JPH07307762A - Frame relay exchange and router - Google Patents

Abstract

PURPOSE:To prevent abolition of the frame of a data link connection identifier (DLCI) with a high priority by imparting the priority degree to the unit of the DLCI. CONSTITUTION:Priority or non-priority is registered in the DLCI management table 13C1 of this frame relay exchange 10 for the respective DLCIs by a report from a terminal in a STEP 1 and the priority of the DLCI of a received frame is decided by referring to the DLCI management table 13C1 in the STEP 2. Whether or not a congestion at the point of time exceeds a prescribed value is judged in the STEP 3 and the abolition processing of the frame of the DLCI of the non-priority received when the congestion exceeds the prescribed value is performed in the STEP 4. By performing such processings, the frame of the DLCI of the priority is inhibited from being abandoned by abandoning the received frame of the DLCI of the non-priority when the frame is received by the frame relay exchange 10 in a congestion state.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a priority control system for frame relay communication in which a destination of a frame is determined by an identifier for identifying the other party for communication.

In particular, a type that guarantees a committed information rate (CIR) is provided for each data link connection identifier (DLCI), and a transfer priority for each frame is set in the address field of a frame in the same DLCI. By specifying the drop priority and by providing a common buffer that can be shared by multiple line interface units in the frame relay switching device, priority control during frame transfer and frame discard can be minimized. It is a priority communication method of a relay communication system.

A frame relay communication system is an X.264 system used in a packet switching system used in data communication.
It is based on the 25 protocol and greatly simplifies the data transmission procedure to enable high-speed data transmission.

At present, since the transmission line is digitized, the reliability of the transmission line is improved and the probability of error occurrence in the transmission line during data transmission is low. In addition, high-speed data transfer is required for connection between LANs (Local Area Networks).

Therefore, in the frame relay communication system, the resending control when an error occurs is left to the checking function between terminals, and the main purpose of the network is to perform high-speed data transfer.

FIG. 24 is a diagram showing a connection between LANs by a frame relay network. 1 in the figure is a PVC (Perman) that connects a plurality of frame relay switching devices 10A to 10D and frame relay switching devices 10A to 10D to each other.
ent Virtual Circuit), 20 is a router for interworking for connecting the LAN 30 and the frame relay network 1, and 20a is a data link for connecting the frame relay switching device 10 and the router 20. Connection (DL
C Data Link Connection).

The router 20 includes a plurality of terminals T on the LAN 30.
Frames are multiplexed and transmitted on the data link connection 20a. A plurality of data link connections 20a are created on one physical line 20b.
Then, the frame relay switching device 10 identifies the ground, that is, the data link connection 20a, by the DLCI on the address field in the frame of the data link connection 20a, and sends the frame.

[0008]

2. Description of the Related Art In the above-mentioned configuration, a plurality of terminals T compete with each other to use the physical line 20b for frame relay communication.
Generally, the request level corresponding to the real-time property of the communication between the terminals T is different for each business. For example, a conversational application in which a person operates a keyboard to perform a business requires a fast response, and conversely, in file transfer, a high throughput is required but a real-time requirement is low.

In order to meet such various demands, when the router 20 sends a frame to the frame relay network 1, it gives priority control to the frame and gives priority to sending data having a high real-time property. .

However, since delay occurs even in the frame relay network 1, only the priority control by the router 20 is not sufficient.
It is necessary to perform priority control even inside.

Priority control is real-time (transfer)
It controls not only the priority of, but also the priority of discarding frames. For example, voice data needs to be set to have high real-time property as in conversational data communication.
Even if the call is partially cut off, it is possible to understand the content. Therefore, partial discarding of data is not required to have the same priority as data communication.

FIG. 25 shows a frame format of the frame relay network. From the beginning, the frame format is a flag field (F Flag Field), address field (A Address Field), frame relay data field (D Frame Relay Data Field), frame check sequence (FCS Frame Check Sequence),
It is composed of a flag field (F Flag Field).

In the address field A, a data link connection identifier (DLCI Data Link Connecti
on Identifier), Forward explicit congestion notification bit (FE
CNForward Explicit Congestion Notification), BECN Backward Explicion Notification bit
it Congestion Notification), discardable indication bit
(DE Discard Eligibility Indicator) is written.

The DLCI on the frame is information for identifying the communication partner to which the frame is transmitted, and the terminal T notifies the frame relay network 1 of the communication partner by the DLCI. Here, one DLCI corresponds to one communication partner, and by continuously sending out frames having different DLCIs, it becomes possible to communicate with a plurality of communication partners on one physical line.

Further, for example, the frame relay switching apparatus determines that congestion has occurred in the network when the usage rate of the processor or the usage rate of the buffer exceeds a certain reference value, and by FECN and BECN, The occurrence of congestion is notified to the receiving side terminal T and the transmitting side terminal T, respectively.

[0016]

When the congestion occurs in the frame relay switching device 10 in the network, the DE bit indicates that frame or a frame that may be discarded. When congestion occurs in the frame relay switching device 10, for example, the frame with DE = 1 is discarded.

There are two cases in which the DE bit is set to "1". One is the terminal T on the calling side.
This is the case where DE = 1 is set in the frame transmitted from the beginning, and the other one is when the data exceeding the information amount defined in the contract flows into the frame relay exchange device 10.
The frame relay switching device 10 uses the DE of the incoming frame.
If the bit is set to DE = 1 and congestion later occurs, D
Discard the frame set to E = 1. When the congestion level becomes higher, frames with higher priority are also discarded.

In this way, high-speed data transfer can be performed in the frame relay network 1, but it is unavoidable that data will be discarded when congestion occurs.

Depending on the type of data, such frame discard may not be permitted. For example, in the transfer of a computer program, if even one frame is discarded, the program may be incomplete and the computer may not operate normally. Therefore, there is a demand for a frame relay priority communication method in which important data is not discarded.

[0020]

FIG. 1 is a diagram for explaining a first embodiment of the present invention. STEP1 registers the priority / non-priority for each DLCI in the DLCI management table of the frame relay switching device according to the notification from the terminal, and STEP2
Determines the DLCI priority of the received frame by referring to the DLCI management table, STEP3 determines whether the congestion degree at that time exceeds a predetermined value, and STEP4 determines the congestion degree exceeds a predetermined value. Non-priority DL received when
The CI frame is discarded.

With such processing, when the frame relay switching device 10 receives a frame in the congestion state, the received frame of the non-priority DLCI is discarded, so that the frame of the priority DLCI is not discarded.

FIG. 2 is a diagram for explaining the second embodiment of the present invention. In the present invention, the DLC for displaying the DLCI assigned to the address field of the frame
Specific bits in the I field are used to indicate transfer priority / discard priority.

Priority control processing is performed in frame units from the following steps. STEP 21 registers the priority control level and the priority control processing of the frame relay switching device according to the priority control level, and STEP 22 assigns the priority to each frame received from the terminal by the router by referring to the table in the router. In STEP 23, the frame relay switching device refers to the address field of the received frame to determine the priority of the frame. In STEP 24, the frame relay switching device performs priority control according to the priority of the received frame.

As a result, the same DLCI sent by the terminal is transmitted.
It becomes possible to perform priority control for each frame above,
It enables priority control at a fine level. Furthermore, FIG.
FIG. 8 is a diagram illustrating a third embodiment of the present invention. Reference numeral 10 is a frame relay switching device, 11 is a management processor that manages the entire system, 12 is a plurality of line interface units 13, and a common buffer that can be used by a plurality of line interface units is provided. A communication multiplexer, 14 is a magnetic disk device, 15 is a magnetic tape device, 16 is a network monitoring device for monitoring the entire network, and 17 is a high-speed system bus.

By adopting such a configuration, when congestion occurs in the frame relay switching device, the frame received at that time is temporarily saved in the common buffer, so that the discarding of the frame can be prevented. Becomes

[0026]

In the first embodiment of the present invention, when the DLCI is registered, the priority / non-priority is registered in correspondence with the DLCI, and when the non-priority frame is received in the congestion state, it is regarded as the received frame. Frames having the same DLCI as the received frame in the transmission buffer are collectively discarded. By such processing, discarding of the priority DLCI can be prevented.

When a priority frame is received in the congestion state, the last received frame having the non-priority DLCI in the transmission buffer is collectively discarded. By such processing, discarding of the priority DLCI can be prevented.

Further, in the second embodiment of the present invention,
By using a specific bit of the DLCI field in the address field of the frame relay as a bit for instructing transfer priority / discard priority control, it becomes possible to perform more detailed priority control.

Further, in the third embodiment of the present invention, the common buffer 1 that can be used by a plurality of line interface units 13 is used.
30 is provided, and when congestion occurs, the frame is temporarily saved in the common buffer 130, and when the congestion is resolved, the frame is extracted from the common buffer 130 and transmitted, so that the frame can be prevented from being discarded.

[0030]

EXAMPLE 3 FIG. 3 is a diagram for explaining a third example of the present invention. In order to understand all of the present invention, the structure of the frame relay switching device 10 will be described with reference to FIG. As shown in the figure, the frame relay switching device 10 includes a system management processor (MPU Manegem) that manages the entire system.
ent Processor) 11, a plurality of line compatible units (LS Line Se
t) a communication multiplexer (CMU Communication) that accommodates 13
Multiplex Unit) 12, Magnetic disk unit (Magnetic D
isk Storage) 14, Magnetic Tape Device (Magnetic Tape St
orage) 15, network monitoring device (NSP Network S
It is composed of an up processor processor 16 and a high speed system bus 17.

Each line interface 13 is a subscriber line,
Or it is connected to a trunk line to another node. FIG. 4 is a diagram for explaining the configuration of the line interface, FIG. 4A shows the configuration of the line interface 13, and the line interface 13 is the bus controller 13.
It is connected to the high-speed system bus 17 via A, and the processing device (CPU Central Processing Unit) 13B is
Exchange processing is executed according to the program written in the memory 13C, and the received frame is sent to the designated line.

Further, the reception buffer 13D is a buffer memory for temporarily storing the frames received from the line, and the transmission buffer 13E is a buffer memory for temporarily storing the frames to be transmitted. Part 1
3F controls the line between the line interface 13 and the line.

(B) shows the structure of the line controller. The line control unit 13F is provided for the line,
A reception processing unit 13G that receives a frame from a reception channel line and a transmission processing unit 1 that transmits a frame to a transmission line
It is composed of 3H and performs line control at the time of frame transmission / reception.

FIG. 5 shows the connection processing of the frame relay network. In the figure, the frame relay switching device 10A having the node number 700 has
0), the frame received from the subscriber line (line number 0) accommodated in
This is an example of transmitting to the frame relay switching device 10B having the node number 701 via 0D.

Here, the terminal 30AT communicates with the terminal 30BT, but it is necessary to register each terminal in the frame relay switching apparatus 10 prior to the communication. The terminals 30AT and 30BT are connected to the frame relay switching device 10
Before connecting with A, 10B, it is necessary to identify the call and register the connection information in the frame relay network to establish the connection. Each line interface unit 13 of the frame relay switching device 10 has a DLCI for connection.
It has a management table and a PVC management table.

The connection information for identifying the call includes the physical line number to which the terminal is connected, the DLC number for identifying the input frame, the call number, the authorized information rate of the call,
Here, it is composed of 256 kbps and the priority added to the DLCI.

These pieces of information are stored in the network monitoring device 1.
6. It is registered in the DLCI management table of each line interface 13 through the system management processor 11.
FIG. 6 shows the structure of the management table of the frame relay network. (A) shows a DLCI management table, and by referring to this table, the frame relay switching apparatus 10A identifies that the call number of the frame having the DLCI 100 received from the 0th line is 100.

Further, (B) shows a PVC management table, and it is necessary to register the table for establishing the PVC in advance. This information consists of the destination node number, the call number at that node, and the line interface number to which the call is connected. FIG. 6B shows an example of the PVC management table registered in the originating node 701 shown in FIG.

After registering such a table, the terminal 30
When the AT calls the terminal 30BT connected to the node number 701 with the line number 0 and the DLCI number 100, the frame relay switching device 10A detects the DLCI 100 to the DLCI.
The management table is referenced to identify that the call number is 100. Then, by referring to the PVC management table with this call number 100, the node number 70
1, the line corresponding part number 10 and the call number 10 are identified.

FIG. 7 shows a format of a frame in the network. As shown in FIG. 7, the frame relay switching apparatus 10A on the originating side has information regarding the destination node, namely, 701, 10, 50 and 1 in the in-network management header thereof (the last numeral indicates the priority of the frame). Frame is inserted and the frame is sent to the next node. A relay node that has received a frame in the network from the preceding node via the relay line, for example, the frame relay switching devices 10C, 10
D refers to the in-network management header to determine a subsequent node and sends the frame.

The node of the communication partner which has identified that the frame is addressed to itself, here, the frame relay switching device 10B determines the line corresponding part number 10 and sets the DL.
The terminal 30BT is determined by referring to the CI management table,
A frame relay frame is assembled from the in-network frame and sent to the terminal 30BT.

In FIG. 7, the priority is provided by the present invention and will be described in detail later. FIG. 8 shows a routing process of the frame relay network. In FIG. 5, the number of routes from the originating node 700 to the destination node 701 is shown, but normally, there are multiple routes from one node to another node.

In FIG. 8, frame relay exchanges 10A, 10B, 10C, 10D and 10E (node numbers are 700, 701, 702, 703 and 7 respectively).
04)) to form a frame relay network. Each node determines the next node by referring to the in-network management header of the received frame with the routing table. Frame relay switching device 10A
When the frame is transmitted to the frame relay switching device 10B, the frame relay switching devices 10C and 10D refer to the in-network management header received from the line with the routing table so that the communication partner communicates with the frame relay switching device 1
It is identified as 0B, the route is determined, and the received frame is transmitted.

In a frame relay network,
The frame relay exchange 10 on the receiving side determines a terminal from the DLCI and sends the frame to the designated terminal. Such registration of DLCI is required by the declaration from the subscriber.
The NSP 16 is the DLC of the memory 13C of the line interface unit 13.
It is executed by writing in the I management table.

The first embodiment is a priority control system of a frame relay network which determines a partner terminal according to such DLCI and performs communication. Here, DLCI
DLC is given priority for each and non-priority DLC when congestion occurs
Discard the I frame and prevent the priority DLCI frame from being discarded.

In the memory 13C of the line interface 13,
In addition to the program for operating the processing device 13B, a DLCI management table for managing each DLCI is written. This DLCI management table is (1) line number, (2) DLCI number, as described in FIG.
(3) It is composed of DLCI connection information (call number) and CIR for the DLCI declared by the terminal.
In the present invention, in addition to the above-mentioned four items, priority information for identifying the priority of each DLCI is added.

When the subscriber wants to communicate with another terminal, it is necessary to declare the information speed to be used with the communication partner terminal in order to set the PVC for connecting the terminals on the frame relay network. Is.

The sum of the information speeds cannot exceed the capacity of each physical line. Therefore, such a registration request is allowed only if the physical line has a sufficient information rate to use.

If the request is granted, the requested data rate is the CIR Committed Information Rate.
Rate)) is called. FIG. 9 is a flowchart of DLCI priority registration.

This processing is executed when the subscriber declares the registration of a new DLCI with the required information rate. In S11, according to the declaration from the subscriber terminal,
It is determined whether or not the DLCI to be registered is the priority DLCI, and if the DLCI is the priority DLCI, the certified information rate (CIR) regarding the priority DLCI is added, and in S13, the sum of the CIRs of the priority DLCIs is added. Does not exceed the transmission capacity assigned to the priority DLCI. In S14, the DLCI to be registered is registered.
Is a non-priority DLCI, the DLC trying to register
Add 1 / n the CIR of I to the committed information rate for non-priority DLCI. Here, 1 / n is a scaling factor determined in consideration of discarding non-priority frames, and n is usually 5 to 10. In S15,
It is determined that the sum of 1 / n of the authorized information rates of the non-priority DLCIs does not exceed the bandwidth allocated to the non-priority DLCIs, and in S16, when the bandwidths allocated to the priority DLCIs or the non-priority DLCIs do not exceed, respectively. DLCI
Registration, priority / non-priority display, and update of CIR summation. In S17, when the sum of the CIRs of the priority DLCI or non-priority DLCI in S13 or S15 exceeds the bandwidth allocated to each, new registration is performed. Rejected and not registered.

FIG. 10 shows CIR for priority / non-priority DLCI.
The figure explaining registration of is shown. In the figure, the bandwidth of the physical line (for example, 1.5 Mbps) is used for the priority DLCI used by the priority DLCI (for example, 1 Mbps) and the non-priority D.
Non-priority DLCI used band used by LCI (for example,
It is divided into two bands of 0.5 Mbps).
In the priority DLCI, the certified information speed is 1: 1 and the priority DLC is set.
I, and non-priority DLCI is assigned n: 1. That is, in this example, 192 kbps has already been allocated to the priority DLCI, and the non-priority DLC has already been assigned.
For I, the total sum of the actual CIR is 896 kbps, but with the scaling factor n = 10, this is 1/10
It is evaluated as 9.6 kbps.

This processing means that registration can be performed up to 10 times the bandwidth allocated to the non-priority DLCI. When there is no congestion in the network, non-priority DLCIs are not discarded, but when congestion occurs, non-priority DLCs are not discarded.
The I frame is discarded.

By such processing, the frame of the preferential DLCI is prevented from being discarded, and the preferential communication can be performed. FIG. 11 shows an embodiment of the priority / non-priority DLCI control. According to the procedure described in FIG. 9, it is determined in advance for each DLCI whether the DLCI is the priority DLCI or the non-priority DLCI in the memory 1 of the frame relay switching device 10.
It is registered in the DLCI management table 13C1 of the 3C.
The CIR cumulative table 13C2 is written with the total of the authorized information rates of the priority / non-priority DLCIs, and this table is referred to when the DLCIs are registered.

In this state, (1) the frame relay switching device 10 receives a frame from the terminal through the subscriber line. (2) The processing device 13B of the line interface 13 uses the DLCI of the received frame as the DLCI management table 13C1.
, The priority / non-priority classification of the DLCI of the received frame, the communication partner are identified, and these are written in the in-network management header of the received frame. (3) The processing device 13B arranges the received frames in the transmission queue order in the transmission buffer 13E.
Write in. The in-network management header and the user data, which are given the priority classification in (4), are transmitted as an in-network frame to the next node through the relay line.

FIG. 12 shows an embodiment of buffer control.
(A) shows frame batch discard control at the time of receiving a non-priority frame. Transmission data is written in the transmission buffer 13E in the frame relay switching device 10. At this time, in order to detect the congestion state of the transmission buffer 13E, the processing device 13B determines the number of frames in the transmission buffer 13E and
The threshold value for determining the congestion state is compared, and when the number of frames in the transmission buffer 13E exceeds the threshold value, the processing device 13B determines the congestion state.

When a frame is received in such a state, the processor 13B identifies the priority / non-priority of the frame from the DLCI, and in the case of the non-priority frame, the received frame and the transmission buffer 13E are stored. The received frame already written to the frame and the frame of the same DLCI are collectively discarded.

(B) shows frame batch discard control at the time of receiving a priority frame. When the received frame is a priority frame, the last received non-priority DLCI frame written in the transmission buffer 13E is collectively discarded, and the received priority frame is transmitted.
Write in.

In this way, when congestion occurs, non-priority DLC
By collectively discarding I frames, communication regarding priority DLCI is guaranteed. In the first embodiment of the present invention, DL
Although priority is set for each CI and priority control is performed,
In the second embodiment, for each frame in the same DLCI,
The control is performed by giving priority.

For example, when performing inter-LAN communication through a frame relay network, the same DLCI is added to the data from a plurality of terminals accommodated in the LAN, and the data is sent out on the same data link. In such a configuration, the priority of frames from a plurality of terminals on the LAN is not the same, and the priority differs depending on the terminal or its application. Therefore, in order to perform priority control, it is necessary to give a priority to each frame of the same DLCI.

The priority of each frame is determined according to the address on the LAN and the protocol used by referring to the table of the router. At this time, L
The AN data is assembled as a frame relay frame in which a priority is given to each frame and sent to the frame relay network 1. The frame relay switching device 10 identifies the priority of the received frame and executes the priority control designated by the priority.

The structure of the frame used in the frame relay communication is as described in FIG. 25, and the communication partner is identified by the DLCI in the address field for communication.

FIG. 13 shows the structure of the address field of the frame relay frame. As shown in (A), the address field is made up of 2 octets (16 bits), the DLCI field in the address field is 10 bits, the FECN, BECN, and DE described above are each 1 bit, and the command / response indication bit C is included. / R (not used in frame relay) 1 bit, 1 per octet as address field extension indicator bit EA
Are using a bit. EA = 0 indicates expansion, and E
A = 1 indicates no expansion. Here, the address field consists of 2 octets, and EA = 1 of the second octet.

The DLCI field is 10 as shown.
Since it is a bit, the DLCI number is 0 to 10
Up to 23 can be registered. Considering the actual operation of the frame relay switching device 10, it is considered that it is almost unnecessary to set 1024 DLCIs for one line. Therefore, in the second embodiment of the present invention, the DLCI
The DLCI field is configured so that a specific bit among the 10 bits of the field is used as a bit for performing priority control.

(B) shows the structure of the DLCI field of the present invention. In the second embodiment of the present invention, a 10-bit DLC is used.
4 bits in the I field are used as priority control bits and 6 bits are used as DLCI.

In the present embodiment, as shown in the figure, the priority control bits are arranged so that the upper 2 bits are used as the transfer priority control bits and the next 2 bits are used as the discard priority control bits.

For example, when performing communication between LANs 30 through the frame relay network 1, a router 20 is provided between the LAN and the network. Router 20
Is provided with an address on the LAN 30 for each terminal and a table for setting the above-mentioned priority from the application being used. By referring to this table,
Transfer priority and discard priority are given in the address field.

(C) shows a priority control management table. This table is written in the frame relay switching device 10 for performing priority control. In the present invention,
Two bits are used for transfer priority control and two bits are used for discard priority control. Therefore, the priority control can be performed in four stages, and the priority of transfer or non-discard is set to be higher in the order of “11”, “10”, “01”, “00”.

FIG. 14 shows the structure of the transmission buffer. Frames sent from the frame relay switching device 10 to the relay line and the terminal line are stored in the transmission buffer 13E arranged for each line in the line corresponding unit 13, and are sequentially transmitted from the previously stored frames. It is transferred to the processing unit 13F and sent to the line.

The priority control related to the transfer is realized by performing the transfer process according to the priority when transferring the frame from the transmission buffer 13E to the transmission processing unit 13F. In the present invention, the transfer priority is divided into four levels.

Specifically, the processing device 13B identifies the transfer priority written in the address field of the received frame, performs chaining of the frame for each transfer priority, and writes it in the transmission buffer 13E. go.

The processing device 13B executes the transfer priority control by controlling the frequency of reading the frame from the transmission buffer 13E according to the four transfer priorities. In FIG. 14, the transmission buffer 13E is set to 2000.
In this example, each unit buffer 13e is composed of a data area having discard priority, effective data length, next frame start address, next buffer address, and area of 256 bytes as shown in the figure. Has been done.

A buffer management table for managing the transmission buffer 13E is provided in the memory 13C. In the buffer management table, a flag indicating busy / unused and a memory 13 for each unit buffer 13e.
The address for designating the buffer on C is written.

FIG. 15 shows the structure of the transmission management table.
In the present invention, the transfer priority is divided into four stages of the first priority to the fourth priority. The transmission management table is 4
It is for controlling transfer in accordance with one priority. The total number of unit buffers in use at all priorities is written at the beginning of the transmission management table, and then, for each priority, the unit buffer in use in the transmission buffer 13E of each priority is written. Of the total number of frames, the HEAD address indicating the address of the head unit buffer 13e in which the frame to be transmitted next in the same transfer priority is stored, and the frame to be transmitted last in the same transfer priority is stored. Unit buffer 13
A TAIL address indicating the address of e is written. In this way, the priority control of the written frame is performed based on the transmission management table.

FIG. 16 shows buffer chaining.
The frame of the user data exchange-processed in the frame relay exchange device 10 is stored in the transmission buffer 13E. In the present embodiment, since the data part of the buffer 13e is set to 256 bytes, when the frame of user data exceeds 256 bytes, one unit buffer 13e cannot be accommodated, and therefore one frame is composed of a plurality of frames. The data is divided and stored in the buffer 13e.

In FIG. 16, the data of the frame of the user data is (for example, 256 byte
e, is composed of 128 bytes),
The data is divided into units of 256 bytes, and each is stored in the three unit buffers 13e. Then, the address of the buffer 13e in which the next data is stored is written in the column of the next buffer address (BUF ADD).

In this way, when one frame is divided and written in a plurality of unit buffers 13e, if there is a designation in the next buffer address column, the unit buffers 13e are continuously transmitted. , It becomes possible to continuously output one frame divided and stored in the plurality of unit buffers 13e.

FIG. 17 shows chaining of frames.
The usage state of the transmission buffer 13E is written in the transmission management table described with reference to FIG. Therefore, when a new frame is to be stored in the transmission buffer 13E, the transmission management table is referenced to refer to the required number of unit buffers 1
Secure 3e. Specifically, the buffer management table is scanned, and the flag corresponding to the unused unit buffer 13e is determined to be in use, and the address of each unit buffer 13e is confirmed. Then, the number of secured buffers 13e is added to the total number of used buffers on the transfer management table and the value of the number of used buffers for each priority.

Further, referring to the TAIL address corresponding to the priority on the transmission management table, the next frame address in the unit buffer 13e indicated by the address is set to the head address of the buffer of the frame to be newly stored. Change to.

Then, the value of the TAIL address on the transmission management table is also changed to the head address of the buffer of the frame to be newly stored. In this way, by storing the frame on the transmission buffer 13E,
The entire frame on the unit buffer 13e can be chained as shown in FIG. The frame transmission from the transmission buffer 13E configured as above is executed as follows.

The processor 13B of the line interface 13 uses the buffer 1 indicated by the HEAD address on the transmission management table.
The data section 3e is transferred to the transmission processing section 13H. The HEAD address in the transmission management table is rewritten to the next frame address in the transmission buffer 13E in order to change to the subsequent frame.

When the next buffer address is written in the unit buffer 13e indicated by the HEAD address, the designated next unit buffer 13e also continuously transfers.

Unit buffer 13 for which transfer processing has been completed
e releases the buffer 13e by changing the flag indicating the used state on the buffer management table from the used state to the unused state.

The frame transfer starts from the unit buffer 13e designated by the HEAD address, the data in the buffer 13e designated by the next address is transferred, and when the buffer having no next address is transferred, one frame is transferred. finish.

The transfer priority control can be easily executed by controlling the frame transfer from the transmission buffer 13E provided for the transfer priority to the transmission processing unit 13H. That is, the priority control is performed by the transmission buffer 13
This can be realized by changing the ratio of frame transfer from E to the transmission processing unit 13H according to the transfer priority.

In the transmission processing, when the transfer of one frame from the transmission buffer 13E to the transmission processing unit 13H is completed, the transfer of the next frame is started. Therefore, the processing device 13B first refers to the transmission management table to obtain the head address (HEAD address) of the unit buffer in which the head of the frame is stored. The priority control is realized by performing the reference ratio of the HEAD address for each transfer priority.

For example, when the transfer priority is set as follows, the first priority: the second priority: the third priority: the fourth priority = 1.
By setting the number of times the HEAD table of each transfer priority of the 6: 8: 4: 1 transmission management table is referenced to 16: 8: 4: 1, the frames of each transfer priority are transmitted on the line according to this ratio. Sent to. At this time, for example, if the frame with the first priority is not present in the transmission buffer 13E, the frames with the second priority or lower are sent to the line at this ratio.

When the frame having the higher priority is absolutely prioritized with respect to the frame having the lower priority, the frame is not transferred according to the reference ratio of the HEAD address, but the frame is transferred. At the same time, the HEAD address of a higher priority than the priority of the frame to be transmitted is also referred to, and the frame to be transmitted is transferred only when the HEAD address of the higher priority does not exist. I will explain. In the present invention, the processing device 13B of the line interface 13 is configured as shown in FIG.
2 indicating the discard priority in the address field described in
Discard priority control is performed by referring to the bits.

The discard priority control is performed by the frame relay switching device 1
Perform in conjunction with a congestion level of zero. In the present invention, the congestion of the frame relay switching device 10 ranges from light congestion to heavy congestion, from 0th order congestion, 1st order congestion, 2nd order congestion,
It is managed in four stages of third-order congestion.

When congestion occurs, the BECN, FECN bit and integrated link layer management (CLLM) in the address field described in FIG. 13 are defined as specified in the ITU-T recommendation for frame relay.
Consolidated Link Layer Manegement) message notifies the terminal side that congestion has occurred.

When the congestion is not stopped and the congestion level rises even though the terminal is notified that congestion has occurred, the frame relay switching device 10 has a lower discard priority according to the congestion level. Discard from the frame.

In the present invention, the disposal process is performed as follows. 0th-order congestion; only congestion notification and no discard processing. Primary congestion: Frames of the third priority are discarded.

Secondary congestion: Discard the frames of the second and third priorities. Third-order congestion; frames of the first priority, the second priority, and the third priority are discarded. (Discarding All Frames) FIG. 18 shows a frame discarding process. The processing device 13B controls the reception processing unit 13G according to the congestion level, and performs the discard processing according to the discard priority of the received frame. When the processing device 13B transfers a frame from the transmission buffer 13E to the transmission processing unit 13H, the processing device 13B refers to the discard priority of the frame in the buffer 13e indicated by the HEAD address. Then, based on the congestion level and the discard priority at that time, it is determined whether or not the frame to be transferred should be discarded. If the frame is to be discarded, the transmission processing unit 1
No transfer to 3H is performed.

Then, after storing the next frame address value in the unit buffer 13e as the HEAD address of the transmission management table, the buffer 13e and the buffer 13e chained with the next buffer address are released from the buffer management table, Discard the frame.

The line interface 13 is provided with a transmission buffer 13E as shown in FIG. In such a configuration, in order to minimize the discard of frames, it is necessary to prepare a large memory area as the transmission buffer 13E of each line interface 13, and further, the memory area required for the frame relay switching device 10 as a whole. Will be huge.

Therefore, in the third embodiment of the present invention, the common buffer 13 that can be commonly used by a plurality of line interface units 13 is used.
0 is set. FIG. 19 shows the structure of the common buffer.
In the figure, a common buffer 130 is connected to a high-speed system bus 17 that transfers frames between the line interface 13 and the frame relay exchange 10. The common buffer 130 includes a bus control unit 13A, a processing device 13B, and a memory 13C, similarly to the line interface unit 13, and further includes a common buffer memory 13CB.

When the processing device 13B of the line interface 13 detects congestion, the frame received from the line is transferred to the common buffer 130, and when the processing device 13B stops detecting the congestion state, it is written in the common buffer 130. It is possible to prevent the discarding of the frame by taking out the stored frame and writing it in the transmission buffer 13E toward the transmission line.

FIG. 19 shows an example in which the area of the common buffer 130 is divided and assigned to each line interface 13. Such a configuration is also effective in reducing the cost of the line interface 13 as described below.

In FIG. 4, the transmission buffer 13E has a fixed capacity. However, the line interface 13 is connected to various lines having different data rates, such as a relay line and a subscriber line, and in order to realize communication without frame discarding, the memory capacity depends on the type of line. It is necessary to change. The frame relay switching device 10 accommodates a plurality of line corresponding units 13. Here, if the capacity of the transmission buffer 13E different for each line interface 13 is changed, the efficiency in the manufacturing process is reduced. Therefore, the common buffer 130 is provided with a memory area required for each line interface 13.
By providing it above, an effect of cost reduction is produced.

When congestion occurs, the processing unit 13B of the line interface 13 adds its own address and the address of the common buffer to the received frame and sends it to the high speed system bus 17. The processing device 13B of the common buffer 130 identifies and fetches the frame having the address addressed to the common buffer 130, and writes the frame in an arbitrary position of the common buffer memory 13CB. On the memory 13C, this common buffer memory 1
A table is created that manages the address of the transmission source line interface 13 for the frame written in 3CB and the write position of the common buffer memory 13C in association with each other, and divides the area of the common buffer 130 for each line interface. It can also be used in common without doing. With this configuration, the usage efficiency of the common buffer 130 can be improved.

When the processing device 13B of the line interface 13 stops detecting congestion, the line interface 13 of the transmission source requests the transmission of the frame written in the common buffer 130. The processing unit 13B of the common buffer 130 takes out the frame, adds the address of the line interface 13 of the transmission source, and transmits the frame onto the high-speed system bus 17. The line interface 13 that detects the frame having the address addressed to itself captures the frame.

After the frame is fetched, the frame is transferred by the same process as the frame received from another line and another line corresponding unit 13, and such a process can prevent the frame from being discarded.

FIG. 20 shows a flowchart of frame transfer to the common buffer. The figure illustrates the transfer and retrieval of frames to and from the common buffer 130 in the configuration of FIG. 19. The line interface 13 receives the frame in S31, and the processor 13B of the line interface 13 receives the frame in S32. In step S33, the received frame is transferred to the common buffer 130 in the case of the congestion state.
When the congestion continues, the received frame is continuously transferred to the common buffer 130 and written. S34
Then, the processing device 13B of the line handling unit 13 constantly monitors the congestion state and determines whether or not the congestion state has been resolved. When the congestion state is resolved in S35, the frame is taken out from the common buffer 130. In step S36, the extracted frame is written in the transmission buffer 13E to perform the transmission process. When it is determined in S32 that the congestion state is not present, the process proceeds to S36, and the received frame is written in the transmission buffer 13E as it is.

In FIG. 19, the common buffer 130 is independently provided, but the common buffer 130 of the line interface 13-1 accommodating an important line or a line in which congestion is likely to occur can handle the maximum congestion. It is also possible to set the capacity so that the area of the common buffer is used when congestion occurs in another line interface 13. With such a configuration, it is possible to minimize the capacity of the common buffer 130 without discarding frames of all lines.

Further, it is also possible to control the frame discarding process at the time of congestion of the common buffer 130 by the priority for each DLCI of the first embodiment and the four levels of priority designated in the address field of the second embodiment.

FIG. 21 is a diagram for explaining a router according to the fourth embodiment of the present invention. The function of the router 20 will be described with reference to FIG. When performing communication between the LAN 30A and the LAN 30B via the frame relay network 1,
The router 20A assembles the LAN frame of the LAN 30A into a frame relay frame and then sends it to the frame relay network 1, and the router 20B assembles the frame received from the frame relay network 1 into a LAN frame and then sends it to the LAN 30B. Fourth of the present invention
In this embodiment, a table for setting the priority from the address of each terminal on the LAN and the protocol used is registered in advance on the router 20, and the priority is given to the frame transmitted to the frame relay network 1. .

FIG. 22 shows an example of a LAN frame format. The frame format begins with the LAN protocol field for controlling within the LAN. The header field is a layer 2 protocol MAC (Media Acce
ss Cntrol) address, communication terminal 30 on LAN 30B
IP (Inter) as a layer 3 protocol for instructing BT
net Protocol) is included minimally. The user field following the header field contains the address of the terminal 30AT that is the transmission source as the application layer protocol.

A unique MAC address is given to the router 20A connected to the LAN 30A as one of the LAN terminals, and the router 20A uses the M address in the LAN frame.
LAN frame transmission / reception circuit 21 by identifying the AC address
A LAN frame is received through. The frame analysis unit 22 receives the LA received by the LAN frame transmission / reception processing unit 21.
The N frame is analyzed, and the LAN 30B of the communication partner is identified from the IP address in the LAN frame. Remote L connected to the LAN of the communication partner via the trunk lines B and C
In the case of AN, the frame analysis unit 22 refers to the registration table 27A in FIG.
Determine the value of I. This DLCI value is used to set a PVC for connecting to a communication partner in the frame relay network 1. Frame analysis unit 22
21 refers to the priority table 27B shown in FIG. 21 from the source address or the LAN protocol in the user data of the LAN frame, and also sets the priority for each LAN frame.

FIG. 23 shows a registration table in the router used together with the frame relay network using the exchange of the first embodiment described above. In this case, the source L
For communication between the AN 30A and the communication partner LAN 30B, a plurality of DLCIs having different priorities registered in the frame relay network 1 (B, B'in FIG. 21).
Are registered in advance in the frame relay network. For example, for each frame relay frame destined for the same remote LAN 30B, the DLCI having a different priority by referring to the priority table 20B.
Is given. Here, LAN 30B of the same communication partner
DLCIs with different priorities for, eg, 50, 5
1, 52 are written in the registration table in advance.

Frame relay (FR) frame assembly section 2
3 is LAN data, frame relay network 1
DLC corresponding to frame transfer / discard priority in the network
A LAN frame is assembled only by adding I. The assembled frame relay frame is sent to the frame relay line through the frame relay (FR) transmission / reception processing unit 24.

Next, the features of the router 20 used with the frame relay switch according to the second embodiment of the present invention will be described. In this case, the same DLCI is set in the frame relay frame addressed to the LAN of the same communication partner based on the registration table 27A shown in FIG. 23 (c) regardless of the priority, and the priority from the priority table 27B. The frame priority level is also set. The frame relay frame assembling section 23 assembles a LAN frame into a frame relay frame by giving identifiers of transfer priority and discard priority within the frame relay network. The assembled frame relay frame is a frame relay transmission / reception circuit 24.
Through the frame relay line.

As described above, the priority table for setting the priority from the LAN address and the LAN protocol is the router 2
Since it is provided in 0, the DLCI having the priority described in the first embodiment or the identifiers indicating the transfer priority and the discard priority described in the second embodiment are set in the above table. By reference, it is set in the address field of the frame when assembling a LAN frame into a frame relay frame. Then, the assembled frame is transmitted to the frame relay network 1.
Therefore, the frame relay network 1 performs priority control according to the DLCI or the priority identifier in the address field of the received frame. By such processing, discarding of important frames can be prevented, and frames with high real-time requirement can be preferentially transferred.

In the router 20B on the communication partner side, the frame received through the frame relay transmission / reception processing unit 24 from the frame relay line is sent to the frame analysis unit 25,
The frame is analyzed, sent to the LAN frame assembling unit 26, assembled as a LAN frame, and then sent to the LAN through the LAN frame transmission / reception processing unit 21.

[0113]

According to the present invention, by assigning a priority to a DLCI unit, it is possible to prevent discarding of a DLCI frame having a high priority.

[0114] Further, even for frames of the same DLCI, priority control can be performed even for frames of the same DLCI by giving transfer priority and discard priority for each frame. At this time, the transfer priority and the discard priority can be given by using a plurality of bits, so that finer priority control can be performed.

Further, by providing a common buffer accessible from a plurality of line corresponding parts, when a congestion occurs, the received frame is saved in the common buffer, and when the congestion is resolved, the frame is taken out from the common buffer and transmitted. Frame discard can be reduced.

Then, in the router, the LAN address, L
The AN protocol support is provided with a priority table for setting the priority, and when the LAN frame is assembled into the frame relay frame, the transfer priority and the discard priority are set by referring to the priority table, and the frame relay network is set. Then, priority control is performed based on the priority set in the address field of the received frame, so that discarding of important frames can be prevented and frames with high real-time property can be preferentially transmitted.

[Brief description of drawings]

FIG. 1 is a diagram illustrating a first embodiment of the present invention.

FIG. 2 is a diagram illustrating a second embodiment of the present invention.

FIG. 3 is a diagram illustrating a third embodiment of the present invention.

FIG. 4 is a diagram illustrating a configuration of a line interface.

FIG. 5: Frame relay network connection processing

FIG. 6 Structure of management table of frame relay network

[Fig. 7] Format of frame in network

FIG. 8: Routing processing of frame relay network

FIG. 9 is a flowchart of DLCI priority registration.

FIG. 10 is a diagram illustrating registration of CIR for priority / non-priority DLCI.

FIG. 11: Example of priority / non-priority DLCI control

FIG. 12: Example of buffer control

FIG. 13: Structure of address field of frame relay frame

FIG. 14: Structure of transmission buffer

FIG. 15 Structure of Transmission Management Table

FIG. 16: Buffer chaining

FIG. 17: Chaining of frames

FIG. 18: Frame discard processing

FIG. 19 Configuration of common buffer

FIG. 20 is a flowchart of frame transfer to a common buffer.

FIG. 21 is a diagram illustrating a router according to a fourth embodiment of this invention.

FIG. 22: Example of LAN frame format

[Fig. 23] Example of router table

FIG. 24: LAN by frame relay network
Connection

FIG. 25: Frame format of frame relay network

[Explanation of symbols]

 10, 10A to 10E Frame Relay Switching Device 11 Management Processor 12 Communication Multiplexing Device 13 Line Corresponding Unit 130 Common Buffer 13A Bus Control Unit 13B Processing Device 13C Memory 13C1 DLCI Management Table 13C2 CIR Cumulative Table 13CB Common Buffer Memory 13D Reception Buffer 13E Transmission Buffer 13e Unit buffer 13F Line control unit 13G Reception processing unit 13H Transmission processing unit 14 Magnetic disk device 15 Magnetic tape device 16 NSP 17 High-speed system bus 20, 20A, 20B router 20a Data link connection 20b Physical line 21 LAN frame transmission / reception processing unit 22, 25 frame analysis unit 23 FR frame assembly unit 24 FR frame transmission / reception processing unit 26 LAN frame assembly unit 27A registration table 27B Priority table 30, 30A, 30B LAN T, 30AT, 30BT terminal

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI Technical display location H04L 12/66 29/06 9466-5K H04L 11/20 B 9371-5K 13/00 305 C

Claims (17)

[Claims] 1. A frame relay network comprising a plurality of frame relay exchanges and a line connecting the exchanges, and a terminal which communicates through the frame relay network, and identifies a communication partner provided in the frame. In a foum relay communication system that communicates with a partner terminal based on an identifier, a table having priority information for the identifier is registered, and based on the priority information of the received frame, the table is referred to for communication processing of the frame. A frame relay exchange characterized in that priority is determined and communication processing for the frame is performed according to the determined priority. 2. The table is registered with an identifier to which priority information of high priority is given for discarding a frame, and the sum of communication speeds of frames having the identifier should be transmitted. The frame relay switch according to claim 1, wherein the frame relay switch is registered so as not to exceed a physical communication speed of the line. 3. The table further stores an identifier to which priority information of low priority is given for discarding a frame, and 1 / n (n is a sum of communication speeds of frames having the identifier). A number larger than 1) is registered so as not to exceed the difference between the physical communication speed of the line for transmitting the frame and the sum of the communication speeds of the frames having the high priority identifier. The frame relay switch according to claim 2. 4. The exchange has a transmission buffer for temporarily storing a frame to be transmitted, and when a frame is received in a congestion state, if the identifier of the received frame corresponds to low priority information, , The received frame and the frame having the same identifier as the received frame in the transmission buffer are collectively discarded, and when the identifier of the received frame corresponds to high priority priority information, 2. The frame relay switch according to claim 1, wherein frames having an identifier corresponding to priority information of low priority are discarded in a lump. 5. A frame relay network comprising a plurality of frame relay exchanges and a line connecting the exchanges, and a terminal which communicates via the frame relay network, and identifies a communication partner provided in the frame. In the Fum relay communication system that communicates with the other terminal based on the identifier, the frame is provided with a second identifier indicating the priority of the frame, and the exchange receives the frame, based on the second identifier. A frame relay exchange characterized in that a priority of communication processing for the received frame is determined, and communication processing for the frame is preferentially performed according to the determined priority. 6. The communication processing according to the priority includes a transfer priority control processing for transmitting a high priority frame first and a discard priority control processing for discarding a frame with a low priority frame first. The frame relay switch according to claim 1 or 5, wherein 7. The frame relay according to claim 5, wherein the second identifier indicating the priority is a transfer priority control bit and a discard priority control bit provided in a DLCI field of a frame relay frame. switch. 8. In the transfer priority control, a transmission management table for managing position information of frames stored in a transmission buffer is provided for each priority in the exchange, and a high priority table among the transmission management tables is provided. 7. The frame relay switch according to claim 6, wherein the frame relay exchange is performed by preferentially reading out the frame transmission control. 9. The frame relay switch according to claim 7, wherein the transfer priority control transfers a frame having a next priority when there is no frame having a higher priority in the transmission buffer. . 10. A frame relay network comprising a plurality of frame relay exchanges and a line connecting the exchanges, and a terminal which communicates via the frame relay network. The communication partner provided in the frame is identified. In the Foom relay communication system that communicates with the partner terminal based on the identifier, A frame relay switch having a common buffer that temporarily saves a frame to be stored in the transmission buffer when the number exceeds a fixed amount. 11. The frame relay switch according to claim 9, wherein the storage area of the common buffer is divided and allocated for each transmission buffer. 12. A management table for managing the frame to be saved in an arbitrary storage area of the common buffer, and managing the information in the storage area and the information for identifying the transmission buffer of the save source in association with each other. The frame relay switch according to claim 9, characterized by comprising. 13. The management table for writing the saved frame in an arbitrary storage area of the common buffer, and managing the information of the storage area and the information for identifying the transmission buffer of the save source in association with each other. The frame relay switch according to claim 9, characterized by comprising. 14. The frame relay switch according to claim 9, wherein the common buffer saves only frames having a high discard priority. 15. A router provided between a LAN and a frame relay network for transmitting and receiving data between each other is provided with a priority table in which priorities are registered in correspondence with LAN frame identifiers and received. A router characterized in that priority information of the LAN frame is obtained from the priority table based on the identifier of the LAN frame, and the LAN frame is converted into a frame relay frame to which the priority information is added and transmitted. 16. The router according to claim 15, wherein the priority information is added by including the second identifier according to claim 5 in a frame relay frame. 17. The addition of the priority information is performed by changing an identifier for identifying a communication partner to be added when converting a received LAN frame into a frame relay frame according to the priority. Characterized router.