CN1300985C - Method for processing extra-long message in two-layer virtual special-purpose network - Google Patents

Abstract

The invention discloses a method for solving the maximum number of bytes of a link transmission unit in a two-layer virtual private network, including adding an ICMP protocol on a PE device, the ICMP protocol can issue an error report for the network, and when the message sent by the user When the length exceeds 1522 bytes, the ICMP protocol can automatically generate a destination unreachable error report and feed back the maximum number of bytes that can be sent to the user host, so that the user can resend a message with a suitable length according to the feedback information. Using this method can limit the size of the message sent by the user computer, prevent the message exceeding 1522 bytes from the source, and effectively ensure that the message will not be discarded.

Description

The Method of Handling Overlong Messages in Layer 2 Virtual Private Network

field of invention

The invention generally relates to network transmission technology, in particular to a method for solving the problem of limiting the number of bytes of the maximum transmission unit of a link in a two-layer virtual private network.

Background technique

In L2VPN (Layer 2 Virtual Private Network), when a user message passes through a tunnel, it must be encapsulated in a new link layer frame header and tunnel bearer protocol layer, for example, based on the MPLS (Multiprotocol Label Switching) protocol Tunnels must be equipped with two layers of MPLS label headers. If both the link on the user side and the link on the tunnel side are established on the Ethernet, the maximum frame length of the Ethernet user packet after adding the tunnel protocol header will far exceed the maximum frame length of the Ethernet protocol of 1522 bytes , but because it is impossible for the Layer 2 protocol to fragment the packet, such super-long packets are generally sent to the edge device of the service provider (referring to the tunnel edge device in this article, hereinafter referred to as PE equipment) or service provider equipment (referred to as the tunnel intermediate equipment in this document, hereinafter referred to as P equipment) to discard. This problem will be described in detail below in conjunction with the accompanying drawings.

Figure 1 is a schematic diagram of a typical L2VPN network. As shown in Figure 1, the local vlan (virtual local area network) 8 is connected to the remote vlan 12 through the L2VPN tunnel. In this way, the packets sent from vlan 8 to vlan 12 need to pass through PE1 to install the tunnel carrying protocol header, and then they can sent into the tunnel. Because during this process, PE1 generally does not know what protocol is encapsulated in the user's Layer 2 packet, so it is impossible for L2VPN to perform fragmentation processing. The L2VPN protocol stipulates that if the length of the user packet plus the length of the tunnel encapsulation header exceeds the value of the tunnel MTU (the number of bytes in the maximum transmission unit of the link), the PE device cannot fragment it, but simply discards it.

But there are big defects in this processing. In order to illustrate the problem more clearly, we assume that the tunnel is established on the MPLS protocol. FIG. 2 is a schematic diagram of packets encapsulated in an L2VPN tunnel based on the MPLS protocol. As shown in Figure 2, the original user message will generally increase by 22 or 26 (in the case of Vlan tag) bytes after the tunnel protocol header is added. In this way, the maximum length of the message entering the MPLS tunnel can reach 1522 bytes. +26=1548 bytes, which has greatly exceeded the MTU value of 1522 allowed by the Ethernet. Since the L2VPN protocol does not allow fragmentation processing, these ultra-long packets must be discarded according to the previous Ethernet protocol regulations.

The aforementioned flaw is a major problem faced by all L2VPN protocols. For this problem, some solutions have been proposed in the prior art. One of the solutions is: Regardless of the limitations of the Ethernet protocol, forcibly change the MTU value of all equipment outlets on the tunnel to more than 1548 bytes. At present, most Ethernet interface chips support sending and receiving Ethernet data exceeding 1522 bytes. Network packets, so many devices support this setting. CISCO also specially named this kind of packets, called Biggiant packets. However, the problem with this solution is that all P devices and PE devices along the route must support receiving and sending such ultra-long Ethernet packets, otherwise, no matter which intermediate node does not support these protocols, the frame will be sent unconditionally. Throw away or shard incorrectly. In the case of a complex network, there may be many P nodes in the middle of the tunnel, and they may be provided by different equipment vendors. Therefore, it cannot be guaranteed that all P devices involved in the L2VPN tunnel support this setting. There is also a solution, that is, to encapsulate the Ethernet text by using the jumbo frame (jumbo frame) encapsulation protocol. This protocol is an informal standard protocol, which can support the transmission of packets up to 9022 bytes on Ethernet. It was originally a protocol specially developed for Gbit Ethernet, but L2VPN can use this protocol to solve the problem. Own MTU problem. However, this solution also has problems similar to the above-mentioned first solution, that is, all P devices and PE devices along the route must support the jumbo frame encapsulation protocol, otherwise, no matter which intermediate node does not support these protocols , the frame will be unconditionally thrown away or fragmented incorrectly. In the case of a complex network, there may be many P nodes in the middle of the tunnel, and they may be provided by different equipment vendors. Therefore, it cannot be fully guaranteed that all P devices, PE devices and links involved in the L2VPN tunnel support this protocol.

In addition, as a reference to the present invention, there is a method for processing super-long messages in the related prior art, and the method adopts ICMP (Internet Control Message Protocol) to control super-long messages. FIG. 3 shows a schematic diagram of such a method for processing superlong packets. As shown in Figure 3, when the operating system of the computer transmits files on the network through the router, the sent message will use the function of prohibiting fragmentation, that is, the DF (fragmentation prohibited flag) flag in the IP header of the message is set, because the DF flag is an important flag in the known IP protocol, so it will not be described again. If the router receives such a packet and finds that the packet length is larger than the interface MTU, since the packet has been explicitly specified not to be fragmented, the IP protocol stipulates that the packet must be discarded, and the router will respond to The source host sends an ICMP message of type=3, code=4, and carries the MTU value of the interface in the message, so that after receiving the message, the host will actively adjust its own MTU value in the ICMP message The value carried, all messages sent by the host in the future will be smaller than this value. However, this method in the prior art only provides the application in the three-layer network transmission, but does not give the application in the two-layer network transmission technology such as L2VPN.

Contents of the invention

In order to solve the above-mentioned defects in the prior art, the purpose of the present invention is to provide a method that can fundamentally prevent the packet length in the Layer 2 virtual private network from being too long.

In order to achieve the above object, the present invention provides a method for processing super-long messages in a two-layer virtual private network, the method comprising the following steps: 1) binding a message control protocol on a PE device; 2) when the slave sends When the message sent by the end to the PE device is a message that is prohibited from being fragmented and its length exceeds a predetermined length, the PE device uses the message control protocol to automatically generate a feedback message and return it to the sender 3) The sending end adjusts the message length according to the allowable length of the next message in the feedback information and resends the message.

By using the method described in the present invention to bind a message control protocol on the PE equipment, and using this protocol to control the length of the message, the length of the message sent by the user computer can be limited, so that at the source Prevent messages exceeding 1522 bytes. In this way, regardless of whether the P device supports the above two protocols, it can ensure that the packet will not be discarded.

Description of drawings

Figure 1 is a schematic diagram of a typical L2VPN network;

Fig. 2 is a schematic diagram of packets encapsulated in an L2VPN tunnel based on the MPLS protocol;

Fig. 3 is a schematic diagram of the processing of a super-long message set by DF in the prior art;

FIG. 4 is a schematic diagram of a method for processing superlong messages in a Layer 2 virtual private network according to an embodiment of the present invention;

FIG. 5 is a schematic structural diagram of an ICMP message in an embodiment of the present invention.

Detailed ways

Embodiments of the present invention will be described in detail below with reference to FIGS. 4 and 5 .

In order to reduce the time-consuming fragmentation operation of the router and improve network efficiency, the current mainstream operating systems basically set the DF bit of the message when communicating on the network, and are ready to adjust their MTU value according to the situation at any time. The intention of the present invention is to utilize this feature, add an ICMP (Internet Control Message Protocol) on the PE equipment and actively adjust the MTU value of the host computer by using the protocol to respond to the ICMP message, so as to solve the problem of the L2VPN MTU value . Specific embodiments of the present invention will be described in detail below.

First, an ICMP protocol needs to be bundled on the PE device. The ICMP protocol is an existing three-layer protocol, not all in the L2VPN protocol. One of its most basic functions is to issue an error report for the network, such as automatically feeding back to the user an ICMP message whose destination is unreachable (Type=3, code=4). Currently, almost all routers or Layer 3 switches support this protocol. As for the specific process of bundling the ICMP protocol, since it is well known in the art, it will not be repeated here.

Then, as shown in Figure 4, when the original message sent by the PC enters the tunnel through the PE device, according to the L2VPN protocol, the PE device must add a tunnel protocol header to it. After installing the tunnel protocol header, the PE device will compare the length of the packet after adding the header with the MTU value of the tunnel (that is, the maximum length of Ethernet is 1522 bytes), and judge the DF of the packet. Whether to set. If the packet is an IP packet with DF set, and its length exceeds the MTU value of the tunnel (that is, 1522 bytes) after the tunnel protocol header is added, the PE device will discard the packet and report to the PC The computer responds with an ICMP message, and carries the MTU value of the outgoing interface in the ICMP message to feed back to the PC. For example, in the case that the tunnel is established on the MPLS protocol, the MTU value of the tunnel carried in the ICMP message may be 1500-26=1474. Note that in the above example, because the ICMP protocol stipulates that the maximum length of the layer-3 payload must be responded, it is necessary to remove the layer-2 frame header (18 bytes) + layer-2 frame CRC checksum (4 bytes) before removing the tunnel The length of the protocol header (26 bytes), so as to correctly indicate the MTU value of the tunnel. In addition, since the DF of the messages in the current mainstream network communication are all set, the above step of determining the DF set can also be omitted.

Figure 5 shows the structure of an ICMP message. As shown in FIG. 5 , the ICMP frame header of the ICMP message frame indicates the MTU value of the message when it is sent next time (1474 in this example). After receiving the above-mentioned ICMP message with the MTU value, the user PC only needs to automatically reduce the MTU value of its corresponding network interface according to the value (1474) fed back in the ICMP message. Since any host supporting the TCP/IP protocol must support the ICMP protocol at present, and the operation process of the protocol is also known, it will not be described again. In this way, all subsequent user packets will not exceed the normal Ethernet MTU value even after the tunnel protocol header is added.

The invention utilizes the error report feedback function of the ICMP protocol to enable the user host to actively reduce its own MTU value, and solve the defect that the MTU value does not match and cannot be fragmented in the L2VPN without updating the device, and can effectively ensure that the message is not will be discarded. This is a very meaningful addition to the L2VPN protocol.

It should be noted that although the above description of the present invention has been made with reference to specific examples thereof, it is not meant to limit the invention thereto. Those skilled in the art should understand that various modifications and changes can be made on the basis of the idea of the present invention. In a word, various modifications without departing from the spirit of the present invention are within the protection scope of the appended claims of the present invention.

Claims (7)

1. in two-layer virtual private network, handle the overlength method of message for one kind, may further comprise the steps:

1) on tunnel edge device PE, adds a message control protocol;

2) when the message that is sent to described PE equipment from transmitting terminal be that the message of the burst that is under an embargo and its length are when surpassing a predetermined length, the described message control protocol of described PE equipment utilization produces a feedback information automatically and it is returned to transmitting terminal, comprises the permission length of next message in this feedback information at least;

3) transmitting terminal is according to the permission length adjustment message length of the next message in the described feedback information New Development literary composition of delivering newspaper of laying equal stress on.

2. method according to claim 1 is characterized in that, described message control protocol is Internet Control Message Protocol ICMP.

3. method according to claim 1 is characterized in that, described predetermined length is the maximum frame length of Ethernet.

4. method according to claim 1 is characterized in that, described step 2) in also comprise and judge that whether described message length surpasses the step of described predetermined length.

5. method according to claim 1 is characterized in that, described step 2) in also comprise the step whether burst flag bit DF is set of forbidding in the IP head of judging described message.

6. method according to claim 1 is characterized in that, described step 2) in the permission length of next message be legal Ethernet message length, maximum is less than described predetermined length.

7. according to any described method in the claim 1 to 6, it is characterized in that, comprise further in the described step 3) that transmitting terminal is adjusted into message length the step of the permission length of the described message that is included in the described feedback information.

Images