JPH11234306A - Data transferring device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To process one piece of user's data on a cell base without dividing it by storing user's information in a data unit in a payload area for (n) double length to transmit, switching the (n) double length data in an (n) double length data unit according to store and forward switching and reproducing the user's information from the payload area. SOLUTION: N double length data transmitted from plural transmitting terminals T1 and T2 arrive at relaying devices L1 and L2 on the way to a communication network. The multiplexing device L1 once stores data D1 from the terminal T1 and data D2 from the terminal T2 and first transfers the data D1 from terminal T1 and later the data D2 from the terminal T2. The device L2 performs multiple separation of the data D1 and D2 from the terminals T1 and T2. In such a case, the data D1 and D2 from the terminals T1 and T2 surely arrive with data as lumps of the data D1 and D2 in turn respectively.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is used in a system for transmitting and receiving a MAC frame and an IP packet including an Ethernet frame, and converting data and the like into cells by using an ATM in communication between computers. In particular, the present invention is used for a transfer method for outputting a large number of MAC frames or IP packet cells input at high speed to a destination.

[0002]

2. Description of the Related Art In ATM (Asynchronous Transfer Mode) communication, all information is divided and written and transferred to fixed-length cells. As described above, it is widely known that high-speed asynchronous transfer is enabled by writing and transferring all information in a fixed-length cell.

A conventional example will be described with reference to FIGS. 25 and 26. FIG. 25 is a conceptual diagram of a conventional ATM switch. FIG. 26 is a block diagram of a main part of a cell buffer unit used in a conventional ATM switch. As shown in FIG. 25, the cell is a cell buffer section 1-1 of the ATM switch.
And are temporarily stored in the cell buffers 1-2 and read out from the cell buffer units 1-1 and 1-2 after the contention control for avoiding the collision between cells going to the same output line. The cells read from the cell buffer units 1-1 and 1-2 are connected to the address filter 2 when the address written in the header area matches the address permitting transmission by the address filter 2. Output line.

[0004] As shown in FIG.
It comprises a cell buffer 10, a cell discarding unit 11, a control unit 12, and a queue length monitoring unit 13. Cells are stored in the cell buffer 10. The queue length of the cell buffer 10 is monitored by the queue length monitoring unit 13. The result of monitoring the queue length is reported to the control unit 12. In the control unit 12,
According to the queue length, incoming cells are stored in the cell buffer 1.
It is determined whether or not to accumulate 0. A threshold value is provided for the queue length, and whether to discard the queue length is determined based on whether the monitoring result reported from the queue length monitoring unit 13 exceeds the threshold value.

A conventional ATM communication network will be described with reference to FIG. FIG. 27 is a conceptual diagram of a conventional ATM communication network.
In this example, the data D1 and D2 of the transmitting terminals T1 and T2 are IP (Internet Protocol) packets. This IP packet is divided and written into the cell. This cell is input to the relay device L1. Relay device L
1 is provided with an ATM switch as shown in FIG. 25, and the cell has an address (VPI, VPI) written in its header area.
CI) and the output line is allocated to a desired output line. At this time, as shown in FIG. 27, even when a series of cells are transmitted from the transmitting terminals T1 and T2, the cells are exchanged irrespective of the relay device L1. A state occurs in which cells from the transmitting terminal T2 are mixed.

[0006] These cells are distributed to predetermined receiving terminals R1 or R2 by the relay device L2. The receiving terminals R1 and R2 reproduce the data D1 and D2, which are IP packets, by assembling the cells that have arrived.

A specific configuration example of the conventional relay devices L1 and L2 will be described with reference to FIG. FIG. 28 is a diagram showing a specific configuration example of conventional relay devices L1 and L2. As shown in FIG. 28, the relay devices L1 and L2 are configured by a first stage, a second stage, and a third stage,
In the first stage, tagging for switching in the second stage is performed. In the second stage, switching is performed according to the tag. The ATM switch shown in FIG. 25 shows the configuration of this second stage. In the third stage, tag deletion for returning to the original cell is performed.

[0008]

The user information of one data unit may be a unit such as an IP packet or a MAC frame, but these units have a variable length. When trying to transfer variable-length user information in a variable-length cell (one that stores user information in communication block units), high-speed hardware processing cannot be performed because of variable-length processing.

[0009] Therefore, if a fixed-length cell larger than the size of the user information (one that accommodates user information in communication block units) is used, many areas in the cell that do not accommodate user information are generated. Then, in the network, the exchange rate of useless information transfer increases.

[0010] For this reason, when a fixed-length cell smaller than the size of the user information is used, high-speed processing can be performed, and a useless unaccommodated area can be reduced.

[0011] IT used in the current ATM communication network
The cell structure of the UT standard has a header area of 5 bytes and a payload area of 48 bytes, and the cell length is fixed.
If the cell has a fixed length, the relay device in the middle of the network can buffer (temporarily store) the cell in a fixed length unit or transfer the cell, which is suitable for hardware processing. High-speed transfer is possible.

The transfer of data by such an ATM is as follows:
Since the transfer mode is intended to transfer traffic from a variety of information sources in a unified manner, there is a portion where adaptation to computer communication traffic is insufficient.

For example, if the size of the user information is longer than the cell length, the user information cannot be stored in one cell unit.
Dividing into a plurality of cells at transmitting terminals T1 and T2,
In the receiving terminals R1 and R2, it is necessary to store a plurality of cells until they are all at one end, and reassemble the user information.

[0014] The division of the plurality of cells causes what is called a cell tax problem. This is an existing ATM
OH (Over head) rate of cells used in the scheme, that is,
Frame synchronization and OAM (Operation Adm
(Instration and Maintenance). The OH rate of frames and packets before the appearance of ATM is less than 4% on average,
The OH rate of the cell is always 9.4% even excluding the overhead of the AAL, which is considerably higher than the OH rates of various frames before that.

Further, a dead cell problem has also occurred. This means that since the cell length is extremely short compared to the length of an IP packet or a MAC frame, even if only one cell is lost in the transmission path, the IP packet of the upper layer is discarded on the receiving side. I have. When the packet is discarded, the received information of the packet is lost in the UDP. In addition, retransmission processing is performed in TCP, which results in further increase in cell traffic. Therefore, A
Under the current situation where computer communication traffic occupies most of the TM communication network, it is necessary to improve compatibility with computer communication.

As a technique proposed from such a viewpoint, there is a technique disclosed in JP-A-7-254904. Here, a technique is disclosed in which user data is transferred without being divided by using n-fold length data, and the probability of retransmission of data due to partial loss of data is reduced. The above-mentioned publication discloses a technique for creating n-fold data, but does not consider a relay connection by store-and-forward and a process when the cell length is too large.

The present invention has been made in view of such a background, and the present applicant has specifically studied the structure of n-fold data which is optimal in view of the current state of the ATM communication network. And proposes its configuration. That is, an object of the present invention is to provide a data transfer device that can process data of one user on a cell basis without dividing the data and can solve the cell tax problem. SUMMARY OF THE INVENTION It is an object of the present invention to provide a data transfer device capable of improving data throughput in an ATM communication network and solving the dead cell problem. SUMMARY OF THE INVENTION It is an object of the present invention to provide a data transfer device capable of reducing a data transfer delay time in an ATM communication network. An object of the present invention is to provide a data transfer device to which an existing ATM switch can be applied.

[0018]

According to the present invention, as a method of transferring variable-length packets for computer communication, the length of n-fold data accommodating variable-length packets is set to be an integral multiple of the basic length data length. Features. Since the basic length data length can be processed in units of up to about 64 bytes at present, the header area length is set to 8 bytes and the payload area is set to an integral multiple of 8 bytes. Is one IP packet or one MAC frame
It is assumed that it is accommodated in two n-fold length data. In consideration of the consistency with the conventional ATM, the basic length data length is set to 53 bytes of the current cell length, and the n-fold length data is set to 5 bytes.
It can be an integer multiple of 3 bytes. As a result, one piece of data can be transferred, multiplexed, and discarded on a cell basis, so that the transfer delay time can be reduced in units of data.

That is, the present invention relates to a data transfer apparatus, which is characterized by the fact that user information in one data unit is divided into a payload area of n times long data in which n basic length data are cascaded. A transmission terminal provided with means for accommodating and transmitting the n-fold data to a communication network, a relay device having means for exchange-connecting the n-fold data by storage and exchange in n-fold data units, and Receiving means for receiving the data in units of double-length data and reproducing the user information from the payload area.

At this time, it is preferable that the relay apparatus and the receiving terminal each include means for discarding the n-times data that arrives in n-times data units.

Preferably, the transmitting means includes a means for inserting a padding part into a surplus length when the length of the payload area is longer than the length of the user information.

The relay device and the receiving terminal have n
A cell buffer for temporarily storing double-length data, and means for comparing the free space of the cell buffer with the length of the arrived double-length data according to information indicating the length included in the header area of the double-length data. The free space of the cell buffer is set to n
When the length of the double-length data is smaller than that of the double-length data, it is desirable to include a unit for discarding the entire n-fold length data. Accordingly, it is possible to prevent data in which some data of the user information transmitted from the transmitting terminal is missing from arriving at the receiving terminal.

The n-times long data has a cell structure of a cell length of L bytes as basic length data, and a length of a bit string L.times.n (L is a natural number) obtained by connecting n variable integers to each other.
The structure can be as follows.

For example, one basic length data located at the head of the n-length data is set as a header area of the n-times data, and all other basic length data are set as a payload area of the n-times data. Or L is 8 m bytes, m is an integer of 2 or more, and m is 8
Or the L byte is 53 bytes, and the first basic length data has a header area of 5 bytes of the n times long data and has a length of another (L × n) -5 bytes. A structure is used as a payload area of this n-fold data, or a structure using n-fold data in which n integers of 53-byte basic length data having a header area of 5 bytes and a payload area of 48 bytes are connected. be able to. The header area may include information indicating the length of the n-fold data.

At this time, the information indicating the length may be included in a part of the VPI area or a part of the VCI area in the header area.

When the basic length data of 53 bytes each having a header area of 5 bytes and a payload area of 48 bytes has a structure using n times integer data connected by n integers, the header area has the following structure. Information indicating whether or not the basic length data including the header area is the last position basic length data in the n-fold length data may be included. At this time, information indicating whether the data is the basic length data at the last position is a VPI in the header area.
Part of the area, or part of the VCI area, or PT
It can be included in a part of the region. Further, the header area may include information on the number of basic length data following the basic length data including the header area. Further, the header area may include information indicating whether or not the basic length data including the header area is located at the head of the n-fold length data.

[0027] In addition to the header area, the payload area may include information indicating the length of the n-fold data. At this time, the information indicating the length may be included in a part of the AAL area in the payload area.

In the case where the 53-byte basic length data having a header area of 5 bytes and a payload area of 48 bytes has a structure using n times integer data connected by n integers, the payload area includes: Information indicating whether or not the basic length data including the payload area is the basic length data at the last position in the n-fold length data can be included. Also in this case, the information indicating whether the data is the basic length data at the final position can be included in a part of the AAL area in the payload area.

As described above, when n-times data including information indicating whether the data is the basic length data at the final position is used, the relay apparatus and the receiving terminal temporarily store the n-times data. According to the cell buffer to be stored and the information indicating whether the data is the basic length data at the last position, the free space of the cell buffer becomes smaller than the length of the basic length data, and n times as long as the current writing is being performed. If the basic length data constituting the data is not yet the basic length data at the final position, it is desirable to include a unit for discarding the entirety of the n-fold length data currently being written.

[0030] The relay device may include means for extracting route information of n-fold data, means for dividing the n-fold data into basic length data, and means for extracting the basic length data. A first stage including a means for respectively providing the obtained route information, a second stage including a means for switching and connecting the basic length data transmitted from the first stage in accordance with the route information, It is also possible to provide a third stage including means for deleting the route information of the basic length data exchanged and connected by the stage, and means for assembling the basic length data into the n-fold length data.

At this time, the route information includes information on the input port at which the n-fold data has arrived and information indicating the basic length data at the final position, and the assembling means includes the information on the input port and the final It is desirable to include means for assembling n-fold length data for each input port according to the information indicating the basic length data of the position. Further, the basic length data may be an ATM cell length.

Thus, the first, same as the ATM switch,
By using the second and third stages, the existing AT
The M switch can be directly applied to the present invention.
Thereby, the data transfer device of the present invention can be realized in a short time. In particular, by making the basic length data the same length as the ATM cell, the existing ATM switch can be applied to the present invention as it is. Also, even if the basic length data is not an ATM cell length, but is an arbitrary length, simply changing the second stage to one that matches the length will cause the first and third stages to use the existing length. The ATM switch device can be applied as it is.

In the first stage, when n-length data to which various VPI / VCI are added is input to the input port, the n-length data is divided into basic length data,
A switching tag corresponding to PI / VCI is attached to the head of the basic length data. This tag is VP
It is set in advance so as to output to a predetermined output port corresponding to the I / VCI. In the second stage, switching is performed according to the tag. In the third stage, the tags are deleted, and at each output port, the basic length data is sorted for each input port. The sorted basic length data is assembled into n-fold length data.

[0034]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of the present invention will be described with reference to FIGS. 1 and 2, FIGS. 4 to 13, FIGS. FIG. 1 is a configuration diagram of a main part of a communication network according to an embodiment of the present invention. FIG. 2 is a block diagram of a main part of the cell buffer unit according to the embodiment of the present invention. 4 to 6 show n of the embodiment of the present invention.
FIG. 3 is a diagram illustrating a configuration of double-length data. FIG. 7 is a diagram showing the relationship between the data of the user information and the payload area. FIG.
FIG. 3 is a diagram showing a configuration of a variable length packet having a plurality of header areas. FIG. 9 is a diagram showing the configuration of the header area. FIG. 10 is a diagram showing an information writing position in the header area. FIG. 11 shows the AAL area. 12 and 13
FIG. 4 is a diagram showing an example of header information of a variable length packet having a plurality of header areas. FIG. 18 is a block diagram of a main part of an ATM switch corresponding to a relay device and a relay network according to a second embodiment of the present invention. FIG. 19 is a block diagram of a main part of an ATM switch corresponding to a relay device and a relay network according to a third embodiment of the present invention.

The present invention relates to a data transfer apparatus, which is characterized in that, as shown in FIG. 1, data D1 and D2, which are user information in one data unit, are cascaded with n basic length data. Transmitting terminal T1 that accommodates in the payload area of n-fold data and transmits to the communication network connected to
And T2, and the relay devices L1 and L2 for exchange-connecting the n-fold data by the storage exchange in the n-fold data unit
And receiving terminals R1 and R2 that receive the n-fold data in n-fold data units and reproduce the user information from the payload area. The relay devices L1 and L2 and the receiving terminals R1 and R2 discard incoming n-fold data in n-fold data units. When the length of the payload area is longer than the lengths of the data D1 and D2, the transmitting terminals T1 and T2
The padding part is inserted to the surplus length.

The relay devices L1, L2 and the receiving terminals R1,
As shown in FIG. 2, R2 has a cell buffer 10 for temporarily accumulating n-fold data, and information indicating the length included in the header area of the n-fold data read by the header identification unit 15, as shown in FIG. A control unit 14 for comparing the free space of the cell buffer detected by the queue length monitoring unit 13 with the length of the n-times long data that has arrived, and a free space of the cell buffer 10 according to the comparison result of the control unit 14 When the capacity is smaller than the length of the n-fold data,
It includes a cell discard unit 11 which is means for discarding all of the n-times long data.

The following is an example of the structure of n-fold data used in the data transfer device of the present invention. In the example shown in FIG. 4, the n-fold length data has a cell structure of a cell length of L bytes as basic length data, and the basic length data is connected to a variable number of variable integers of a bit string L × n (L is a natural number). ). At this time, one basic length data located at the head of the n-fold length data is set as a header area of the n-fold length data, and all other basic length data is set as a payload area of the n-fold length data.

In particular, in the embodiment of the present invention, FIG.
The L is 8 m bytes, and m is an integer of 2 or more.
It is assumed that double-length data is used. At this time, m can take a value such as "2", "4", "6", "8". For example, by setting the value of m to "8" and setting the basic length data length to 64 bytes, it is possible to configure n times long data which is n times 64 bytes. In this way, by increasing the processing unit, high-speed transfer processing can be realized. According to the current technical level, it is sufficiently possible to set the basic length data length to 64 bytes. Also, by increasing the processing unit in this way,
The idea of realizing high-speed transfer processing is disclosed in
There is no technology disclosed in Japanese Patent No. 4904.

In the example shown in FIG. 5, the L byte is 8 bytes, and the integer n is 2 or more.

In the example shown in FIG. 6, the L bytes are 53 bytes, and the basic length data at the beginning has a header area for the n-times length data of 5 bytes and has another (L × n) -5 bytes. Is the payload area of this n-fold data.

The example shown in FIG. 8 uses n-fold length data in which n integers are connected to a basic length data of 53 bytes having a header area of 5 bytes and a payload area of 48 bytes.

As shown in FIG. 9, the header area includes:
It includes information indicating the length of the n-fold data. At this time,
The information indicating the length is, as shown in FIG. 10B, a part of the VPI area in the header area or FIG.
As shown in (c), it is included in a part of the VCI area.

When the n-fold data structure shown in FIG. 8 is used, the header area contains the basic length data including the header area as the basic length data at the last position in the n-fold data. Contains information indicating whether or not there is. The information indicating whether the data is the basic length data at the last position is shown in FIG.
As shown in FIG. 10B, a part of the VPI area in the header area, or as shown in FIG. 10C, a part of the VCI area, or as shown in FIG. Included in part of the PT region. Further, the header area may include information on the number of basic length data subsequent to the basic length data including the header area. Further, the header area may include information indicating whether or not the basic length data including the header area is located at the head of the n-fold length data.

In the example shown in FIG. 11, the payload area may include information indicating the length of the n-fold data. The information indicating the length is included in a part of the AAL area in the payload area. When the n-fold data structure shown in FIG. 8 is used, whether or not the basic length data including the payload area is the basic length data at the last position in the n-fold data is included in the payload area. May be included. The information indicating whether the data is the basic length data at the last position is included in a part of the AAL area in the payload area. In this case, FIG.
As shown in the figure, the relay devices L1, L2 and the receiving terminal R
1 and R2, the cell buffer 10 for temporarily storing n-times long data, and the queue length monitoring unit 13 according to the information read by the header identification unit 15 indicating whether or not the basic length data at the final position is present. If the detected free space of the cell buffer 10 is smaller than the length of the basic length data, and the basic length data constituting the n-fold length data being currently written is not the basic length data at the final position, The configuration may include a read cell discarding unit 16 which is a unit for discarding all the n-times long data currently being written.

The relay devices L1 and L2 are n
Variable length cell that is double length data has basic length data length is ATM
In the case of being constituted by a basic length cell having the same length as the cell, as shown in FIGS. 18 and 19, route information of the variable length cell is extracted, and the variable length cell is divided into basic length cells by means. A first stage 21 including a certain buffer section 32 and a tag buffer 33 which is a means for adding route information to the basic length cell, and a basic length cell transmitted from the first stage 21 A second stage 22 exchanged and connected according to the following, an output buffer 42 which is means for deleting the route information of the basic length cell exchanged and connected by the second stage 22, and assembling the basic length cell into the variable length cell. And a third stage 23 including

The route information includes information on the input port at which the variable length cell has arrived and information indicating the basic length cell at the final position. The output buffer 42 stores the information on the input port and the basic length at the final position. A variable length cell is assembled for each input port according to the information indicating the cell. Here, the basic length cell has been described as having the same length as the ATM cell. However, it is possible to operate even if the length is arbitrary, such as 64 bytes.

[0047]

(First Embodiment) A first embodiment of the present invention will be described. In the transmission terminals T1 and T2 shown in FIG. 1, user information is stored in n-fold data. At the time of this storage, the user information is packed into the payload area of the n-fold data from the header area of the n-fold data. As a matter of course, the size of the payload area of the n-fold data may be the same as the size of the user information, or the payload area may be larger than the user information. In the latter case, a vacant area of the payload area is generated, so padding is performed with a bit string that can be distinguished from user information. For the bit string of the padding part, padding such as all "0" or "1" can be selected. After the user information is stored in the n-fold data as described above,
Sent to the communication network.

At the relay devices L1 and L2 in the middle of the communication network, n-times long data transmitted from the plurality of transmitting terminals T1 and T2 arrives. The relay device L1 of FIG.
The data from T2 is multiplexed as each block.
In this case, the cells of the data D1 from the transmitting terminal T1 and the cells of the data D2 from the transmitting terminal T2 do not coexist in the relay device L1 as in the related art. To do so, the multiplexing relay device L1 temporarily stores the data D1 from the transmitting terminal T1 and the data D2 from the transmitting terminal T2, transfers the data D1 from one transmitting terminal T1 first, and then transfers the data D1 from another transmitting terminal T1 to the other. The data D2 from the transmitting terminal T2 is transferred.

The relay device L2 in FIG.
1. Demultiplex the data D1 and D2 from T2.
At this time, the data D1, D2 from the transmitting terminals T1, T2
Always arrives at the relay device L2 in the order of n-fold data as a set of the data D1 and D2. Therefore, in the relay device L2, if only the transfer destination is known, the transfer can be sequentially performed by the data D1 from the transmitting terminals T1 and T2 and the n-fold length data in the unit of D2. In FIG. 1, the transmitting terminal T1,
The data D1 and D2 in n-fold length data units from T2 only need to be sequentially transferred to the receiving terminals R1 and R2, respectively.

Finally, in the receiving terminals R1 and R2, the n times longer data of the data D1 and D2 received from the transmitting terminals T1 and T2 are simply removed from the user information data D1 for one original data simply by removing the header area and the padding part. , D2 can be reproduced. Here, in the related art, a cell delay as a time interval for receiving from the first cell to the last cell is large because cells of another communication are interleaved between cells, whereas in the present invention, the transmission terminals T1, When the data D1 and D2, which are n times as long as the data D1 and T2, are received, the user information data D1 and D2 are fixed, so that the delay can be reduced.

Further, consider a case where data is lost in a part of the ATM communication network (cell loss occurs in the related art). In the prior art, the original data D1 and D2 of the transmitting terminals T1 and T2 are divided into a plurality of cells and transferred. Therefore, even when one cell loss occurs, the other cells constituting the same original data D1 and D2 are transferred by the respective relay devices and received by the receiving terminal R
1 and R2. Then, after the receiving terminals R1 and R2 perform de-celling (reconstruct the original data from all the received cells), the original user information data D1 and D2 as a result of the de-celling are found to be strange. The receiving terminals R1 and R2 request the transmitting terminals T1 and T2 to resend the data D1 and D2.

On the other hand, in the present invention, when there is a loss in the basic length data of the unit corresponding to the cell in the prior art,
If the basic length data is lost in the lossy relay device L1, L2 or the transmission line, the transfer process of the n-times long data including the lost basic length data can be stopped in the relay device following the transfer. That is, useless data cell transfer such that the data D1 and D2 cannot be reproduced to the receiving terminals R1 and R2 is not performed.
Therefore, an inadvertent increase in traffic in the communication network can be suppressed.

The operation of the cell buffer according to the first embodiment of the present invention will be described with reference to FIGS. FIG. 3 is a flowchart showing the operation of the cell buffer unit according to the first embodiment of the present invention. The cell buffer unit shown in FIG. 2 is provided in the relay devices L1 and L2 and the receiving terminals R1 and R2 shown in FIG.

The header identification section 15 shown in FIG. 2 refers to the header area of the cell that has arrived (S1). As a result of the reference,
Whether the information written in the header area is data length information,
Alternatively, it is determined whether the data is information of the last data (S2). That is, as shown in FIG. 4, if the data is n-length data having one header area at the head position of the n-length data, information on the data length of the n-length data is written in the header area. ing. Also, as shown in FIG. 8, if the header area is n-fold data in which a header area is written for each of a plurality of sections of n-fold data, the header area includes the number of subsequent basic length data. In some cases, information on the data length of n-times long data is written, and in some cases, final data information indicating the last basic length data is written. The header identification unit 15 shown in FIG. 2 can cope with any of those cases.

When the data length information is obtained from the reference result of the header identification unit 15, the data length is recognized (S
3) By comparing with the queue length information monitored by the queue length monitoring unit 13, it is determined whether or not to enter the cell buffer 10 (S4). As a result of the determination, if the data does not enter the cell buffer 10, all the basic length data constituting the n-fold length data are discarded (S6). If it enters the cell buffer 10, it is stored in the cell buffer 10 (S5).

When the final data information is obtained from the reference result of the header identification unit 15 (S2), it is determined whether or not the basic length data is the final basic length data (S2).
7). When it is determined that it is the final basic length data,
It is determined whether to enter the cell buffer 10 (S4). That is, if the capacity of the cell buffer 10 is smaller than the basic length data length, the final basic length data cannot be stored in the cell buffer 10. In this case, a part of the n-fold data already stored in the cell buffer 10 is discarded when the read cell discard unit 16 reads out the data from the cell buffer 10. As a result, the n-fold data with a part of the information missing does not arrive at the next-stage relay device or receiving terminal. If the capacity of the cell buffer 10 already runs out before the final basic length data arrives, the n-fold length data currently being written at that time is discarded. As a result, in the communication network of the first embodiment of the present invention shown in FIG. 1, switching is performed in units of one n-fold data, and an empty cell may be mixed in one n-fold data or a cell of another data may be mixed. Is not mixed.

In the following, the structure of n-fold length data will be described. In the example shown in FIG. 4, when L is the basic length data length and n is a natural number, the length of the n-fold length data can always be expressed as L × n. A description will be given of the use of the n-fold data in which the cell structure is determined as described above. In the transmitting terminals T1 and T2, first,
1 data D1 and D2 of user information (IP packet or MA
Based on the size Lu of the C frame, the size Lv of the n-times long data is determined with the most compact size. Here, Lb is the basic length data size, and Lh is the header size of n-times long data. At this time, the size Lv of the n-fold data can be determined based on the smallest min (N) satisfying N ≧ (Lu + Lh) / Lb. That is, Lv = Lb × n = Lb × min (N). Here, the information on the size Lu of the user information includes the MA in the IP header area when handling the IP packet.
Even when the C frame is user information, the header area has size information, and can be used to determine the size Lv of n-fold data.

As described above, in the first embodiment of the present invention,
In FIG. 4, it is assumed that L is 8 m bytes and m is an integer of 2 or more and is n-times long data. At this time, m is “2”, “4”, “6”,
It can take a value such as "8". For example, by setting the value of m to "8" and setting the basic length data length to 64 bytes, it is possible to configure n times long data which is n times 64 bytes. In this way, by increasing the processing unit, high-speed transfer processing can be realized. According to the current technical level, it is sufficiently possible to set the basic length data length to 64 bytes. In addition, there is no idea of increasing the processing unit and realizing high-speed transfer processing in the technique disclosed in Japanese Patent Application Laid-Open No. 7-254904.

The example shown in FIG. 5 is an example in which the basic length data length has a header area of 8 bytes and a payload area of an integral multiple of 8 bytes. The processing in the n-fold data is
Since the length of the n-fold data is an integral multiple of 8 bytes,
By processing in units of bytes, header area processing and transfer processing can be performed.

Thus, the header area of the cell is 8 bytes,
Assuming that the payload area is an integral multiple of 8 bytes, data in units of 8 bytes is generally a unit that can be easily handled by a processor or a dedicated communication LSI. Therefore, the transmitting terminals T1, T
2, when the user information is stored in the n-fold data, when the relay devices L1, L2 perform the transfer process, when the receiving terminals R1, R2
Various processing such as when user information is extracted from n-times long data in step 2 can be easily executed by a general-purpose processor or a communication-dedicated LSI.

As a result, in order to realize the communication network of the present invention, it is only necessary to slightly improve the ATM switch used in the existing relay device. The improvement is that the basic length data of the number corresponding to the length of the n-times long data is buffered or transferred continuously.

The example shown in FIG. 6 is an example in which the basic length data length is the current cell structure including a header area of 5 bytes and a payload area of 48 bytes. The transfer processing of the n-fold data is performed by setting the length of the n-fold data to an integral multiple of 53 bytes.

FIG. 7 shows an example in which a bit string at a position corresponding to the header area of the basic length data of n times longer data than the basic length data length is used as a payload area continuous with the preceding and succeeding payload areas. This is an example. The entire payload area of the n-fold data can be used so that user information can be stored. Therefore, as described in the example of FIG. 4, it can be determined by the length Lv of the n-fold data.

In the example shown in FIG. 8, the bit string at the position corresponding to the header area of the basic length data of n times longer data than the basic length data length is used as the header area without using it as the payload area. That is, this is an example of a structure in which n-fold data can be regarded as a group of concatenated basic length data. In this case, an area for writing header information is provided for each length Lb of the basic length data from the head position of the n-fold length data. Therefore, it differs from the example shown in FIG. 4 in that the size Lv of the n-fold data is determined as follows.

First, since the size of the payload area of the basic length data is Lb−Lh, N ″ ≧ Lu / (Lb−Lh) Therefore, one data of the user information (IP packet or M
Based on the size Lu of the AC frame, etc., the size of the most compact n-times long data Lv ″ is Lv ″ = Lb × min (N ″).

The user information is divided for each size Lb-Lh, and Lh, Lb + L
h, 2Lb + Lh, 3Lb + Lh,... Up to this point, the transmitting terminal T
1 is different in the processing of T2. Receiving terminal R1, R2
In extracting user information from n-fold data, Lh, Lb + Lh, 2Lb + Lh, 3
The original user information can be reproduced by extracting the data of the size Lb-Lh from the position of Lb + Lh,.

The example shown in FIG. 9 shows a case where an area indicating whether or not it is the length of the n-fold data or the part corresponding to the last basic length data is newly provided in the header area of the n-fold data. This is an example. A new area for adding information indicating the length Lv * (or Lv * ″) of the n-fold data is provided in the area for adding the header information of the n-fold data described above, or as shown in FIG. Information indicating whether the header area is the first header area in the n-fold data or the header area at the minimum position in the n-fold data is added to the header area provided for each length of the basic length data. A new area is provided.

When such information is added, the relay device L
1, L2 and the receiving terminals R1 and R2 can know the size of the n-fold data while receiving the n-fold data. By using this information, the relay apparatuses L1 and L2 can reduce the buffer amount required for each processing of accumulation and transfer, and the receiving terminals R1 and R2 can reproduce the original user information, and can shorten the processing time.

At this time, the size of the n-fold data is Lv *
(Or Lv * ″) is obtained as follows: the header size of the n-fold data: Lh + Lh *, a new area to add the length information of the n-fold data: Lh *, and the header area of the n-fold data. The length Lv * of the n-fold data in the case of only the head position is as follows: Lv * = Lb × min (N *) where N * ≧ (Lu + Lh + Lh *) / Lb. The length Lv * ″ of n-fold length data in the case of the basic length data for each length Lb from the beginning of the data is: Lv * ″ = Lb × min (N * ″), where N * ″ ≧ Lu / (Lb -Lh-Lh *).

In the example shown in FIG. 10, the area indicating the length of the n-fold data or whether or not it is the part corresponding to the last basic length data is set as a part of the VPI area, a part of the VCI area, or the PT area. This is an example of a case in which a part is provided. 10
(B) shows a case where 10 (c) is provided in a part of the VCI area and 10 (d) shows a case where it is provided in a part of the PT area. It is illustrated as a comparison object. When provided in the PT area of the conventional header area, functions other than "111" are currently assigned, so that this is not used as a bit string indicating whether or not it is the last basic length data equivalent part. There is no.

In the example shown in FIG. 11, the AAL (ATM adaptation layer) is provided with an area indicating whether it is the length of the n-fold data or the portion corresponding to the last basic length data. Specifically, this is an example of a case where an area is provided at the beginning of the payload area.

FIG. 12 shows the transmission terminals T1 and T2 that add the subsequent length to the area indicating the length of the n-fold data in each of the header areas periodically arranged in the n-fold data. This is an example of a case where the data is described in terms of the data equivalent number, the byte number, or the bit number and transmitted, and the subsequent length is represented by the basic length data equivalent number.

FIG. 13 shows the length of the first header area of the area indicating the length of the n-length data of each of the header areas periodically arranged in the n-length data. Is described in terms of the number of equivalent basic length data, the number of bytes, or the number of bits, and in the second and subsequent header areas of the same n-fold length data, bits indicating the second and subsequent bits are described and transmitted. This is an example of the case where the bits indicating the second and subsequent bits are indicated by “0 to 0”.

FIG. 14 is a diagram showing the order of cells after multiplexing at the ATM switch when IP packets are transferred by cells. FIG. 14A shows the case of a conventional 53-byte fixed-length cell, and FIG. 14B shows the case of n-times long data. In FIG. 14, four IP packets are formed into cells, each IP packet length is equal, and the input phase to the ATM switch is the same. The ATM switch performs polling from the upper input in the figure and outputs it to one output port.

In FIG. 14A, since a conventional fixed-length cell of 53 bytes is multiplexed, a cell containing a part of each IP packet in a payload area is mixed. On the other hand, FIG.
In (b), one IP packet is accommodated in one cell longer than the fixed-length cell with n-times longer data.
It is not divided by P packets. For the sake of comparison described later, the structure of the n-fold length data is such that the bit string at the position corresponding to the header area of the basic length data in the payload area as shown in FIG. Use as an area. In other words, the structure is such that n-fold data can be regarded as a consolidation of basic length data. Therefore, the cell length of one IP packet is the same in both cases.

FIG. 15 is a diagram showing a transfer delay time of an IP packet when the multiplexed cell stream shown in FIG. 14 is received. The propagation delay time and the assembling time of the IP packet are set to zero. Similarly to FIG. 14, FIG. 15A shows the case of a conventional 53-byte fixed-length cell, and FIG. 15B shows the case of n-times long data. It can be seen that the data transfer device of the present invention shown in FIG. 15B has a shorter transmission time of one IP packet than the conventional example shown in FIG. 15A.

FIG. 16 is a diagram showing the average transfer delay time of the IP packet shown in FIG. The horizontal axis is the number of fixed-length cells constituting one IP packet, or the ratio (n times) of the length of n-times data to the basic data length. The vertical axis is the average transfer delay time of the IP packet. Comparing the average transfer delay time of the IP packet, it can be seen that the data transfer apparatus of the present invention can transfer the packet in about 0.63 times shorter time than the fixed length cell system.

FIG. 17 is a diagram showing a computer simulation result obtained by comparing the average transfer delay time of an IP packet under different conditions. It is assumed that IP packets are randomly input from input ports of a plurality of ATM switches, and are multiplexed and output to one output port at a transmission line usage rate of 0.8. The length of the IP packet is 32 when the cell is formed into a conventional fixed-length cell, and the ratio of the length of the n-fold data to the basic length is 32 when the cell is formed into the n-length data. . The horizontal axis is the number of input ports of the ATM switch,
The vertical axis is the average transfer delay time of the IP packet. Comparing the average transfer delay time of the IP packet, it can be seen that the data transfer device of the present invention can transfer the data in about 0.6 times as short as the fixed length cell system.

(Second Embodiment) FIG. 18 shows a second embodiment of the present invention.
This will be described with reference to FIG. In the second embodiment of the present invention, the existing A
This is an embodiment for realizing a data transfer device applied to the present invention by adding functions to these based on a TM switch. Therefore, here, the n-fold data is called an integer multiple cell, the basic length data is called a basic length cell, and this length is called a basic cell length.

In the second embodiment of the present invention, the portion through which the main signal passes is constituted by three stages. Here, the first stage 21, the second stage 2
2. Called third stage 23.

First, in the first stage 21, when an integer multiple length cell with various VPI / VCIs is input to the input port, the integer multiple length cell is temporarily stored in the buffer unit 32. In this buffer unit 32, VPI / VC
I is extracted and transferred to the table 34. Table 3
In 4, a switching tag corresponding to the extracted VPI / VCI is generated and transferred to the tag buffer 33. Further, in the buffer unit 32, “n of the basic cell length” described in the “region indicating the cell length” in the integral multiple-length cell header is used.
Is detected, and the detection result is transferred to the counter 31. The counter 31 identifies the head of one integer multiple length cell by the multiple. Recognition of an integer multiple-length cell header is performed in the same manner as in the conventional fixed-length cell system.
This is performed by checking the operation of the C bit. The counter 31 recognizes that the input of the first integer multiple cell after the reset of the first stage 21 is the head of the integer multiple cell. The counter 31 transfers the multiplex number to the table 34, and the table 34 generates tags for the multiplex number.

In the buffer unit 32, the integer multiple cell is divided into a plurality of basic length cells, and the adder 35 divides a switching tag corresponding to VPI / VCI added to the head of the integer multiple cell. Is added to the head of each basic length cell.

This basic length cell is sent to the second stage 22. This tag is set in advance so as to be output to a predetermined output port corresponding to VPI / VCI. In the case of an integer multiple length cell, the length is n times the basic length cell. However, when the header is located only at the beginning of the integer multiple length cell, the subsequent portion is transferred to the basic length cell in order to transfer the subsequent portion to the integer multiple length cell. The tag obtained in the processing of the head part of the cell is stored, and is used to transfer the subsequent part for each basic length cell.
When the cell header is at the head of each integer multiple length cell, the same processing as when there is no cell header may be performed.
Tag information corresponding to the PI / VCI may be read, a tag may be added, and the tag information may be transferred to the second stage 22. For this reason, the processing method of the second stage 22 is based on the conventional AT for fixed-length cells.
Same as the M switch. Here, assuming that the basic length cell length is the same as that of the existing ATM cell, the second stage 22 can be realized by the existing ATM switch circuit (LSI).

In the third stage 23, the transferred basic-length cells are temporarily stored in the buffer unit 40. The buffer unit 40 deletes the tag from the basic length cell. This is the same as a conventional fixed-length cell ATM switch. Next, it is necessary to assemble the basic length cells into integer multiple length cells. In the second stage 22, transfer processing is performed in the same manner as in the conventional ATM switch for fixed-length cells, so that the basic-length cells arriving at each output port of the third stage 23 are the basic-length cells of other integer multiple-length cells. Arrives interleaved with. Therefore,
In the third stage 23, an output buffer 42 is provided for each input port at each output port, and outputs when the last portion of one integer multiple length cell arrives. For this operation, in addition to the path information used in the conventional fixed-length cell ATM switch, a bit indicating the input port number and the basic length data of the final position is added to the tag.
These controls in the third stage are performed by the output control unit 41.

The configuration of the second embodiment of the present invention follows the conventional relay devices L1 and L2 shown in FIG. 28 as it is, and the existing ATM switch can be easily converted to the data transfer device of the present invention. Can be. Thus, the data transfer device of the present invention can be realized with a small amount of time and cost.

(Third Embodiment) FIG. 19 shows a third embodiment of the present invention.
This will be described with reference to FIG. The third embodiment of the present invention is a block diagram of a main part of a data transfer device in the case of an integer multiple length cell having no data length information as shown in FIGS.
Further, in the third embodiment of the present invention, similarly to the second embodiment of the present invention, based on existing ATM switches, a function is added thereto to realize a data transfer device for an integer multiple cell. That is, as in the second embodiment of the present invention, the portion through which the main signal passes is constituted by three stages.

First, the input port of the first stage 21 is
When an integer multiple-length cell with various VPI / VCIs is input, the integer multiple-length cell is temporarily stored in the buffer unit 32. The input control unit 50 reads the “last basic length cell equivalent” described in the “area indicating whether or not the last basic length cell” in the header of the integral multiple length cell stored in the buffer unit 32. Is displayed, the end of one integer-long cell is identified, and the input of the next integer-long cell is recognized as the beginning of the integer-long cell. Recognition of an integer multiple-length cell is performed by using the CR
This is performed by checking the operation of the C bit. Note that the input of the first integer multiple length cell after the reset of the first stage 21 is also recognized as the head of the integer multiple length cell.

A tag corresponding to VPI / VCI is generated by the table 34 for the input of the integer multiple-length cell recognized in this way. This tag is transferred to the tag buffer 33, added to the head of the basic length cell by the adder 35, and sent to the second stage 22. This tag is VPI
/ VCI so as to be output to a predetermined output port in accordance with / VCI. Up to this point, it is the same as the conventional fixed-length cell ATM switch.

In the integer multiple length cell, the length is n of the basic length cell.
However, if the header is only at the beginning of the integer multiple length cell, to transfer the subsequent part for each basic length cell,
The tag obtained by processing the head portion of the integral multiple length cell is stored, and is used to transfer the subsequent portion for each basic length cell.
When the cell header is at the head of each integer multiple length cell, the same processing as when there is no cell header may be performed.
Tag information corresponding to the PI / VCI may be read, a tag may be added, and the tag information may be transferred to the second stage 22. For this reason, the processing method of the second stage 22 is based on the conventional AT for fixed-length cells.
Same as the M switch.

In the third stage 23, the transferred basic-length cells are temporarily stored in the buffer unit 40. In the buffer unit 40, the tag is deleted from the accumulated basic length cells.
This is the same as a conventional fixed-length cell ATM switch. Next, it is necessary to assemble the information divided and transferred for each basic length cell into an integral multiple length cell. In the second stage 22, the transfer process is performed in the same manner as in the conventional ATM switch for fixed-length cells. Arrives interleaved. Therefore, here, in each output port, the output buffer 42
Is output when the last part of one integer multiple length cell arrives. For this operation, the tag needs a bit indicating the input port number in addition to the path information in the second stage 22 used in the conventional ATM switch for fixed-length cells. These controls in the third stage are performed by the output control unit 41.

The output control section 41 determines whether or not the cell is the last part of the integer multiple length cell by (1) determining whether “the number of subsequent basic length data” in the cell header is zero, or (2) The length of the integer multiple-length cell, that is, whether the “length display area” has a value other than zero or (3) the “bit indicating that it is the last part of the integer multiple-length cell” in the tag.
In the case of (3), the “bit indicating the last part of the integer multiple length cell” is required in advance in the tag, and the bit value is set by the input control unit 50 of the first stage 21.

The configuration of the third embodiment of the present invention is a configuration that follows the conventional relay devices L1 and L2 shown in FIG. 28 as it is, and the existing ATM switch can be easily converted to the data transfer device of the present invention. Can be. Thus, the data transfer device of the present invention can be realized with a small amount of time and cost.

(Summary of Embodiment) The transfer efficiency at the cell level is compared between the existing ATM system and the present invention with reference to FIGS. FIG. 20 is a diagram showing the relationship between the cell payload length and the transfer efficiency in the case of a short packet,
The horizontal axis indicates the cell payload length (Byte), and the vertical axis indicates the transfer efficiency. FIG. 21 is a diagram showing the relationship between the cell payload length and the transfer efficiency in the case of a medium packet, where the horizontal axis indicates the cell payload length (Byte) and the vertical axis indicates the transfer efficiency. FIG. 22 is a diagram showing the relationship between the cell payload length and the transfer efficiency in the case of a long packet. The horizontal axis indicates the cell payload length (Byte), and the vertical axis indicates the transfer efficiency. First, when considering the single fixed-length method, the hardware technology is much more advanced than when the cell length of the existing ATM method is determined. Can be However, the single fixed-length method is a method in which high transfer efficiency cannot be expected for variable-length packets.

To show this, the transfer efficiency with respect to the cell payload length for three types of packet lengths of short / medium / long is shown in FIGS. FIG.
As can be understood from FIG. 22, if the cell payload length is shorter than the packet length, the packet is divided and transferred by a plurality of cells, and there are cell payload lengths with low transfer efficiency in some places. Also, where the cell payload length is longer than the packet length, the transfer efficiency gradually decreases as the cell payload length increases. Therefore, it is impossible to select a payload length such that “the transfer efficiency is always high for any packet length”.

Next, FIG. 23 shows a comparison of the transfer efficiency with respect to the packet length between the single fixed length system and the integer multiple length system. FIG. 23 is a diagram illustrating a comparative example of the transfer efficiency, in which the horizontal axis indicates the packet length (Byte) and the vertical axis indicates the transfer efficiency. The header length and the payload length of a single fixed-length cell were set to 5 bytes and 48 bytes, respectively, as in the standard ATM system. The header length and the basic length of the integral multiple length cell were also 5 bytes and 53 bytes, respectively. For the cell method, refer to LL
The C / SNAP bit and the last bit of AAL5 are omitted.

The difference in the transfer efficiency increases as the packet length increases. When the packet length is around 1500 bytes, the transfer efficiency of the integer multiple length system is 96.6 to 99.7%, whereas the transfer efficiency of the integer multiple length method is 96.6 to 99.7%. 90.6% at maximum with fixed length method
It is. In terms of OH rate, the integer multiple length method is 0.3% ~
In contrast to 3.4%, in the single fixed length system, it is at least 9.4%.

The average transfer efficiency of the packet length of 20 to 1500 bytes is 93.8% (OH rate of 6.2%) in the integer multiple length system, while it is 86.0% in the single fixed length system.
3% (13.7%), with a difference of 7.5%.

Therefore, the use of the integer multiple length method can considerably reduce the cell tax problem.

When a cell is transferred through the ATM network, the cell passes through the ATM switch. However, since the buffer amount of the ATM switch is not infinite, cell loss always occurs. FIG. 24 is a diagram showing an example of comparing the packet loss rate with respect to the buffer amount of the single fixed length method and the integer multiple length method, in which the horizontal axis represents the buffer size and the vertical axis represents the packet loss rate.

Assuming a 32 × 32 port ATM switch, each output port is equally loaded from each input port.
Excess arrivals that arrive randomly and exceed the output rate are accumulated in the output buffer. Current I
Since P packet transfer is generally performed in the best effort form, the simulation shows
Assuming that the average use efficiency of the output terminal is 0.999, the buffer amount was changed from 128 to 2048 transfer units (one transfer unit is 53 bytes). The length of the packet to be transferred is set to 1500 bytes from the maximum length of the IP packet transferred by the Ethernet.
Two cells and 32 transfer units.

In the single fixed length system, it is assumed that cells are interleaved in the switch, and even if one cell is lost in the buffer, a packet including the cell is discarded at the receiving terminal. In the integer multiple length method, a block of 32 transfer units is transferred in the switch, and if not completely stored in the buffer, 32 transfer units are discarded.

As can be seen from FIG. 24, the packet loss rate of the single fixed length system is about twice that of the integer multiple length system. This is the case of a single-stage switch, and the difference is further increased in consideration of the cell loss due to the passage through the multi-stage switch. Therefore, the dead cell problem can be considerably improved by using the integer multiple length method.

[0103]

As described above, according to the present invention,
Cell data can be processed without dividing one user's data, and the cell tax problem can be solved. As a result, the data throughput in the ATM communication network can be improved, the dead cell problem can be solved, and the data transfer delay time can be reduced. Further, since the existing ATM switch can be applied to the present invention, the data transfer device of the present invention can be realized in a short time.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a main part of a communication network according to an embodiment of the present invention.

FIG. 2 is a block diagram of a main part of a cell buffer unit according to the embodiment of the present invention.

FIG. 3 is a flowchart showing the operation of the cell buffer unit according to the embodiment of the present invention.

FIG. 4 is a diagram showing a configuration of n-fold data according to the embodiment of the present invention.

FIG. 5 is a diagram showing a configuration of n-fold data according to the embodiment of the present invention.

FIG. 6 is a diagram showing a configuration of n-fold data according to the embodiment of the present invention.

FIG. 7 is a diagram showing a relationship between data of user information and a payload area.

FIG. 8 is a diagram showing a configuration of a variable length packet having a plurality of header areas.

FIG. 9 is a diagram showing a configuration of a header area.

FIG. 10 is a diagram showing an information writing position in a header area.

FIG. 11 is a diagram showing an AAL area.

FIG. 12 is a view showing an example of header information of a variable length packet having a plurality of header areas.

FIG. 13 is a diagram showing an example of header information of a variable length packet having a plurality of header areas.

FIG. 14 illustrates an ATM when an IP packet is transferred in a cell.
FIG. 5 is a diagram showing the order of cells after multiplexing by a switch.

FIG. 15 is a diagram showing a transfer delay time of an IP packet when a multiplexed cell stream is received.

FIG. 16 is a diagram showing an average transfer delay time of an IP packet.

FIG. 17 is a view showing a result of computer simulation.

FIG. 18 is a block diagram of a main part of an ATM switch corresponding to a relay device and a relay network according to a second embodiment of the present invention.

FIG. 19 is a block diagram of a main part of an ATM switch corresponding to a relay device and a relay network according to a third embodiment of the present invention.

FIG. 20 is a diagram illustrating a relationship between a cell payload length and transfer efficiency in the case of a short packet.

FIG. 21 is a diagram illustrating a relationship between a cell payload length and a transfer efficiency in the case of a medium packet.

FIG. 22 is a diagram showing a relationship between a cell payload length and a transfer efficiency in the case of a long packet.

FIG. 23 is a diagram showing a comparative example of transfer efficiency.

FIG. 24 is a diagram showing an example of comparing the packet loss rate with respect to the buffer amount of the single fixed length system and the integer multiple length system.

FIG. 25 is a conceptual diagram of a conventional ATM switch.

FIG. 26 is a block diagram of a main part of a cell buffer unit used in a conventional ATM switch.

FIG. 27 is a conceptual diagram of a conventional ATM communication network.

FIG. 28 is a diagram showing a specific configuration example of a conventional relay device.

[Explanation of symbols]

 1-1, 1-2 Cell buffer unit 2 Address filter 10 Cell buffer 11 Cell discard unit 12, 14 Control unit 13 Queue length monitoring unit 15 Header identification unit 16 Read cell discard unit 21 First stage 22 Second stage 23 Third Stage 31 Counter 32, 40 Buffer unit 33 Tag buffer 34 Table 35 Adder 41 Output control unit 42 Output buffer 50 Input control unit D1, D2 Data L1, L2 Relay device R1, R2 Receiving terminal T1, T2 Transmitting terminal

Claims (28)

[Claims] 1. A transmitting terminal comprising means for accommodating user information in a unit of a data unit in a payload area of n times long data in which n basic length data are cascade-connected, and transmitting the user information to a communication network. A relay device including means for switching and connecting double-length data in n-fold data units by storage exchange, and means for receiving the n-fold data in n-fold data units and reproducing the user information from the payload area A data transfer device comprising: a receiving terminal comprising:
Where n is a natural number 2. The relay device and the receiving terminal,
2. The data transfer apparatus according to claim 1, further comprising means for discarding each of the n-times long data that arrives in n-times data units. 3. The data transfer device according to claim 1, wherein said transmitting means includes means for inserting a padding part into a surplus length when a length of said payload area is longer than a length of said user information. 4. The relay device and the receiving terminal,
a cell buffer for temporarily storing n-fold data, and means for comparing the free space of the cell buffer with the length of the arrived n-fold data in accordance with information indicating the length included in the header area of the n-fold data. 2. The data transfer apparatus according to claim 1, further comprising: when the free space of the cell buffer is smaller than the length of the n-fold data according to the comparison result of the comparing means, discarding the entire n-fold data. . 5. The n-fold length data has a cell structure of a cell length of L bytes as basic length data, and has a length of a bit string L × n (L is a natural number) obtained by connecting n variable integers to each other. The data transfer device according to claim 1. 6. A method according to claim 1, wherein one basic length data located at the head of said n-fold length data is a header area of said n-fold length data, and all other basic length data is a payload area of said n-fold length data. 6. The data transfer device according to 5. 7. The data transfer device according to claim 6, wherein said L is 8 m bytes, and m is an integer of 2 or more. 8. The data transfer device according to claim 7, wherein said m is 8. 9. The L byte is 53 bytes, and the first basic length data has a header area of 5 bytes of the n-times length data, and has a length of another (L × n) -5 bytes. n
6. The data transfer apparatus according to claim 5, wherein the data area is a payload area of double-length data. 10. The data transfer apparatus according to claim 5, wherein the basic length data of 53 bytes having a header area of 5 bytes and a payload area of 48 bytes each uses n times integer data connected by an integer n. 11. The data transfer device according to claim 6, wherein the header area includes information indicating a length of the n-fold data. 12. The header area includes the n
The data transfer device according to claim 10, further comprising information indicating a length of the double-length data. 13. The data transfer device according to claim 11, wherein the information indicating the length is included in a part of a VPI area in the header area. 14. The data transfer device according to claim 11, wherein the information indicating the length is included in a part of a VCI area in the header area. 15. The header area includes information indicating whether or not the basic length data including the header area is the basic length data at the last position in the n-fold length data.
0. The data transfer device according to item 0. 16. The data transfer device according to claim 15, wherein the information indicating whether the data is the basic length data at the final position is included in a part of a VPI area in the header area. 17. The data transfer device according to claim 15, wherein the information indicating whether the data is the basic length data at the final position is included in a part of a VCI area in the header area. 18. The data transfer device according to claim 15, wherein the information indicating whether the data is the basic length data at the last position is included in a part of a PT area in the header area. 19. The data transfer device according to claim 10, wherein the header area includes information on the number of basic length data following the basic length data including the header area. 20. The data transfer device according to claim 10, wherein the header area includes information indicating whether or not the basic length data including the header area is located at the head of the n-fold length data. 21. The data transfer device according to claim 6, wherein the payload area includes information indicating a length of the n-fold data. 22. The data transfer device according to claim 21, wherein the information indicating the length is included in a part of an AAL area in the payload area. 23. The data according to claim 10, wherein the payload area includes information indicating whether or not the basic length data including the payload area is the basic length data at the last position in the n-fold length data. Transfer device. 24. The data transfer device according to claim 23, wherein the information indicating whether or not the basic length data at the last position is included in a part of an AAL area in the payload area. 25. A cell buffer for temporarily storing n-fold data and a free capacity of the cell buffer according to information indicating whether or not the data is basic length data at the last position. Is smaller than the length of the basic length data, and when the basic length data constituting the n-times long data currently being written is not yet the basic length data at the last position, 24. The data transfer device according to claim 15, further comprising means for discarding all data. 26. The relay device, comprising: means for extracting route information of n-fold data, means for dividing the n-fold data into basic length data, and means for extracting the basic length data. A first stage including a means for respectively providing the obtained route information, a second stage including a means for switching and connecting the basic length data transmitted from the first stage in accordance with the route information, 2. The data transfer device according to claim 1, further comprising a third stage including means for deleting the route information of the basic length data exchanged and connected by the stage, and means for assembling the basic length data into the n-fold length data. . 27. The route information includes information on an input port at which the n-fold data has arrived and information indicating basic length data at a final position, wherein the assembling means includes information on the input port and the final position. 27. The data transfer apparatus according to claim 26, further comprising means for assembling n-fold length data for each input port according to the information indicating the basic length data. 28. The data transfer apparatus according to claim 26, wherein said basic length data is an ATM cell length.