CN115955440A - Anti-lost packet method and related device - Google Patents

Abstract

The invention provides a packet loss prevention method and a related device, and relates to the technical field of data networks. The packet loss prevention method is applied to a switching chip of a network card, and firstly, a flow control frame carrying a buffer congestion level of a port of a net piece is generated; and then, sending the flow control frame to an exchange chip of the line card so that the exchange chip of the line card adjusts the rate of sending the data message to the port of the net sheet according to the buffer congestion level of the port of the net sheet, thereby avoiding the occurrence of packet loss of the port of the network card due to congestion.

Description

Anti-packet loss method and related device

technical field

The present invention relates to the technical field of data network, in particular, to an anti-packet loss method and a related device.

Background technique

The network switch rack is composed of multiple line cards and network chips, and each line card and network chip has a switching chip for exchanging data packets. The data packets coming in from the panel port of any line card can be forwarded through this line card, that is, go out from any port of this line card, and the data packets can also be forwarded through line card->network->line card.

Since the transmission rate of the panel port of the line card and the network chip is fixed, when the receiving bandwidth of any line card or network chip is greater than the sending bandwidth, buffer (cache) overflow will occur, and then packet loss will occur.

The current anti-packet loss technology is mainly proposed for the sudden congestion of the line card. When the transmission rate of the panel port of the line card and the network chip is the same, the sudden congestion will occur on the network chip, resulting in the buffer overflow of the network chip. , packet loss occurs.

Contents of the invention

In order to overcome the disadvantages of the prior art, an embodiment of the present invention provides a packet loss prevention method and a related device.

The technical scheme of the embodiment of the present invention can be realized like this:

In the first aspect, the embodiment of the present invention provides an anti-packet loss method, which is applied to the switch chip of the mesh in the network switch rack, and the network switch rack also includes a line card, and the switch chip of the line card is connected to the switch chip of the network switch rack. The exchange chip communication connection of above-mentioned mesh chip, described method comprises:

Generate a flow control frame carrying the buffer congestion level of the port of the mesh;

Send the flow control frame to the switch chip of the line card, so that the switch chip of the line card adjusts the number of data packets sent to the port of the network chip according to the buffer congestion level of the port of the network chip rate to avoid packet loss at the port of the network card due to congestion.

Optionally, before generating the flow control frame carrying the buffer congestion level of the port of the mesh, the method further includes the step of obtaining the buffer congestion level of the port of the mesh, which includes:

Obtaining a packet count value for a port of the network slice, wherein a change in the packet count value indicates that a port of the network slice is transferred into or out of a data packet;

Determine the buffer congestion level of the port of the network slice according to the packet count value.

Optionally, the step of determining the buffer congestion level of the port of the network slice according to the message count value includes:

If the packet count value is less than a first preset value and the port of the mesh is in a mildly congested state, then determine that the buffer congestion level of the port of the mesh is the first preset level;

If the packet count value is less than a second preset value and the port of the mesh is in a moderately congested state, then determine that the buffer congestion level of the port of the mesh is a second preset level;

If the packet count value is less than a third preset value and the port of the network slice is in a heavily congested state, then determine that the buffer congestion level of the port of the network slice is a third preset level.

Optionally, the step of determining the buffer congestion level of the port of the network slice according to the packet count value further includes:

If the packet count value is greater than a second preset value and the port of the mesh is in an uncongested state, then determine that the buffer congestion level of the port of the mesh is a second preset level;

If the packet count value is greater than a third preset value and the port of the mesh is in a mildly congested state, then determine that the buffer congestion level of the port of the mesh is a third preset level;

If the packet count value is greater than a fourth preset value and the port of the mesh is in a moderately congested state, determine that the buffer congestion level of the port of the mesh is a fourth preset level.

In the second aspect, the embodiment of the present invention provides an anti-packet loss method, which is applied to the switching chip of the line card in the network switch rack. The network switch rack also includes a mesh, and the switching chip of the mesh is connected to the The switch chip communication connection of the line card, the method includes:

receiving the flow control frame sent by the switch chip of the mesh, the flow control frame carrying the buffer congestion level of the port of the mesh;

According to the buffer congestion level of the port of the network chip, the rate of sending data packets to the port of the network chip is adjusted, so as to avoid packet loss at the port of the network card due to congestion.

Optionally, the step of adjusting the rate of sending data packets to the port of the mesh according to the buffer congestion level of the port of the mesh includes:

If the buffer congestion level of the port of the mesh is the first preset level, the rate of sending data packets to the port of the mesh is increased to a preset maximum rate;

If the buffer congestion level of the port of the mesh is the second preset level, the rate of sending data packets to the port of the mesh is reduced according to the first preset reduction rate;

If the buffer congestion level of the port of the network slice is the third preset level, the rate of sending data packets to the port of the network slice is reduced according to the second preset reduction rate, wherein the second preset The drop rate is greater than the first preset drop rate;

If the buffer congestion level of the port of the network slice is the fourth preset level, the rate of sending data packets to the port of the network slice is reduced to a preset minimum rate.

Optionally, the method also includes:

After reducing the rate of sending data packets to the port of the network chip, the counter for controlling the line card starts counting;

If it is detected that the value of the counter is greater than the preset threshold, the rate of sending data packets to the port of the network slice is gradually increased to the preset maximum rate within a preset time range.

In a third aspect, an embodiment of the present invention provides an anti-packet loss device, which is applied to a switch chip of a mesh in a network switch rack, and the network switch rack further includes a line card, and the switch chip of the line card is connected to the switch chip of the network switch rack. The exchange chip communication connection of above-mentioned network chip, described device comprises:

A generating module, configured to generate a flow control frame carrying the buffer congestion level of the port of the mesh;

A sending module, configured to send the flow control frame to the switch chip of the line card, so that the switch chip of the line card adjusts the port to the network chip according to the buffer congestion level of the port of the network chip The rate at which data packets are sent to avoid packet loss at the port of the network card due to congestion.

In a fourth aspect, an embodiment of the present invention provides an anti-packet loss device, which is applied to a switching chip of a line card in a network switch rack. The network switch rack further includes a mesh, and the switching chip of the mesh is connected to the The switch chip communication connection of the line card, the device includes:

A receiving module, configured to receive a flow control frame sent by the switch chip of the mesh, the flow control frame carrying the buffer congestion level of the port of the mesh;

An adjustment module, configured to adjust the rate of sending data packets to the port of the mesh according to the buffer congestion level of the port of the mesh, so as to avoid packet loss at the port of the network card due to congestion.

In the fifth aspect, an embodiment of the present invention provides a switch chip, the switch chip has a logic device, and the logic device is burned with a computer program, and when the computer program is executed, the anti-packet loss as described in the first aspect is realized method.

In the sixth aspect, an embodiment of the present invention provides a switch chip, the switch chip has a logic device, and the logic device is burned with a computer program, and when the computer program is executed, the anti-packet loss as described in the second aspect is realized method.

In the seventh aspect, the embodiment of the present invention provides a network switch rack, the network switch rack includes a plurality of meshes and a plurality of line cards, each of the meshes has a switching chip as described in the fifth aspect , each of the line cards has the switch chip as described in the sixth aspect.

In a third aspect, an embodiment of the present invention provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed, implements the anti-packet loss method as described in the first aspect or the method as described in the second aspect Anti-packet loss method.

Compared with the prior art, the embodiment of the present invention provides an anti-packet loss method applied to a switch chip of a mesh in a network switch rack. First, a flow control frame carrying a buffer congestion level of a port of the mesh is generated; Then, the flow control frame is sent to the switch chip of the line card, so that the switch chip of the line card adjusts the rate of sending data packets to the port of the network chip according to the buffer congestion level of the port of the network chip. Because the embodiment of the present invention sends the flow control frame carrying the buffer congestion level of the port of the network chip to the switch chip of the line card, the switch chip of the line card is adjusted to the buffer congestion level of the port of the network chip according to the buffer congestion level of the port of the network chip. The rate at which the port sends data packets, so as to avoid packet loss due to congestion on the port of the network card.

Description of drawings

In order to illustrate the technical solutions of the embodiments of the present invention more clearly, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention, and thus It should be regarded as a limitation on the scope, and those skilled in the art can also obtain other related drawings based on these drawings without creative work.

FIG. 1 is a schematic structural diagram of a network switch rack provided by an embodiment of the present invention;

FIG. 2 is an example diagram 1 of a data packet forwarding process provided by an embodiment of the present invention;

Fig. 3 is an example Fig. 2 of a data packet forwarding process provided by an embodiment of the present invention;

Fig. 4 is an example Fig. 3 of a data packet forwarding process provided by an embodiment of the present invention;

5 is a schematic flow diagram of a method for preventing packet loss applied to a switch chip of a mesh provided by an embodiment of the present invention;

FIG. 6 is a schematic flow diagram of a method for obtaining a buffer congestion level provided by an embodiment of the present invention;

FIG. 7 is a block diagram of functional units of a port resource management device deployed in a switch chip of a network card provided by an embodiment of the present invention;

FIG. 8 is a schematic flowchart of a packet loss prevention method applied to a switch chip of a line card provided by an embodiment of the present invention;

9 is a functional unit block diagram of an anti-packet loss device applied to a switch chip of a mesh provided by an embodiment of the present invention;

FIG. 10 is a block diagram of functional units of an anti-packet loss device applied to a switch chip of a line slice provided by an embodiment of the present invention.

Icon: 100-anti-packet loss device applied to the switch chip of the mesh; 101-generation module; 102-transmission module; 103-acquisition module; 200-anti-packet loss device applied to the switch chip of the line chip; 201-reception module; 202 - adjustment module.

Detailed ways

In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments It is a part of embodiments of the present invention, but not all embodiments. The components of the embodiments of the invention generally described and illustrated in the figures herein may be arranged and designed in a variety of different configurations.

Accordingly, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

It should be noted that like numerals and letters denote similar items in the following figures, therefore, once an item is defined in one figure, it does not require further definition and explanation in subsequent figures.

In addition, terms such as "first" and "second" are used only for distinguishing descriptions, and should not be understood as indicating or implying relative importance.

It should be noted that, in the case of no conflict, the features in the embodiments of the present invention may be combined with each other.

As shown in FIG. 1 , the network switch rack is composed of multiple line cards and network slices, and each line card and network slice has a switching chip for exchanging data packets.

The incoming data packets from the panel port of any line card can be forwarded through this line card, that is, go out from any panel port of this line card, and can also be forwarded in the way of line card->network->line card. It should be noted that packets cannot be lost during the forwarding of data packets.

Usually, the transmission rate of each panel port of a line card is fixed. When the ports of other line cards send data packets to a certain panel port of the same line card at the same time, because the receiving bandwidth of the line card is higher than the sending bandwidth, the line If the card does not have time to send the data packet, the buffer of the line card will overflow, resulting in packet loss.

For example, as shown in FIG. 2 , assuming that the transmission rate of the panel port of each line card is 100 Gbps, the panel port P0 of line card 2 and the panel port P0 of line card 3 simultaneously send data to the panel port P1 of line card 1 The packets are sent at full speed of 100Gbps, that is to say, line card 1 will receive 200Gbps data packets through network chip 1 and send them to P1. Since the port rate of P1 is only 100Gbps, it will inevitably cause P1 port buffer overflow. Then packet loss occurs.

The traditional solution is to distribute tokens on the line card to solve the problem of port buffer overflow. Since the tokens take a long time to respond and process, the port buffer needs to be deep enough.

Exemplarily, as shown in FIG. 3 , it is assumed that the buffer of panel port P1 of line card 1 has 1000 units (each unit can store up to 256 bytes), which are distributed to line card 2, line card 3.. line card m each k, where k*m<1000, if the data packets sent by line cards 2 and 3 have already occupied k buffers, then no data packets will be sent to P1 of line card 1 until the data packets of line card 1 P1 then sends new tokens to line card 2 and line card 3. That is to say, even if the combined rate of line card 2 and line card 3 is sent to P1 of line card 1 at a rate of 200Gbps, but because the sent message occupies a limited buffer, even if P1 has no time to transfer the data message, it will not be lost. The number of buffers exceeding P1 will result in packet loss.

However, traditional solutions only address burst issues from line card to line card. As shown in Figure 4, if the transmission rate of the panel port of network chip 1 and line card 1 is 100Gbps, then burst congestion will occur on network chip 1, and the data packet transfer rate (100Gbps) of the port of the network chip is lower than that of the network chip. Incoming rate (200Gbps), resulting in buffer overflow and packet loss.

In order to solve the packet loss in the forwarding process of the data message due to the sudden congestion of the network chip, the embodiment of the present invention provides an anti-packet loss method applied to the switching chip of the network chip, which will be described in detail below.

Please refer to FIG. 5 , the packet loss prevention method includes steps S101-S102.

S101. Generate a flow control frame carrying the buffer congestion level of the port of the network slice.

Among them, the flow control frame includes but not limited to chip id, port id, Priority and Congestion Level.

When the sending buffer is congested, the switching chip of the network chip fills its own chip id into the flow control frame. After the switching chip of the line card receives the flow control frame, it can know which network chip is congested according to the chip id.

There can be one or more ports where the network chip is congested. In the embodiment of the present invention, the port id of the port can be represented in the form of a bitmap, for example, bit0 represents port 0, bit1 represents port1, and so on. Make different settings according to your needs. When sending congestion, the switch chip of the network chip writes the port id of the congested port into the flow control frame. After the line card chip receives the flow control frame, it can know which port of which network chip is congested according to the port id.

Priority is used to express which priority or sub-channel is congested in the congested port, and can be expanded or canceled according to requirements.

Congestion Level, that is, the buffer congestion level, in the embodiment of the present invention, multiple buffer congestion levels can be set to make the deceleration and speed up relatively smooth, so as not to waste port bandwidth during transition.

S102, sending the flow control frame to the switch chip of the line card, so that the switch chip of the line card adjusts the rate of sending data packets to the port of the network chip according to the buffer congestion level of the port of the network chip, so as to avoid the port of the network card due to congestion There is packet loss.

Among them, the buffer congestion level represents the buffer congestion degree of the port.

For example, assume that there are four levels of buffer congestion, namely level 0, level 1, level 2, and level 3. Level 0 indicates that the port buffer is not congested; level 1 indicates that the port buffer is slightly congested; level 2 indicates that the port The buffer is moderately congested; level 3 means that the port buffer is severely congested.

In the embodiment of the present invention, before step S101 is executed, the buffer congestion level of the port of the network slice needs to be obtained first, as shown in FIG. 6 , the implementation process may include steps S201-S202.

S201. Obtain a packet count value for a port of a network slice.

Wherein, the change of the message count value indicates that the port of the network chip transfers in or out the data message.

The switching chip of the network chip is equipped with a port resource management device. As shown in FIG. 7 , the port resource management device includes a message transfer-in module, a message transfer-out module and a port resource counting module.

When the data message is planned to be forwarded from a certain port, the message transfer module completes the buffer operation of the data message, and initiates a counting request to the port resource counting module, the content is to increase the number of messages or count the length of the message .

When the data message is processed and transferred out to the port, the message forwarding module takes the data message out of the cache and sends it, and initiates a counting request to the port resource counting module, the content of which is to reduce the number of messages or Text length count.

In the embodiment of the present invention, if a data message is planned to be forwarded from a certain port, but because the port is temporarily unable to process it, the data message will be kept in the cache, that is, when a burst traffic occurs , if the port does not have enough capacity to process, then the data packets of these burst traffic will be temporarily stored in the cache, and the counter in the port resource counting module will maintain a relatively high value.

When the burst traffic ends and disappears, the previously cached packets are still taken out from the cache and forwarded. At this time, the counter in the port resource counting module gradually decreases. When the counter decreases to a relatively small value, it indicates that the burst The effect of flow disappears.

Understandably, the packet count value refers to the value of the counter in the port resource counting module.

S202. Determine the buffer congestion level of the port of the network slice according to the message count value.

In the embodiment of the present invention, multiple thresholds can be set for the message count value, and the buffer congestion level of the port of the network slice can be determined according to the comparison result of the message count value and the set threshold value. It can be understood that the size of the threshold value and The number can be set according to actual needs, and correspondingly, the number of buffer congestion levels can be set according to actual needs.

In a possible implementation manner, the buffer congestion level may include a first preset level, a second preset level, a third preset level, and a fourth preset level, and the first preset level, the second preset level , the third preset level and the fourth preset level indicate that the buffer congestion degree of the port is gradually increasing.

Optionally, there are the following six situations in the implementation process of step S202:

Case 1, if the message count value is less than the first preset value and the port of the mesh is in a mildly congested state, then determine that the buffer congestion level of the port of the mesh is the first preset level;

Case 2, if the message count value is less than the second preset value and the port of the mesh is in a moderately congested state, then determine that the buffer congestion level of the port of the mesh is the second preset level;

Case 3, if the message count value is less than the third preset value and the port of the mesh is in a heavily congested state, then determine that the buffer congestion level of the port of the mesh is the third preset level;

Case 4, if the packet count value is greater than the second preset value and the port of the mesh is not congested, then determine that the buffer congestion level of the port of the mesh is the second preset level;

Case 5, if the packet count value is greater than the third preset value and the port of the mesh is in a mildly congested state, then determine that the buffer congestion level of the port of the mesh is the third preset level;

Case 6, if the packet count value is greater than the fourth preset value and the port of the mesh chip is in a moderately congested state, then determine that the buffer congestion level of the port of the mesh chip is the fourth preset level.

Exemplarily, it is assumed that the first preset value, the second preset value, the third preset value and the fourth preset value are threshold0, threshold1, threshold2 and threshold3 respectively. Wherein, their sizes satisfy threshold0<threshold1<threshold2<threshold3.

The first preset level, the second preset level, the third preset level and the fourth preset level are represented by level 0, level 1, level 2 and level 3 respectively.

Use threshold0, threshold1, threshold2, and threshold3 to determine the buffer congestion level of the port, that is, determine the buffer congestion level of the port.

It can be understood that when the packet count value is less than threshold0 and the current port is in a mildly congested state, then the port will change from a lightly congested state to a non-congested state, and the corresponding buffer congestion level is level 0.

When the packet count value is less than threshold1 and the current port is in the moderately congested state, the port will change from the moderately congested state to the lightly congested state, and the corresponding buffer congestion level is level 1.

When the packet count value is less than threshold2 and the current port is in a heavily congested state, the port will change from a heavily congested state to a moderately congested state, and the corresponding buffer congestion level is level 2.

When the packet count value is greater than threshold1 and the current port is in the uncongested state, the port will change from the uncongested state to the lightly congested state, and the corresponding buffer congestion level is level 1.

When the packet count is greater than threshold2 and the current port is in a mildly congested state, the port will change from a lightly congested state to a moderately congested state, and the corresponding buffer congestion level is level 2.

When the packet count is greater than threshold3 and the current port is in a moderately congested state, the port will change from a moderately congested state to a severely congested state, and the corresponding buffer congestion level is level 3.

The embodiment of the present invention also provides an anti-packet loss method applied to a switch chip of a wire slice, which will be described in detail below.

Please refer to FIG. 8 , the packet loss prevention method includes steps S301-S302.

S301. Receive a flow control frame sent by a switch chip of a mesh.

Among them, the flow control frame carries the buffer congestion level of the port with the mesh slice.

S302. Adjust the rate of sending data packets to the port of the network chip according to the buffer congestion level of the port of the network chip, so as to avoid packet loss at the port of the network card due to congestion.

Optionally, there are the following four situations in the implementation process of step S302.

In case 1, if the buffer congestion level of the port of the network chip is the first preset level, the rate of sending data packets to the port of the network chip is increased to the preset maximum rate.

In case 2, if the buffer congestion level of the port of the network chip is the second preset level, the rate of sending data packets to the port of the network chip is reduced according to the first preset reduction rate.

In case 3, if the buffer congestion level of the port of the network chip is the third preset level, the rate of sending data packets to the port of the network chip is reduced according to the second preset reduction rate.

In case 4, if the buffer congestion level of the port of the network chip is the fourth preset level, the rate of sending data packets to the port of the network chip is reduced to a preset minimum rate.

Wherein, the preset maximum rate, the first preset decrease rate, the second preset decrease rate and the preset minimum rate can be set according to actual needs.

For example, assuming that the first preset level, the second preset level, the third preset level and the fourth preset level are represented by level 0, level 1, level 2 and level 3 respectively, and the preset maximum rate is 100Gbps, The preset minimum rate is 0Gbps, the first preset drop rate is 80%, and the second preset drop rate is 20%.

Level 0 means that the port buffer is not congested, and the rate at which the line card sends data packets to the port of the network chip can be increased to 100Gbps; level 1, indicates that the port buffer is slightly congested, and the rate at which the line card sends data packets to the port of the network chip A little reduction is required to reduce to 80% of the preset maximum rate, that is, 100Gbps*80%=80Gbps; level 2 means that the port buffer is moderately congested, and the rate at which the line card sends data packets to the port of the network chip needs to be greatly reduced. Up to 20% of the preset maximum rate, that is, 100Gbps*20%=20Gbps; level 3 means that the port buffer is seriously congested, and the line card needs to stop sending data packets to the port of the network chip immediately, that is, the line card sends data packets to the port of the network chip The rate of sending data packets is reduced to 0Gbps.

In order to alleviate the frequent sending of flow control frames by the switching chip of the network card to the switching chip of the line card, for the above-mentioned cases 2-4, the packet loss prevention method also includes the following procedure.

First, after reducing the rate of sending data packets to the port of the network chip, the counter of the control line card starts counting.

Then, if it is detected that the value of the counter is greater than the preset threshold, the rate of sending data packets to the port of the network slice is gradually increased to the preset maximum rate within the preset time range.

Wherein, the preset threshold can be set according to actual needs.

It is understandable that when the congestion level of the mesh port changes to a more serious level, the switching chip of the line card will send a datagram to the port of the mesh according to the buffer congestion level in the flow control frame after receiving the flow control frame. The rate of data packets decreases, and the counter of the line card starts counting. When the value of the counter reaches the preset threshold value, the rate of sending data packets to the port of the network chip is gradually increased until it reaches the preset maximum rate, so that it can be used in multiple When there is buffer congestion on a port, the switch chip of the network card may frequently send flow control frames to the switch chip of the line card.

In order to execute the corresponding steps in the above method embodiments and various possible implementations, an implementation of an anti-packet loss device 100 applied to a switch chip of a network chip and an anti-packet loss device 100 applied to a switch chip of a line card are given below. Packet discarding device 200.

Referring to FIG. 9 , the anti-packet loss device 100 applied to a switch chip of a mesh includes a generation module 101 , a sending module 102 and an acquisition module 103 .

The generating module 101 is configured to generate a flow control frame carrying the buffer congestion level of the port of the mesh slice.

The sending module 102 is used to send the flow control frame to the switch chip of the line card, so that the switch chip of the line card adjusts the rate of sending data packets to the port of the network chip according to the buffer congestion level of the port of the network chip, so as to avoid network card Packet loss occurs on the port due to congestion.

Optionally, the obtaining module 103 is used to obtain the packet count value for the port of the network slice, wherein the change of the packet count value represents that the port of the network slice is transferred into or out of the data packet; according to the packet count value, determine The buffer congestion level of the port on the mesh.

Optionally, the obtaining module 103 is specifically configured to determine that the buffer congestion level of the port of the mesh is the first preset level if the packet count value is less than the first preset value and the port of the mesh is in a mildly congested state; If the message count value is less than the second preset value and the port of the network slice is in a moderately congested state, then determine that the buffer congestion level of the port of the network slice is the second preset level; if the message count value is less than the third preset value, and the port of the network chip is in a heavily congested state, then it is determined that the buffer congestion level of the port of the network chip is the third preset level.

Optionally, the obtaining module 103 is also specifically configured to determine that the buffer congestion level of the port of the mesh is the second preset level if the packet count value is greater than the second preset value and the port of the mesh is in an uncongested state; If the message count value is greater than the third preset value, and the port of the network slice is in a mildly congested state, then determine that the buffer congestion level of the port of the network slice is the third preset level; if the message count value is greater than the fourth preset value, and the port of the mesh is in a moderately congested state, then determine that the buffer congestion level of the port of the mesh is the fourth preset level.

Referring to FIG. 10 , the anti-packet loss device 200 applied to a switch chip of a line card includes a receiving module 201 and an adjusting module 202 .

The receiving module 201 is configured to receive the flow control frame sent by the switch chip of the mesh, and the flow control frame carries the buffer congestion level of the port of the mesh.

The adjustment module 202 is configured to adjust the rate of sending data packets to the port of the network chip according to the buffer congestion level of the port of the network chip, so as to avoid packet loss at the port of the network card due to congestion.

Optionally, the adjustment module 202 is specifically configured to increase the rate of sending data packets to the port of the mesh to a preset maximum rate if the buffer congestion level of the port of the mesh is the first preset level; If the buffer congestion level of the port of the mesh slice is the second preset level, the rate of sending data packets to the port of the mesh slice will be reduced according to the first preset reduction rate; if the buffer congestion level of the port of the mesh slice is the third preset level level, then the rate of sending data packets to the port of the network chip is reduced according to the second preset reduction rate, wherein the second preset reduction rate is greater than the first preset reduction rate; if the buffer congestion level of the port of the network chip is the fourth If the preset level is selected, the rate of sending data packets to the port of the network chip will be reduced to the preset minimum rate.

Optionally, the adjustment module 202 is also specifically configured to control the counter of the line card to start counting after reducing the rate of sending data packets to the ports of the network slice; The rate at which the ports of the slices send data packets is gradually increased to the preset maximum rate within a preset time range.

Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific details of the above-described packet loss prevention device 100 applied to the switch chip of the network chip and the packet loss prevention device 200 applied to the switch chip of the line card For the working process, reference may be made to the corresponding process in the foregoing method embodiments, which will not be repeated here.

The embodiment of the present invention also provides a switching chip installed on a mesh chip, the switching chip has a logic device, and the logic device is burned with a computer program, and when the computer program is executed, the relevant operations in the method provided by the above-mentioned method embodiment are implemented. .

The embodiment of the present invention also provides a switching chip installed on the line card, the switching chip has a logic device, and the logic device is burned with a computer program, and when the computer program is executed, the relevant operations in the method provided by the above-mentioned method embodiment are realized. .

An embodiment of the present invention also provides a computer-readable storage medium containing a computer program. When executed, the computer program can be used to perform related operations in the method provided by the above-mentioned method embodiment.

To sum up, the anti-packet loss method and related devices provided by the embodiments of the present invention first generate a flow control frame carrying the buffer congestion level of the port of the mesh; then, send the flow control frame to the switch chip of the line card to The switching chip of the line card adjusts the rate of sending data packets to the port of the network chip according to the buffer congestion level of the port of the network chip. Because the embodiment of the present invention sends the flow control frame carrying the buffer congestion level of the port of the network chip to the switch chip of the line card, the switch chip of the line card is adjusted to the buffer congestion level of the port of the network chip according to the buffer congestion level of the port of the network chip. The rate at which the port sends data packets, so as to avoid packet loss due to congestion on the port of the network card.

The above is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Anyone skilled in the art can easily think of changes or substitutions within the technical scope disclosed in the present invention. All should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be determined by the protection scope of the claims.

Claims (13)

1. A method for preventing packet loss, characterized in that, it is applied to the switching chip of the mesh in the network switch frame, and the network switch frame also includes a line card, and the switching chip of the line card is connected to the mesh The communication connection of the switch chip, the method includes: Generate a flow control frame carrying the buffer congestion level of the port of the mesh; Send the flow control frame to the switch chip of the line card, so that the switch chip of the line card adjusts the number of data packets sent to the port of the network chip according to the buffer congestion level of the port of the network chip rate to avoid packet loss at the port of the network card due to congestion. 2. The method according to claim 1, wherein, before the generation of the flow control frame carrying the buffer congestion level of the port of the mesh, the method further comprises obtaining the port information of the mesh The steps of buffer congestion level include: Obtaining a packet count value for a port of the network slice, wherein a change in the packet count value indicates that a port of the network slice is transferred into or out of a data packet; Determine the buffer congestion level of the port of the network slice according to the packet count value. 3. The method according to claim 2, wherein, according to the message count value, the step of determining the buffer congestion level of the port of the mesh slice comprises: If the packet count value is less than a first preset value and the port of the mesh is in a mildly congested state, then determine that the buffer congestion level of the port of the mesh is the first preset level; If the packet count value is less than a second preset value and the port of the mesh is in a moderately congested state, then determine that the buffer congestion level of the port of the mesh is a second preset level; If the packet count value is less than a third preset value and the port of the network slice is in a heavily congested state, then determine that the buffer congestion level of the port of the network slice is a third preset level. 4. the method for claim 3, is characterized in that, described according to described message counter value, the step of determining the buffer congestion level of the port of described network slice also comprises: If the packet count value is greater than a second preset value and the port of the mesh is in an uncongested state, then determine that the buffer congestion level of the port of the mesh is a second preset level; If the packet count value is greater than a third preset value and the port of the mesh is in a mildly congested state, then determine that the buffer congestion level of the port of the mesh is a third preset level; If the packet count value is greater than a fourth preset value and the port of the mesh is in a moderately congested state, determine that the buffer congestion level of the port of the mesh is a fourth preset level. 5. A method for preventing packet loss, characterized in that, it is applied to the exchange chip of the line card in the network switch rack, and the network switch rack also includes a mesh, and the exchange chip of the mesh and the line card The communication connection of the switch chip, the method includes: receiving the flow control frame sent by the switch chip of the mesh, the flow control frame carrying the buffer congestion level of the port of the mesh; According to the buffer congestion level of the port of the network chip, the rate of sending data packets to the port of the network chip is adjusted, so as to avoid packet loss at the port of the network card due to congestion. 6. The method according to claim 5, wherein, according to the buffer congestion level of the port of the mesh, the step of adjusting the rate of sending a data message to the port of the mesh comprises: If the buffer congestion level of the port of the mesh is the first preset level, the rate of sending data packets to the port of the mesh is increased to a preset maximum rate; If the buffer congestion level of the port of the mesh is the second preset level, the rate of sending data packets to the port of the mesh is reduced according to the first preset reduction rate; If the buffer congestion level of the port of the network slice is the third preset level, the rate of sending data packets to the port of the network slice is reduced according to the second preset reduction rate, wherein the second preset The drop rate is greater than the first preset drop rate; If the buffer congestion level of the port of the network slice is the fourth preset level, the rate of sending data packets to the port of the network slice is reduced to a preset minimum rate. 7. The method of claim 6, further comprising: After reducing the rate of sending data packets to the port of the network chip, the counter for controlling the line card starts counting; If it is detected that the value of the counter is greater than the preset threshold, the rate of sending data packets to the port of the network slice is gradually increased to the preset maximum rate within a preset time range. 8. An anti-packet loss device, characterized in that, it is applied to the switching chip of the mesh in the network switch frame, and the network switch frame also includes a line card, and the switching chip of the line card is connected to the mesh The communication connection of the switch chip, the device includes: A generating module, configured to generate a flow control frame carrying the buffer congestion level of the port of the mesh; A sending module, configured to send the flow control frame to the switch chip of the line card, so that the switch chip of the line card adjusts the port to the network chip according to the buffer congestion level of the port of the network chip The rate at which data packets are sent to avoid packet loss at the port of the network card due to congestion. 9. An anti-packet loss device, characterized in that it is applied to the switching chip of the line card in the network switch rack, and the network switch rack also includes a mesh, and the switching chip of the mesh and the line card The communication connection of the switch chip, the device includes: A receiving module, configured to receive a flow control frame sent by the switch chip of the mesh, the flow control frame carrying the buffer congestion level of the port of the mesh; An adjustment module, configured to adjust the rate of sending data packets to the port of the mesh according to the buffer congestion level of the port of the mesh, so as to avoid packet loss at the port of the network card due to congestion. 10. A switch chip, characterized in that the switch chip has a logic device, and the logic device is programmed with a computer program, and when the computer program is executed, the anti-virus protection according to any one of claims 1 to 4 is realized. Packet drop method. 11. A switch chip, characterized in that the switch chip has a logic device, and the logic device is programmed with a computer program, and when the computer program is executed, the anti-virus protection according to any one of claims 5 to 7 is realized. Packet drop method. 12. A network switch frame, characterized in that, the network switch frame includes a plurality of mesh sheets and a plurality of line cards, each of the mesh sheets has a switching chip as claimed in claim 10, each The switching chips as claimed in claim 11 all exist on the line cards. 13. A computer-readable storage medium, characterized in that it stores a computer program, and when the computer program is executed, it implements the anti-packet loss method according to any one of claims 1 to 4 or the method according to any one of claims 5 to 5. 7. The anti-packet loss method described in any one.

Images