JP2010103755A - Frame transfer system and frame transfer method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To select a pass or discard when a header is normal and a payload has an error by specifying which has a frame error, the header including a MAC address or the payload. <P>SOLUTION: In this frame transfer method for detecting an error of a frame comprising a header including an address and a payload by a first FCS (Frame Check Sequence), discarding the frame when the error is detected and making the frame pass when the error is not detected, a frame transmitting side generates the first FCS from the entire frame, and also generates a second FCS from the header, and a frame receiving side checks the first FCS and the second FCS, and specifies whether the header has an error from respective error detection states or whether only the payload has an error without an error in the header. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a packet transport technique for transferring an Ethernet (registered trademark) frame and an SDH signal as an Ethernet frame, a frame transfer system for distinguishing and detecting frame header and payload errors and selecting discard or passage of the frame, and The present invention relates to a frame transfer method.

  In a frame transfer system, there is FCS (Frame Check Sequence) as a method for detecting a frame error using a checksum (Non-Patent Document 1). In this method, all frames are calculated by a checksum on the data transmission side, and transmitted as an FCS at the end of the frame. On the other hand, on the receiving side, data is calculated with a checksum using the same algorithm used on the transmitting side, and the calculation result is compared with the received FCS. Thereby, an error in the frame can be detected. However, the position of the error in the frame cannot be specified and the error cannot be corrected.

  This FCS process depends on a forwarding mode which is a processing method from receiving data to transmitting it (Non-Patent Document 2). In the Ethernet as a forwarding mode, there are the following four types according to the method of performing error detection by FCS and the method of not performing it.

(1) Store & forward Buffers the frame to be forwarded to FCS, checks whether there is a frame error, and sends it to the target port. For example, if even one bit has an error, the frame is discarded and the frame is retransmitted.

(2) Cut-through High-speed transfer is possible because the destination MAC address of the frame to be forwarded is seen and sent to the target port immediately without buffering. However, since an error check by FCS is not performed, an error frame is forwarded, and thus there is a disadvantage that data reliability is lowered.

(3) Fragment free Transfer 64 bytes from the beginning of the frame. As with cut-through, error checking by FCS is not performed. Since 64 bytes is the minimum frame of Ethernet, a short packet can be discarded.

(4) Error-free cut-through Automatically selects store & forward and cut-through according to the frame error rate. When the error rate is high, the store and forward method is highly reliable, and when the error rate is low, the high-speed cut-through method is changed.

  Next, a processing method when there is an error in the MAC address will be described. When the address is rewritten due to a frame error, if the rewritten address exists on the sub-network, a transmission error occurs and becomes a problem. On the other hand, when the rewritten address does not exist in the sub-network, the conventional Ethernet performs flooding for sending data to all addresses in the sub-network. This places a heavy load on the network and increases the possibility of loops. Thus, the MAC address error becomes a big problem in operating the network.

  Also, in PBB-TE (Provider Backbone Bridging-Traffic Engineering), which is currently being standardized, there is a method of discarding a frame without performing flooding even when an address rewritten due to a frame error does not exist in the subnetwork. .

  On the other hand, payload errors can be processed according to the application used. For example, in the case of text data, since a 1-bit error leads to garbled characters, it is better to discard the frame and retransmit it. Some delay in this case is not a problem. However, in two-way real-time communication such as video and audio, the payload error appears as noise, but in real-time communication, delay is an important factor, so it is convenient not to discard and retransmit the frame for the payload error. Is good.

Also, if the store & forward method is applied when encapsulating and transferring SDH / SONET signals by packet transport technology, the frame is discarded with a 1-bit error. Spread and quality is greatly reduced.
崔 Fuyumei, “Network Fundamentals”, Soseisha, pp.7-9, April 10, 2008 http://www.allied-telesis.co.jp/library/nw guide / device / switch.html

  As described above, it is desirable to discard a frame for an MAC address error, but it is desirable to process a payload error according to the type of data carried in the payload. However, in the current FCS processing, it is not possible to specify whether there is an error in the header including the MAC address or the payload. Also, there is no method that can select passing or discarding according to the type of data to be placed on the payload when a frame error is detected.

  The present invention provides a frame transfer system and a frame transfer method that specify whether a frame error is in a header or a payload including a MAC address and can select passing or discarding when the header is normal and the payload has an error. The purpose is to do.

  A first aspect of the present invention is a frame transfer method for detecting an error of a frame including a header including an address and a payload by the first FCS, discarding the frame when the error is detected, and passing the frame when no error is detected. In the frame transmission side, the first FCS is generated from the entire frame and the second FCS is generated from the header, and the first FCS and the second FCS are checked on the frame reception side, From the detection status of each error, it is specified whether there is an error in the header, or there is no error in the header and there is an error only in the payload.

  The frame receiving side discards the frame if there is an error in the header from the detection status of each error by checking the first FCS and the second FCS, and if there is no error in the header and there is an error only in the payload Choose to pass or discard.

  On the frame transmission side, the pass / discard control tag used to select whether to pass or discard the frame is set in the header. On the frame reception side, the pass / discard control is performed when there is no error in the header and only the payload has an error. Perform processing according to the tag settings.

  The second FCS is placed at the end of the header, and on the frame receiving side, priority is given to the second FCS check, and if an error is detected in the second FCS, the frame is discarded and an error occurs in the second FCS. If it is not detected, the pass / discard control tag is confirmed. If it is set to pass, the frame is passed without checking the first FCS, and if it is set to discard, it corresponds to the error detection status by the first FCS. Processing may be performed.

  Also, the second FCS is arranged at the end of the previous frame, and on the frame receiving side, the first FCS check is prioritized, and if an error is detected in the first FCS, the second FCS is checked, If an error is detected in the second FCS, the frame is discarded regardless of the setting of the pass / discard control tag. If no error is detected in the second FCS, processing according to the setting of the pass / discard control tag is performed. You may do it.

  According to a second aspect of the present invention, there is provided a frame transfer system that detects an error of a frame including a header including an address and a payload by the first FCS, discards the frame when the error is detected, and passes the frame when no error is detected. The frame transmission side includes FCS generation means for generating the first FCS from the entire frame and generating the second FCS from the header, and the frame reception side includes the first FCS and the second FCS. Error position specifying means for checking whether there is an error in the header or whether there is an error in the header and there is an error only in the payload based on the detection status of each error.

  The receiving side of the frame discards the frame if there is an error in the header, and passes or discards the frame if there is no error in the header and there is only a payload, depending on the error position specified by the error position specifying means. Frame transfer control means for selecting is provided.

  The frame transmission side has a passing / discarding control tag setting means for setting a passing / discarding control tag used to select whether to pass or discard the frame in the header. The frame transfer controlling means on the receiving side of the frame has an error in the header. When there is an error only in the payload, processing is performed according to the setting of the pass / discard control tag.

  The present invention makes it possible to detect whether an error is in the header or in the payload, and in the case of a payload-only error, it is possible to select whether to pass or discard the frame, so that it is possible to transfer the frame according to the situation. .

  In particular, in communication with an emphasis on real-time characteristics, if there is only an error in the payload, it is possible to perform transfer with little delay without performing retransmission by setting to pass. In addition, when transferring an SDH signal by the packet transport technology, similarly, if a payload only error is set to pass, a 1-bit error that has been discarded in the past does not spread to a bit error for the payload. Quality improvement is expected. This is also useful when transferring a dedicated line with a strict bit error rate (BER).

  Further, by discarding a header error, erroneous transmission due to a destination error and flooding with an unknown address can be avoided, and the load on the network can be greatly reduced.

  The present invention is based on an error locating technique for identifying whether a frame error is in a header including an address or a payload, and a transfer control technique capable of selecting passing or discarding when the header is normal and only the payload has an error. Composed.

FIG. 1 shows a frame configuration example in the present invention.
In the figure, the frame is composed of a header 11 including an address and a payload 12. In order to specify the error position, as shown in FIGS. 1 (1) and (2), it is generated from the header 11 in addition to the existing FCS (first FCS) 13 generated from the entire conventional frame. The header FCS (second FCS) 14 is used. The header FCS 14 is arranged at the end of the header 11 or the end of the frame. If there is an error in the header 11, an error is detected from the existing FCS 13 and the header FCS 14. If the header is normal and there is an error only in the payload 12, the error is detected only from the existing FCS 13.

  In order to select passing or discarding when the header 11 is normal and only the payload 12 has an error, as shown in FIG. 1 (3), passing or discarding is set according to the data type of the payload 12 Use the discard control tag 15 The passing / discarding control tag 15 is arranged inside the header 11 or at the end of the header 11. The pass / discard control tag 15 is handled as a part of the header 11 and is checked by the header FCS 14.

  As an example of the pass / discard control tag 15, a 1-bit tag is used. When only the payload 12 has an error, a setting of “pass” is set to “0”, and a setting of “discard” is set to “1”. . Thus, when an error is detected only from the payload 12, it is possible to control passage or discard. For example, when the SDH signal is transferred by the packet transport, the passing / discarding control tag 15 is set to “0”, and when the IP signal is transferred, it is set to “1”. Thus, when an error is detected from the header 11, the frame is discarded regardless of the setting of the pass / discard control tag 15, but when an error is detected only from the payload 12, the pass / discard control tag 15 Depending on the setting, pass or discard is selected.

  By the way, although the position of the header FCS 14 is shown in two ways in FIGS. 1 (1) and (2), it is more efficient depending on the traffic volume of communication in which the passage / discard control tag 15 is set to pass or discard. Is selected.

  FIG. 2 shows the position and processing procedure of the header FCS 14 when there is a lot of traffic in the passage setting. This example is applied, for example, when there are many SDH signal transfers in packet transport.

  In this case, since there is a lot of traffic that does not need to inspect the payload 12, it is efficient to arrange the header FCS 14 at the end of the header 11. The processing procedure first reads up to the header FCS 14 and performs an FCS check (S1). If an error is detected, the frame is discarded regardless of the setting of the pass / discard control tag 15 (S2). On the other hand, if there is no error, the pass / discard control tag 15 is confirmed (S3), and if the setting is “pass”, the existing FCS 13 is passed without being checked (S4). On the other hand, if the setting is “discard”, the existing FCS 13 is read and the FCS check is performed (S5). If an error is detected, it is discarded (S2), and if there is no error, it is passed (S4). Thus, if the passing / discarding control tag 15 is set to “passing”, the transfer can be performed without checking the payload 12, which is efficient.

  FIG. 3 shows the position and processing procedure of the header FCS 14 when there is a lot of discard setting traffic. This example is applied, for example, when there are many IP signal transfers in packet transport.

  In this case, since it is desired that both the header 11 and the payload 12 have no error, it is better to check from the existing FCS 13. Therefore, it is efficient to arrange the header FCS 14 immediately after the existing FCS 13, that is, at the end of the frame. In the processing procedure, the existing FCS 13 is first read and FCS check is performed (S11), and if there is no error, the frame is passed (S12). On the other hand, if an error is detected, the FCS 14 for header is read and FCS check is performed (S13), and if an error is detected, the frame is discarded (S14). On the other hand, if there is no error in the header FCS 14 check, there is an error only in the payload 12, so the pass / discard control tag 15 is confirmed (S15), and if the setting is “pass”, it is passed (S12). If the setting is “Discard”, it is discarded (S14). As described above, if there is no error in the check of the existing FCS 13, it is not necessary to check the header FCS 14, so it is more efficient to arrange the header FCS 14 at the end of the frame.

FIG. 4 shows a configuration example of the frame transfer system of the present invention.
In the figure, a client network 21 on the transmission side and the server network 22 are connected via an edge switch 30 on the transmission side, and a server network 22 and a client network 23 on the reception side are connected via an edge switch 40 on the reception side. The edge switches 30 and 40 have a symmetric configuration, and include client-side SDH interfaces 31 and 41, client-side Ethernet interfaces 32 and 42, packet switches 33 and 43, and server-side Ethernet interfaces 34 and 44. The server-side Ethernet interfaces 34 and 44 include MAC layer processing circuits 35 and 45 and physical layer processing circuits 36 and 46. In the server network 22, a relay switch 50 including a reception port 51, an ingress processing circuit 52, a switch fabric circuit 53, an egress processing circuit 54, and a transmission port 55 is arranged. Hereinafter, the processing related to the present invention will be described.

  The received signal that has arrived at the client-side SDH interface 31 and the client-side Ethernet interface 32 of the sending-side edge switch 30 is divided by the packet switch 33 according to the payload size set by the server-side Ethernet interface 34, and route control is performed. A label or tag is assigned, and a pass / discard control tag is further given. The MAC layer processing circuit 35 assigns a MAC transmission source / destination address, calculates and assigns a header FCS together with the existing FCS for the entire frame, and forms a frame. The header FCS is generated by a CRC (Cyclic Redundancy Check) method or a parity method. In the case of the CRC method, 4 bytes or less is sufficient. In the case of the parity method, a check is made by assigning a parity bit of 1 bit to each of the destination address, source address, and other fields in the header, and adding a total of 3 parity bits to the header. Is possible.

  The ingress processing circuit 52 of the relay switch 50 checks the existing FCS and the header FCS according to the processing procedure of FIG. 2 or FIG. 3, and further determines whether to pass or discard by the pass / discard control tag.

  Similarly, the MAC layer processing circuit 45 of the receiving edge switch 40 checks the existing FCS and the header FCS according to the processing procedure of FIG. 2 or FIG. 3, and further processes whether to pass or discard by the pass / discard control tag. And discard frames other than those addressed to the local station. The packet switch 43 adapts the server side frame in accordance with the client side SDH interface 41 and the client side Ethernet interface 42.

  Although the CRC method and the parity method are shown as the header FCS, when CRC is used, especially when CRC16 is used, 1-bit error correction and arbitrary 2-bit error detection can be performed. Therefore, the header may be corrected when there is an error of 1 bit in the header, and the frame may be discarded when an error of 2 bits or more is detected.

The figure which shows the example of a frame structure in this invention. The figure which shows the position and processing procedure of FCS14 for headers when there is much traffic of passage setting. The figure which shows the position and processing procedure of FCS14 for headers when there is much traffic of discard setting The figure which shows the structural example of the frame transfer system of this invention.

Explanation of symbols

11 Header 12 Payload 13 Existing FCS
14 FCS for header
DESCRIPTION OF SYMBOLS 15 Passing / discarding control tag 21 Transmission side client network 22 Server network 23 Reception side client network 30 Transmission side edge switch 31, 41 Client side SDH interface 32, 42 Client side Ethernet interface 33, 43 Packet switch 34, 44 Server Ethernet interface 35, 45 MAC layer processing circuit 36, 46 Physical layer processing circuit 40 Edge switch on reception side 50 Relay switch 51 Reception port 52 Ingress processing circuit 53 Switch fabric circuit 54 Egress processing circuit 55 Transmission port

Claims (8)

In a frame transfer method in which an error of a frame including a header including an address and a payload is detected by a first FCS (Frame Check Sequence), the frame is discarded when an error is detected, and the frame when no error is detected is passed. ,
On the transmission side of the frame, generate the first FCS from the entire frame, and generate the second FCS from the header,
On the receiving side of the frame, the first FCS and the second FCS are checked, and there is an error in the header from each error detection status, or there is no error in the header and only the payload has an error. A frame transfer method characterized by identifying whether or not there is. The frame transfer method according to claim 1,
The frame receiving side discards the frame if there is an error in the header from the detection status of each error by checking the first FCS and the second FCS on the receiving side of the frame, and the payload without error in the header. If there is only an error, the frame transmission method is selected to pass or discard the frame. The frame transfer method according to claim 2, wherein
On the transmission side of the frame, set a passing / discarding control tag used for selection of passage or discarding of the frame in the header,
The frame transfer method characterized in that, on the frame receiving side, when there is no error in the header and there is an error only in the payload, processing according to the setting of the passing / discarding control tag is performed. The frame transfer method according to claim 3, wherein
Placing the second FCS at the end of the header;
The receiving side of the frame gives priority to checking the second FCS, discards the frame if an error is detected in the second FCS, and passes if the error is not detected in the second FCS. Check the discard control tag, and if it is set to pass, pass the frame without checking the first FCS, and if it is set to discard, perform processing according to the error detection status by the first FCS. And a frame transfer method. The frame transfer method according to claim 3, wherein
Placing the second FCS at the end of the previous frame;
The frame receiving side gives priority to checking the first FCS, and if an error is detected in the first FCS, the second FCS is checked, and an error is detected in the second FCS. The frame is discarded regardless of the setting of the passing / discarding control tag, and if no error is detected in the second FCS, processing according to the setting of the passing / discarding control tag is performed. Transfer method. In a frame transfer system that detects an error of a frame consisting of a header including an address and a payload by a first FCS (Frame Check Sequence), discards the frame when an error is detected, and passes the frame when no error is detected ,
The frame transmission side includes FCS generation means for generating the first FCS from the entire frame and generating the second FCS from the header,
The receiving side of the frame checks the first FCS and the second FCS, and there is an error in the header from the detection status of each error, or there is no error in the header and only the payload has an error. A frame transfer system comprising an error position specifying means for specifying whether or not there is an error. The frame transfer system according to claim 6, wherein
The receiving side of the frame discards the frame if there is an error in the header according to the error position specified by the error position specifying means, and if there is no error in the header and there is an error only in the payload A frame transfer system comprising frame transfer control means for selecting whether to pass or discard the frame. The frame transfer system according to claim 7, wherein
The transmission side of the frame includes a passing / discarding control tag setting means for setting a passing / discarding control tag used for selecting whether to pass or discard the frame in the header,
The frame transfer control means on the receiving side of the frame is configured to perform processing according to the setting of the pass / discard control tag when there is no error in the header and only the payload has an error. Frame transfer system.

Images