JP2002516038A - Multipoint-to-multipoint echo processing in a network switch - Google Patents

Abstract

SUMMARY An apparatus and method are disclosed that does not interfere with valuable switching resources of a network switch used in a multipoint-to-multipoint switching scenario. The network switch is connected between the switch body, a plurality of input links and the switch body, and has an input processing port having a plurality of data buffer queues, and is connected between the switch body and the plurality of output links. And an output processing port having a data buffer queue. All data buffer queues have a connection identification code, and the output processing port data buffer queues have a data cell processing code. The input processing port stores a data cell received on one of the input links, a data cell, a link number indicating the input link at which the data cell has arrived, a port number indicating the input processing port, and a data cell buffered. Processing is performed by adding a connection identification code associated with the data buffer queue of the input processing port. The output processing port compares the link number of the data cell processed by the input processing port and transferred to the output processing port via the switch body with the link number of the link connected to the output processing port, and compares the port number with the output processing. The processing is performed by comparing with the port number of the port and comparing the connection identification code with the connection identification code associated with the data buffer queue of the output processing port. The data cells are then stored in the data processing queue of the output processing port according to the adaptation scheme indicated by the data cell processing code between the link numbers, between the port numbers, and between the connection identification codes.

Description

DETAILED DESCRIPTION OF THE INVENTION Multipoint-to-multipoint echo processing in a network switch Cross-reference to related application Claim priority to provisional application No. 60 / 001,498 (Title of Invention "Communication Method and Apparatus", filed on July 19, 1995). Field of the invention The present invention relates generally to network switching, and more particularly, to an apparatus and method that does not interfere with valuable switching resources used in a multipoint-to-multipoint switching scenario. Background of the Invention Telecommunications networks, such as asynchronous transfer mode (ATM) networks, are used for the transfer of voice, video, and other data. ATM networks deliver data by routing data units, such as ATM cells, from a source to a destination through a switch. Switches typically include multiple input / output (I / O) ports. ATM cells are transmitted and received via I / O ports. After the received ATM cells have been routed, the appropriate output port to be transmitted is determined based on the ATM cell header. In a multipoint-to-multipoint switching scenario, ATM cells from various sources are transferred from multiple input queues to multiple output queues inside the switch. In such a scenario, specific ATM cells passing through the switch can be eliminated by eliminating duplicate processing of the ATM cells or by selectively filtering the ATM cells before allowing the ATM to be forwarded through the switch. It is often beneficial to prohibit this flow. Only by allowing certain ATM cells to be forwarded through the switch, valuable switching resources are not obstructed. Therefore, in a multipoint-to-multipoint switching scenario, it is desirable to devise a scheme in which valuable switching resources of the network switch are not hindered. Summary of the Invention Apparatus and methods are disclosed that do not interfere with valuable switching resources in a network switch used in a multipoint-to-multipoint switching scenario. A network switch is connected between the switch body, a plurality of input links and the switch body, an input processing port having a plurality of data buffer queues, and connected between the switch body and the plurality of output links. An output processing port having a queue for data buffers. All data buffer queues have a connection identification code, and the data buffer queue of the output processing port has a data cell processing code. The input processing port handles the processing of a data cell received on one of the input links with a link number indicating the input link at which the data cell has arrived, and a data buffer queue of the input processing port where the data cell is buffered. This is performed by adding the attached connection identification code to the data cell. The output processing port processes the data cell processed by the input processing port and transferred to the output processing port via the switch body, compares its link number with the link number of the link connected to the output processing port, The number is compared with the port number of the output processing port, and the connection identification code is compared with the connection identification code associated with the output processing port. Next, the data cell is stored in the data buffer queue of the output port in accordance with the link number, port number, and connection identification code matching scheme indicated by the value of the data cell processing code. If the data cell processing code is the first value, the output processing port adaptation scheme is such that the link number, port number, and connection identification code match because the data cell is stored in the data processing queue of the output processing port. Request that When the data cell processing code is the second value, the output processing port adaptation scheme does not match the link number, port number, and connection identification number because the data cell is stored in the data processing queue of the output processing port. Request that This mechanism allows ATM cells to be transferred from multiple input queues to each output queue, and each output queue can independently receive sets of ATM cells from various sources. From the above summary, it is clear how the apparatus of the present invention can save valuable switching resources in a network switch. Accordingly, it is a primary object of the present invention to provide an apparatus and method that does not interfere with valuable switching resources of a network switch used in a multipoint-to-multipoint switching scenario. The above objects of the present invention, together with other objects, features and advantages, will become apparent from the following detailed description, which should be read in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE FIGURES For a more complete understanding of the present invention, reference is made to the accompanying drawings. These drawings are not to be construed as limiting the invention, but are intended only as examples. FIG. 1 is a block diagram of a network exchange. FIG. 2 shows the structure of the input queue. FIG. 3 shows the structure of the scheduling list. FIG. 4 shows a standard bus format of a data cell. FIG. 5 shows an exchange internal data cell format of the converted data cell. FIG. 6 shows the format of the input queue descriptor. FIG. 7 shows the format of the output queue descriptor. FIG. 8 is a table showing the various echo fields and the corresponding output port processor functions associated with these codes. FIG. 9 illustrates a “no echo” multipoint-to-multipoint exchange scenario. Detailed description of the invention Referring to FIG. 1, a network switch including a data crossbar 10, a bandwidth arbiter (BA) 12, a plurality of input port processors 14, a plurality of output processors 16, and a plurality of multipoint topology controllers (MTCs) 18. 1 is shown. The data crossbar 10 1, which may be an N × N crosspoint switch, is used for data cell transmission, and in this embodiment, generates a throughput of N × 670 Mbps. The BA 12 controls the switch interconnect, dynamically schedules instantaneously unused bandwidth, and resolves multipoint-to-point bandwidth contention. Each input port processor 14 schedules transmission of data cells from a plurality of connections to the data crossbar 10. Each output port processor 16 receives data cells from data crossbar 10 and organizes these data cells on output links. When traversing switch 1, data cells 22 first enter switch 1 via link 24, arrive at input port processing port 14, and are buffered in queue 26 of the input buffer. Next, the data cells 22 are transmitted from the input buffer queue 26 to the output buffer queue 28 of the output port processor 16 through the data crossbar 10. From the queue 28 of the output buffer, the data cells 22 are transmitted to a link 30 outside the switch 1 and transmitted, for example, to another switch. To facilitate traversing the switch 1, each input port processor 14 includes a cell buffer RAM 32 and each output port processor 16 includes a cell buffer RAM 34. Cell buffer RAMs 32 and 34 are organized into an input queue 26 and an output queue 28, respectively. All data cells 22 of the connection path must pass through one input queue and one output queue for the duration of the connection. Thus, queues 26 and 28 preserve the cell order. This measure also allows quality of service (QoS) guarantees on a per connection basis. To facilitate traversing the switch, three communication paths are used via probe / feedback messages. That is, the probe crossbar 42, the XOFF crossbar 44, and the XON crossbar 46. In this embodiment, the probe crossbar 42, which is an N × N crosspoint switch, is used to transmit the multiplex queue number from the MTC 18 to the output port processor 16. Each input port processor 14 includes a plurality of scheduling lists 47. Each of the scheduling lists 47 is a circular list containing the input queue numbers for a particular connection. Each multiplex queue number is obtained from information supplied to the MTC 18 from the scheduling list 47 of the input port processor 14. The multiple queue number identifies one or more output queues 28 to which data cells are transmitted when a connection is created. The output port processor 16 uses the multiple queue number to direct the request probe message to the appropriate output queue 28, thereby determining if there are enough output buffers available in the output queue 27 for data cells. In the present embodiment, the XOFF crossbar 44, which is an N × N crosspoint switch, is used to communicate a “transmission prohibited” type feedback message from the output port processor to the input port processor. The XOFF feedback message stops the request probe message from being sent from the input port processor 14 to the output port processor 16 via the probe crossbar 42 and puts the scheduling list 47 in the receiving input port processor 14 into the XOFF state. This means that it is not possible to provide multiple queue numbers using the scheduling list 47. As described below, the scheduling list 47 is held in the XOFF state until an XON message is received from the output port processor 16. Input port processor 14 responds to the issued XOFF message by changing the XOFF status bit in the descriptor of scheduling list 47. The XOFF bit inhibits the input port processor 14 from sending a request probe message to the output port processor 16 until the output port processor 16 is notified that the output buffer is available for the corresponding connection. . A "prohibit transmission" type feedback message also indicates that a data cell from input port processor 14 to output port processor 16 may not have enough buffer space available to receive the data cell. Prohibit transmission. In this case, the input port processor 14 does not transmit data cells via the data crossbar 10. Instead, an idle cell containing a complementary cyclic redundancy check (CRC) calculation is transmitted. In this embodiment, the XON crossbar 46, which is an N × N crosspoint switch, is used to communicate a “send permission” type feedback message from the output port processor 16 to the input port processor 16. More specifically, XON crossbar 46 communicates XON feedback messages from output port processor 16 to input port processor 14. When an XOFF feedback message is issued by the output port processor 16 in response to a request probe message from the input port processor 14, the output port processor 16 sets the status bit of the queue descriptor of the corresponding output queue 28. When the number of data cells in the output queue falls below the XON threshold, an XON message is sent from the output port processor 16 to the input port processor 14. The XON message allows the scheduling list 47 of the input port processor 14 to be used for sending request probe messages, and therefore for sending data cells. The probe / XOFF communication path operates in a pipeline manner. First, the input port processor 14 selects the scheduling list 47. Then, using the information associated with the scheduling list, the output port processor 16 or the output queue 28 to which the data cell is to be transmitted is determined. More specifically, the multiple queue number obtained from the information supplied to the MTC 18 from the scheduling list 47 of the input port processor 14 is transmitted from the MTC 18 to one or more output port processors 16 using the probe crossbar 42. Is done. Next, each output port processor 16 checks the availability of the buffer, and if the output buffer is not available for the connection, issues an "off-transmission" type feedback message via the XOFF crossbar 44. If an output buffer is available for the connection, input port processor 14 transmits the data cell to one or more output queues 28 via data crossbar 10. Each input port processor 14 in switch 1 also includes a switch assignment table (SAT) 20 that maps the bandwidth allocation. SAT 20 is the basic mechanism behind data cell scheduling. Each SAT 20 includes a plurality of consecutively ordered cell timeslots 50 and a pointer 52 that always points to one of the cell timeslots 50. All pointers 52 of switch 1 are synchronized such that, at any given time, each of pointers 52 points to the same cell time slot of the corresponding SAT 20, for example, the first cell time slot. In operation, pointers 52 are advanced one by one such that each cell time slot 50 is active for 32 clock cycles at 50 Mhz. If the pointer 52 points to the cell time slot 50, the input port processor 14 uses the corresponding entry 51 of the cell time slot 50 to obtain data for transmitting to the data crossbar 10. The content of each SAT entry 51, if valid, points to the scheduling 47. The content of each (non-empty) entry in the scheduling list 47 is composed of an input queue number. Each input queue number points to an input queue descriptor that contains state information specific to a particular connection. Each input queue descriptor points to the beginning and end of the corresponding input queue 26 that contains data cells for transmission via the data crossbar 10. If the SAT entry 51 does not contain a pointer to the scheduling list, that is, if the SAT entry 51 is set to zero, the corresponding cell time slot 50 of the SAT 20 is not allocated and its cell time slot 50 Is available for dynamic bandwidth. If the SAT entry 51 does not contain a pointer to the scheduling list 47 but the input queue number is not listed in the scheduling list 47, there is no data cell currently available for transmission, and the corresponding cell Time slot 50 is available for dynamic bandwidth. All unallocated bandwidth is called dynamic bandwidth. The dynamic bandwidth is granted by the BA 12 for certain types of connections so that the efficiency of the exchange 1 is improved. The exchange 1 is configured to manage connections having various quality of service attributes so that the characteristics of any connection do not interfere with any other connection. To realize this performance, the input port processor 14 manages each connection together with a set of data structures unique to each connection. To manage various resources, input port processor 14 uses two main data structures. One data structure is the input queue 26 and the other data structure is the scheduling list 47. Schematically, the input queue 26 is used to manage buffers. The input queue 26 is configured as a FIFO and includes a group of one or more buffers operated as a linked list structure using pointers. The incoming data cell 22 is added (enqueued) to the end of the input queue 26. The data cells sent to the data crossbar 10 are removed (dequeued) from the head of the input queue 26. The order of the data cells is always maintained. For a given connection, the order of the data cells sent to the data crossbar 10 is the same as when it arrived at the input port processor 14. However, the time interval between the departing data cells may be different from the time interval between the arriving data cells. FIG. 2 shows the structure of the input queue 26. The scheduling list 47 is used to manage bandwidth. The scheduling list 47 includes one or more input queue numbers configured as a circular list. Like the input queue 26, the scheduling list 47 is operated as a linked list using pointers. The input queue number is added to the end of the scheduling list 47 and is removed from the beginning of the scheduling list 47. For any given scheduling number 47, one input queue number appears only once. In addition to adding and removing, the input queue number can be circulated on the scheduling list 47 by removing the input queue number from the beginning of the scheduling list 47 and then returning the input queue number to the end of the scheduling list 47. In this case, the round robin service of the input queue 26 on the specific scheduling list 47 is performed. FIG. 3 shows the structure of the scheduling list 47. When a data cell 22 is received at the input port processor 14, the first processing performed by the input port processor 14 checks for errors in the header of the data cell and then associates the data cell with a valid connection. Is to check for The integrity of the cell header is confirmed by calculating a header error check (HEC) for the bytes in the header of the received data cell and comparing the calculated HEC to the HEC field of the header of the received data cell. If the calculated HEC and HEC fields do not match, a header error is present and the data cell is dropped. For each incoming data cell, input port processor 14 uses the VPI / VCI field specified in the header of the data cell as an index into input port processor 14's translation table. The translation table correlates the valid connection with the input queue number. The input port processor 14 first checks whether the data cell belongs to a valid connection, ie, a connection established by the switch control software. If the connection is valid, the input queue number is assigned to the data cell from the translation table. If the connection is not valid, the data cell is dropped or assigned an exception queue number from the translation table for further processing. The data cells are checked and converted from the standard data bus format to the switch internal data cell format. FIG. 4 shows a standard data bus format for data cells. FIG. 5 shows an exchange internal data cell format of the converted data cell. As described above, the input queue number is used to point to a queue descriptor that contains state information specific to a particular connection. There is a queue descriptor for each queue in switch 1, ie, for both input queue 26 of input port processor 14 and output queue 28 of output port processor 16. The queue descriptor is held by the switch control software. FIG. 6 shows the format of the input queue descriptor. FIG. 7 shows the format of the output queue descriptor. After an input queue number has been assigned to a data cell, input port processor 14 looks at the corresponding queue descriptor for more information on how to process the data cell. The input port processor 14 first attempts to allocate a buffer for a data cell. If a buffer is available, the data cell buffer number is enqueued at the end of the queue and the data cell is written out to cell buffer RAM 32. If no buffer is available, the data cell is dropped and the statistics are updated. In addition to processing and buffering the incoming data cell stream, the input port processor 14 must transfer data cells from the cell buffer via the data crossbar 10 to one or more output port processors 16. As described above, the data cell transfer is performed using the probe crossbar 42, the XOFF crossbar 44, and the data crossbar 10. More specifically, a multiplex queue number obtained from the information supplied to the MTC 18 from the scheduling list 47 of the input port processor 14 is transmitted from the MTC 18 to one or more output processors 16 using the probe crossbar 42. You. Each output port processor 16 checks for buffer availability and issues an "off-transmission" type feedback message via the XOFF crossbar 44 if no output buffer is available for that connection. If an output buffer is used for the connection, the input port processor 14 transmits the data cells via the data crossbar 10 to one or more output queues 28. However, before the data cells are enqueued in the output queue, output port processor 16 processes each data cell based on information contained in the trailer of converted data cells. With particular reference to FIGS. 6 and 7, both the input queue descriptor and the output queue descriptor include a connection identification (Conn ID) field 60. This field 60 is assigned by the switch control software and contains an arbitrary code indicating one of the eight possible data flow paths to perform the cell mask. When processing a data cell, input port processor 14 inserts this code from the connection identification field of the input queue descriptor into a similar connection identification field (Conn ID) of the converted data cell (see FIG. 6). The translated data cell also includes an incoming port number field 64 indicating the number of the input port processor 14 at which the data cell 22 was received, and an incoming link number field 66 indicating the number of the input link 24 at which the data cell 22 arrived. In. Note that the output queue descriptor also includes a 2-bit code "echo" field 68 that indicates what the output port processor 16 should do when processing data cells transmitted from the input port processor 14. It should be. This will be described later in detail. The code in the echo field 68 is also assigned by the switch control software. For each data cell transmitted from the input port processor 14, the output port processor 16 processes the data cell, and converts the port number, link number, and connection identification code of the data cell itself into the port number, link The number is compared with the connection identification code. The comparison result is used together with the 2-bit code of the echo field 68 of the output queue descriptor to determine whether or not the data queue arriving at the corresponding output queue 28 should be enqueued. FIG. 8 is a table showing the various echo fields and the corresponding output processor functions associated with these codes. For example, if the echo field 68 of the output queue descriptor is set to "00", the output port processor 16 will always enqueue data cells. In contrast, if the echo field 68 of the output queue descriptor is set to "11", the output port processor 16 will never enqueue a data cell. More important, however, is the operation of output port processor 16 when the echo field of the output queue descriptor is set to "01" or "10". Specifically, echo processing of data cells received by output port processor 16 saves switch 1 resources in a multipoint-to-multipoint switching scenario. To illustrate that the switching resources are conserved in the multipoint-to-multipoint switching scenario described above, it must be understood that switch 1 is often used within a network of similar switches. Data cells are routed through this network. In a multipoint-to-multipoint switching scenario, data cells from various sources are transferred from multiple input queues 26 within switch 1 to multiple output queues 28. In such a scenario, it would be beneficial to eliminate redundant processing of data cells or otherwise free up valuable switching resources by inhibiting the flow of specific data cells through switch 1. The echo processing of the data cells received by the output port processor 16 basically achieves the above object by selecting the converted data cells according to the port number, link number, and connection ID code accommodated therein. To achieve. If the echo field 68 of the output queue descriptor is set to "01", that is, in the "no echo" state, the output port processor 16 determines the port number, link number, and connection ID of the data cell. If the code does not match the port number, link number, and connection ID code of output port processor 16, the data cell is always enqueued. On the other hand, if the echo field 68 of the output queue descriptor is set to "10", i.e., in the "echo only" state, the output port processor 16 determines the port number, link number, and The data cell is enqueued only when the connection ID code matches the port number, link number, and connection ID code of the output port processor 16. FIG. 9 shows an example of a “no echo” multipoint-to-multipoint exchange scenario. In FIG. 9, a plurality of data cells (A, B, C, and D) correspond to a plurality of sources (T1, T2, T3, and S1) to a plurality of destinations (R1, R2, R3, and S2). Has been sent. More specifically, T1, T2, and T3 indicate terminal station transmitters, R1, R2, and R3 indicate terminal station transmitters, and S1 and S2 indicate other switching elements in the network. I have. The data cells are received by input port processors 14a and 14b, where they are processed and enqueued into input queues 26a, 26b, 26c, and 26d. The input port processor 14a and the output port processor 16a have the same port number. The input port processor 14b and the output port processor 16b have the same port number. Links 24a, 24c, 30a, and 30c all have the same link number. Further, the links 24b, 24d, 30b, and 30d all have the same link number. All input queues 26 and output queues 28 are assigned an arbitrary connection identification code “6”. As described above, processing of the data cells includes modifying the trailer of each data cell to include an arbitrary connection identification code, link number, and port number. In this example, the data cell A is assigned an arbitrary connection identification code “6”, a link number “24a”, and a port number “14a”. Similarly, an arbitrary connection identification code “6”, a link number “24b”, and a port number “14a” are assigned to the data cell B, and the arbitrary connection identification code “6” and the link number “24c” to the data cell C. , And a port number “14b”, and the data cell D is assigned an arbitrary connection identification code “6”, a link number “24d”, and a port number “14b”. For each data cell, the multiple queue number is simultaneously sent to output port processors 16a and 16b, whereby output port processors 16a and 16b each check for buffer availability. That is, output port processor 16a checks buffer availability for output queues 28a and 28b, and output port processor 16b checks buffer availability for output queues 28c and 28d. If sufficient buffers are available, the data cells are transmitted via data crossbar 10 and processed by corresponding output port processors 16a and 16b. In the "no echo" scenario, the output port processors 16a and 16b do not match the port number, link number, and connection identification number of the data cell with the port number, link number, and connection identification number of the output port processor 16a and 16b. If enqueuing a data cell. Thus, output queue 28a enqueues data cells B, C, and D, output queue 28b enqueues data cells A, C, and D, and output queue 28c enqueues data cells A, B, and D. , Output queue 28d will enqueue data cells A, B, and C. The connection identification code provides another control for filtering the source or destination from sending or receiving, respectively. This control enhances the selection of physical port and link numbers. As described above, echo processing uses a single set of connection resources, namely, input queues 26a-26d, scheduling list 47, and output queues 28a-28d, while each output queue 28a-28d stores data cells. Allows reception from various sets of sources. As described above, according to the echo processing, it is possible to prevent a valuable exchange resource from being hindered in the network exchange used in the multipoint-to-multipoint exchange scenario. It will be understood that various modifications and changes can be made in the method and apparatus described above without departing from the disclosed inventive concepts. Therefore, the present invention should not be construed as being limited to the above embodiments.

────────────────────────────────────────────────── ─── Continuation of front page    (81) Designated countries EP (AT, BE, CH, DE, DK, ES, FI, FR, GB, GR, IE, IT, L U, MC, NL, PT, SE), OA (BF, BJ, CF) , CG, CI, CM, GA, GN, ML, MR, NE, SN, TD, TG), AP (KE, LS, MW, SD, S Z, UG), EA (AM, AZ, BY, KG, KZ, MD , RU, TJ, TM), AL, AM, AT, AU, AZ , BB, BG, BR, BY, CA, CH, CN, CZ, DE, DK, EE, ES, FI, GB, GE, HU, I L, IS, JP, KE, KG, KP, KR, KZ, LK , LR, LS, LT, LU, LV, MD, MG, MK, MN, MW, MX, NO, NZ, PL, PT, RO, R U, SD, SE, SG, SI, SK, TJ, TM, TR , TT, UA, UG, US, UZ, VN (72) Inventor Caldara, Stephen A             Massachusetts, United States             01776, Sudbury, Horse Pond Low             C 220 (72) Inventor Hauser, Stephen A             Massachusetts, United States             01803, Burlington, Farms Dora             Eve 106 (72) Inventors Manning, Thomas A             Massachusetts, United States             01532, Northborough, Summer Street               26th (72) Inventor Vate, Stephen Earl             United States, Missouri, 63033, Bu             Rack Jack, Bristol Rock             Road 5082 [Continuation of summary] Processing the output data cell and link number Compares the link number of the link connected to the port and Compare the port number with the port number of the output processing port And output the connection identification code to the output processing port. Connector associated with the data buffer queue of the This is processed by comparing with the version identification code. next, Data cells are between link numbers, between port numbers, and Data cell processing code between action identification codes Output processing according to the indicated adaptation scheme It is stored in the port's data buffer queue.

Claims (1)

[Claims] 1. A network switch is connected to the input processing port and An output processing port having a queue for data buffers. Each of the data buffer queues has a connection identification code, and The data buffer queue of the processing port has a data cell processing code. A method that does not interfere with the valuable exchange resources of the network switch,   Receiving a data cell at an output processing port, wherein the data cell A link number indicating the input link at which the data cell has arrived, and A port number indicating the input processing port and the data cell Code associated with the data buffer queue of the input processing port. Receiving a connection identification code;   Link of the link connected to the output processing port with the link number Number and compare the port number with the port number of the output processing port. Comparing the connection identification code with the data of the output processing port. Comparing with a connection identification code associated with the buffer queue;   Between the link numbers, between the port numbers, and between the connection identification codes According to the adaptation scheme specified by the value of said data cell processing code, The data cell to the data processing queue of the output processing port. Storing. 2. The adaptation scheme is such that the data cells are queued in the data buffer queue. The link number, the port number, and the connection identification to be stored The method of claim 1 wherein the code requires a match. 3. The adaptation scheme is such that the data cells are queued in the data buffer queue. The link number, the port number, and the connection identification to be stored The method of claim 1 requiring that the code not match. 4. A network switch is connected to the input processing port and An output processing port having a queue for data buffers. Each of the data buffer queues has a connection identification code, and The data buffer queue of the processing port has a data cell processing code. A method that does not interfere with the valuable exchange resources of the network switch,   Receiving a data cell at an output processing port, wherein the data cell A link number indicating the input link at which the data cell has arrived, and A port number indicating the input processing port and the data cell Code associated with the data buffer queue of the input processing port. Receiving a connection identification code;   The link of the link number and the link connected to the output processing port Between the port number and the port number of the output processing port. And the data of the connection identification code and the output processing port. Between the connection identification code associated with the data buffer queue. The data cell according to the matching scheme specified by the value of the data cell processing code. Processing the file. 5. The step of processing the data cell includes the link number, the port number And, if the connection identification code matches, the data cell is output. Of the data processing port 5. The method of claim 4, further comprising the step of storing in a queue. 6. The step of processing the data cell includes the link number, the port number And if the connection identification code does not match, the data cell Storing in the data buffer queue of an output processing port. The method according to claim 4. 7. A network exchange that does not interfere with valuable exchange resources in the network exchange. What   Exchange body,   An input processing port connected between the plurality of input links and the switch body; Each of the data buffer queues has a connection identification code. The input processing port is received on one of the plurality of input links. Link to the input cell where the data cell arrived at the data cell. Port number indicating the input processing port, and the data cell Corresponds to the data buffer queue of the input processing port where Input processor to process by adding the attached connection identification code. Singport,   A plurality of data buffers connected between the switch body and a plurality of output links; An output processing port having a queue for   Each of the data buffer queues has a connection identification code, and The data buffer queue of the input processing port is a data cell processing code. Wherein the output processing port is connected to the input processing port. Processed and transferred to the output processing port via the switch body. The data cell is linked to the link number connected to the output processing port. Link number and the port number The connection identification code is compared with the port number of the Connected to the data buffer queue of the output processing port By comparing with the port identification number, Signal, and between the connection identification code, the value of the data cell processing code Converting the data cells to the output processor according to a matching scheme specified by A network switch for storing the data in the data buffer queue of the network port. 8. The output processing port adaptation scheme is such that the data cells are To be stored in the data buffer queue, the link number, the port number, 8. The network of claim 7, wherein the connection identification code is required to conform. Exchange. 9. The output processing port adaptation scheme is such that the data cells are To be stored in the data buffer queue, the link number, the port number, 8. The network according to claim 7, wherein said connection identification code does not match. switch.