length management unit 10
02 and the operation of each unit in the congestion determination unit 1003 are as follows:
The cell buffers 1101-1 to 1101-n, the queue length management units 1102-1 to 1102-n, and the congestion determination units 1103-1 to 1103- provided in the communication element 1100 according to the first embodiment. This is the same as the operation of each unit in n. Therefore, the description of the operation of these components will be omitted.

Here, generally, the GFC bit is used in access control of a shared media network.
For example, the GFC bit is used when a plurality of terminals are connected on one bus in the UNI of B-ISDN. In such a case, the terminal closest to the network may not be able to transmit cells. Therefore, the GFC bit is generally used for contention control performed to avoid such a risk and realize fair use of the transmission path.

However, in the communication network of FIG. 7, a point-to-point terminal configuration is assumed, and in such a configuration, it is not necessary to perform the above-described contention control. Therefore, in the communication network according to the present embodiment, paying particular attention to the point that the GFC bit is not used,
Are tagged using a bit different from the CLP bit. To this end, according to the present embodiment, the transmitting terminal uses the CLP to indicate the priority of the cell.
Even if the bit is set, MC for each VC
R can be guaranteed.

In the communication network according to the present embodiment, a point-to-point terminal configuration is assumed.
The configuration is not limited to a point-to-point terminal configuration, since any configuration may be used as long as the bit is not used for a purpose other than that used in the present embodiment.

In the communication network according to the present embodiment, GF
Although C bits are used, bits in any area may be used as long as they are not used for purposes other than those used in the present embodiment. For example, VPI (Virtual Pa)
th Identifier), VCI (Virtua)
l Channel Identifier) and P
TI (Payload Type Identifier)
An arbitrary bit may be used in each area of r). However, HEC (Header ErrorCont)
Since the bits in the (roll) area are used for error checking, they cannot be used in the present embodiment.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a communication network including a communication element that implements a flow control method according to a first embodiment of the present invention.

FIG. 2 is a block diagram illustrating a configuration of a communication network including a communication element that implements a flow control method according to a second embodiment of the present invention.

FIG. 3 is a block diagram illustrating a configuration of a communication network including a communication element that implements a flow control method according to a third embodiment of the present invention.

FIG. 4 is a configuration diagram of a transmission order control unit according to a third embodiment of the present invention.

FIG. 5 is a diagram showing an example of the contents of a transmission order table according to a third embodiment of the present invention.

FIG. 6 is a diagram illustrating an example of the contents of a transmission order table according to a third embodiment of the present invention.

FIG. 7 is a block diagram illustrating a configuration of a communication network including a communication element that implements a flow control method according to a fourth embodiment of the present invention.

FIG. 8 is a block diagram illustrating a configuration of a communication network including a communication element that implements a flow control method according to the first related art.

FIG. 9 is a block diagram illustrating a configuration of a communication network including a communication element that implements a flow control method according to a second conventional technique.

FIG. 10 is a block diagram illustrating a configuration of a transmission order control unit according to a second conventional technique.

FIG. 11 is a diagram showing how parameters change with time in the second related art.

FIG. 12 is a diagram showing transmission timing of each VC according to the second conventional technique.

FIG. 13 is a diagram illustrating an example of a change in an input rate of each VC when the flow control method according to the first related art is used.

FIG. 14 is a diagram illustrating an example of a change in the transmission rate of each VC in a communication path when the flow control method according to the first related art is used.

FIG. 15 is a diagram illustrating an example of a change in the transmission rate of each VC in a communication path when the flow control method according to the first related art is used.

FIG. 16 is a diagram illustrating an example of a change in the transmission rate of each VC in a communication path when the flow control method according to the first related art is used.

[Explanation of symbols]

 1000 communication element 1001 cell buffer 1002 queue length management unit 1003 congestion determination unit 1004 GFC bit observation unit 1005 cell discard unit 1010 communication element 1020 to 1040 transmission terminal 1050 packet policer GFC bit setting unit 1100 communication element 1101-1 to 1101-n cell Buffer 1102-1 to 1102-n Queue length management unit 1103-1 to 1103-n Congestion determination unit 1104-1 to 1104-n Cell discard unit 1105 Cell buffer determination unit 1106 Transmission order control unit 1107 Transmission timing determination unit 1108 Active buffer Management unit 1109 MCR determination unit 1110 Communication element 1120 Transmission terminal 1130 Transmission terminal 1140 Transmission terminal 1200 Communication element 1201-1 to 1202-n Cell buffer 1202-1 to 1 02-n queue length management unit 1203-1 to 1203-n congestion determination unit 1204-1 to 1204-n cell discard unit 1205 VC determination unit 1206 transmission order control unit 12061 VC weight management table 12062 transmission order table 12063 transmission order pointer unit 12064 Transmission VC number management unit 12065 Entry insertion start position determination unit 1266 Entry insertion interval calculation unit 12067 Entry insertion unit 1207 Transmission timing determination unit 1208 Active VC management unit 1209 MCR determination unit 1210 Communication element 1220 to 1240 Transmission terminal 41100 Communication element 41101- DESCRIPTION OF SYMBOLS 1 Cell buffer 4111-2 Cell buffer 4112-1 Queue length management unit 410102-2 Queue length management unit 41103-1 Congestion determination unit 41103-2 Congestion determination unit 41 104-1 Cell Discarding Unit 41104-2 Cell Discarding Unit 41105 Cell Buffer Judgment Unit 41106 Transmission Order Control Unit 41107 Transmission Timing Determination Unit 41108 Active Buffer Management Unit 41109 Protocol Judgment Unit 41110 Communication Elements 41120 to 41170 Transmission Terminal 41181 to 41183 Router

Claims (17)

[Claims] 1. GFR (Granted Frame
Rate) In an ATM (asynchronous transfer mode) communication network accommodating a plurality of VCs (virtual channels) for transferring information in the form of fixed-length cells using service categories, a plurality of input cells of the VCs A MCR determining step of determining an MCR (minimum cell rate) of the cell of the VC input to the communication element, and a flow control method corresponding to the determined MCR size. A cell storage step of distributing and storing the cells in different cell buffers provided in the communication element; and multiplexing the cells stored in the cell buffer to the ATM communication network at a speed higher than a predetermined output rate. And a cell output step of outputting the data. 2. The method according to claim 1, wherein the predetermined output rate in the cell output step is a plurality of VCs stored in the cell buffer.
2. The flow control method according to claim 1, wherein the value is a value obtained by summing the MCRs. 3. The cell accumulating step includes: distributing cells input to the communication element to different cell buffers provided in the communication element in accordance with the determined MCR size; When the number of cells in the buffer is larger than a preset congestion determination threshold, the step of discarding cells input to the communication element and the step of storing cells not discarded in the cell buffer The flow control method according to claim 1, comprising: 4. A GFR (Granted Frame)
Rate) In an ATM (asynchronous transfer mode) communication network accommodating a plurality of VCs (virtual channels) for transferring information in the form of fixed-length cells using service categories, a plurality of input cells of the VCs A flow control method used by a communication element that multiplexes the communication protocol, wherein a protocol determination step of determining a type of an upper layer communication protocol in the cell of the VC input to the communication element; A cell storage step of distributing and storing the cells in different cell buffers provided in the communication element; and multiplexing the cells stored in the cell buffer to the ATM communication network at a speed higher than a predetermined output rate. And a cell output step of outputting the data. 5. The method according to claim 1, wherein the predetermined output rate in the cell output step is a plurality of VCs stored in the cell buffer.
The flow control method according to claim 4, wherein the MCR is a value obtained by summing the MCRs. 6. The cell accumulating step includes: distributing a cell input to the communication element to a different cell buffer provided in the communication element in accordance with the determined MCR size; When the number of cells in the buffer is larger than a preset congestion determination threshold, the step of discarding cells input to the communication element and the step of storing cells not discarded in the cell buffer The flow control method according to claim 4, comprising: 7. The communication system according to claim 7, wherein the upper layer communication protocol is TCP.
The flow control method according to claim 4, comprising a transmission control protocol (UDP) and a user datagram protocol (UDP). 8. A transmitting terminal for transmitting information, a receiving terminal for receiving information, a plurality of communication paths for transmitting information, and a communication element for multiplexing information arriving from the plurality of communication paths into one communication path. A flow control method used by the communication element, wherein the information arriving at the communication element is sorted according to a predetermined criterion, and stored in different queues; A transmission order management step of managing the order of transmitting information from the transmission path to the communication path by a transmission order table; and determining whether information is accumulated in each queue in the order managed by the transmission order table. And transmitting the information in a multiplexed manner at a predetermined timing to the communication path when the information is stored. In the step, when a new entry indicating the transmission order of the queue is to be inserted into the transmission order table, the existing entry number M, which is the number of entries already registered in the transmission order table, is newly inserted. A division step of calculating the quotient Q and the remainder R by executing an operation of dividing by the number of new entries N, which is the number of entries to be set, and a plurality of entry insertion intervals (I) (where I is N-1 The following natural number is calculated by the following equation (A): Entry insertion interval (I) = Q + 1 (where 1 ≦ I ≦ R) Entry insertion interval (I) = Q (where R + 1 ≦ I ≦ N−1) (A) (A) calculating an entry insertion interval, and each entry insertion position (I) of an entry to be inserted in the transmission order table is an integer value and is determined by a predetermined method. Using the entry insertion start position, the following equation (B): entry insertion position (L) = Mod {entry insertion start position + {(1 ≦ k ≦ L−1) entry insertion interval (k), total number of entries} (B) (However, L is a natural number equal to or less than N, and when L-1> 0, １ (1 ≦ k ≦ L−1) entry insertion interval (k)
Is from the entry insertion interval (1) to the entry insertion interval (L
-1), and represents 0 when L-1 = 0. A) an entry insertion position calculation step of calculating by the above equation (B); and a table creation step of newly inserting a corresponding entry at each of the entry insertion positions (L) in the transmission order table. Control method. 9. The entry insertion position calculating step,
9. The flow control method according to claim 8, wherein the entry insertion start position is determined by a method of selecting from randomly generated integer values of 1 or more and M or less. 10. The entry insertion position calculation step includes: setting an insertion start position counter taking a value from 1 to a predetermined integer K, from 1 to K every time a new entry is to be inserted into the transmission order table. Incrementing by 1 until the insertion start position counter reaches K, and managing it to return to 1 at the next increment, and the following equation (C): Number of existing entries × insertion start position counter / K... ( C) calculating a real number as a calculation result by the above equation (C), calculating an integer value that is equal to or smaller than the real value and is the maximum, and determining the integer value as an entry insertion start position. The flow control method according to claim 8, wherein 11. The communication network is an ATM (asynchronous transfer mode) communication network accommodating a plurality of VCs (virtual channels) for transferring information in a fixed-length cell format. Is based on the VC
11. The MCR (minimum cell rate) of each cell is determined according to the magnitude of the MCR (minimum cell rate).
The flow control method according to any one of the above. 12. An ATM (Asynchronous Transfer) accommodating one or a plurality of VCs (virtual channels) for transferring information in the form of fixed-length cells using a GFR (Granted Frame Rate) service category. Mode) A flow control method used in a communication network, comprising: an observation step of observing an input rate of a cell of the VC at a UNI (user network interface); and an observed input rate being an MCR (minimum cell rate). If greater, a tagging step of marking a predetermined bit of a GFC (Generic Flow Control) area in a cell in a portion exceeding the MCR, and when congestion occurs in the ATM communication network. In the GFC region, A cell discarding step of discarding cells whose bits are marked. 13. An ATM (Asynchronous Transfer) accommodating one or a plurality of VCs (Virtual Channels) for transferring information in a fixed-length cell format using a GFR (Granted Frame Rate) service category. Mode) A flow control method used in a communication network, comprising: an observation step of observing an input rate of a cell of the VC at a UNI (user network interface); and an observed input rate being an MCR (minimum cell rate). If it is larger, the VPI (virtual path identifier) region, VCI (virtual channel identifier) region or PTI (payload type identifier) region in the cell exceeding the MCR In any of A tagging step of marking a predetermined bit; and a cell discarding step of discarding a cell in which the bit is marked in the area when congestion occurs in the ATM communication network. , Flow control method. 14. ATM (Asynchronous Transfer Mode) accommodating a plurality of VCs (virtual channels) for transferring information in a fixed-length cell format using a GFR (Granted Frame Rate) service category. A communication element for multiplexing a plurality of input cells of the VC in a communication network, wherein an MCR (minimum cell rate) of the input cells of the VC is input.
A plurality of cell buffers for storing input VC cells; and a cell buffer determination unit for distributing cells to the plurality of cell buffers according to the determined MCR size. And multiplexing and outputting cells in a predetermined order to the plurality of cell buffers so that the cells stored in the plurality of cell buffers are output to the ATM communication network at a speed higher than a predetermined output rate. And a transmission order control unit for controlling the communication. 15. An ATM (Asynchronous Transfer Mode) accommodating a plurality of VCs (Virtual Channels) for transferring information in the form of fixed-length cells using a GFR (Granted Frame Rate) service category. A communication element for multiplexing a plurality of cells of the input VC in the communication network, the protocol determining unit determining a type of an upper layer communication protocol in the cells of the input VC; A plurality of cell buffers for storing cells; a cell buffer determination unit for distributing cells to the plurality of cell buffers in accordance with the determined type of the upper layer communication protocol; and a plurality of cells stored in the plurality of cell buffers. Are output to the ATM communication network at a speed higher than a predetermined output rate. And a transmission order controller for controlling to output the multiplexed cells in a predetermined order for a communication element. 16. In a communication network including a transmitting terminal for transmitting information, a receiving terminal for receiving information, and a plurality of communication paths for transmitting information, information arriving from the plurality of communication paths is transmitted to one communication path. A plurality of queues for storing input information; a determination unit for distributing and storing input information in a plurality of queues according to a predetermined type; A transmission order control unit for controlling information to be multiplexed and output to a plurality of queues in a predetermined order so that information stored in the queues is output to the communication network at a speed higher than a predetermined output rate. A transmission order table that stores an order in which information is output; and a case where a new entry indicating the transmission order of the queue is to be inserted into the transmission order table. In this case, an operation of dividing the number of existing entries, which is the number of entries already registered in the transmission order table, by the number of new entries N, which is the number of entries to be newly inserted, is performed in an integer range.
The quotient Q and the remainder R are calculated, and a plurality of entry insertion intervals (I) (where I is a natural number equal to or less than N-1) is calculated by the following equation (D): Entry insertion interval (I) = Q + 1 (where 1 ≦ I ≦ R ) Entry insertion interval (I) = Q (where R + 1 ≦ I ≦ N−1) (D) Entry insertion interval calculation unit calculated by the above equation (D), and each entry calculated by the entry insertion interval calculation unit The insertion position (I) is expressed by the following formula (E) using the entry insertion start position which is an integer value and determined by a predetermined method: entry insertion position (L) = Mod {entry insertion start position +} (1 ≦ k ≦ L-1) Entry insertion interval (k), total number of entries エ ン ト リ (E) (where L is a natural number equal to or less than N and when L-1> 0, Σ (1 ≦ k ≦ L− 1) Entry insertion interval (k)
Is from the entry insertion interval (1) to the entry insertion interval (L
-1), and represents 0 when L-1 = 0. A) an entry insertion position calculation unit that calculates by the above equation (E); and an entry insertion unit that newly inserts a corresponding entry at each of the entry insertion positions (L) in the transmission order table. element. 17. An ATM (Asynchronous Transfer) accommodating one or a plurality of VCs (virtual channels) for transferring information in a fixed-length cell format using a GFR (Granted Frame Rate) service category. Mode) In a UNI (user network interface) of a communication network, when an input rate of a cell of the VC is larger than an MCR (minimum cell rate), a GFC in a cell in a portion exceeding the MCR is used.
(Generic flow control) A communication element for communicating with a transmitting / receiving terminal via a packet policer GFC bit setting unit for marking a predetermined bit in an area, wherein said bit in the GFC area of an input cell A GFC bit observing unit for observing the data, a congestion judging unit for judging that congestion has occurred in the ATM communication network, and observing the GFC bit observing unit when the congestion judging unit judges that congestion has occurred. As a result, if the bit is marked, comprising a cell discarding unit that discards the cell that has been marked,
Communication element.