WO2018113436A1

WO2018113436A1 - Method and device for network congestion processing

Info

Publication number: WO2018113436A1
Application number: PCT/CN2017/110063
Authority: WO
Inventors: 彭信木
Original assignee: 深圳创维数字技术有限公司
Priority date: 2016-12-19
Filing date: 2017-11-09
Publication date: 2018-06-28
Also published as: CN106506392B; CN106506392A

Abstract

A method and a device for network congestion processing. The method comprises: acquiring the current central processing unit (CPU) occupancy when a soft interrupt occurs, and determining, according to the current CPU occupancy, whether to perform packet discard processing; when it is determined to perform packet discard processing, determining the current packet loss rate according to the current CPU occupancy; determining, according to preset packet selection criteria, the number of packets that satisfy a packet discard condition; and selectively discarding packets according to the current packet loss rate and the number of packets that satisfy the packet discard condition.

Description

Network congestion processing method and device

Technical field

The present disclosure relates to computer network technologies, for example, to a network congestion processing method and apparatus.

Background technique

With the continuous development of network big data, the situation in which system servers are stuck due to network congestion often occurs.

When there are a large number of data streams in the router, the usage of the central processing unit (CPU) of the frequent network soft interruption may reach 100%. Because the level of the soft interrupt is higher than that of the normal application process, the router is locally localized. The service is not implemented.

Commonly used solutions: First, the Ethernet driver supports the NAPI (New Application Program Interface) scheduling method. This method cannot selectively drop packets when the network is congested, resulting in loss of key service data packets. At the same time, when the network data volume is moderate, frequent switching between soft interrupt and polling scheduling modes occurs, which affects system performance. Second, the application of Quality of Service (QOS) to limit the interface bandwidth. Because the TC flow control method itself consumes a large amount of CPU, and when the CPU is tight, the queue operation is inaccurate, and the control of the data stream is lost. When the total bandwidth cannot be suppressed, the CPU load is aggravated. Third, reduce the size of the buffer area received and sent by Ethernet. However, this method cannot implement dynamic adjustment of the packet loss rate, and cannot selectively packetize packets.

Summary of the invention

The embodiment of the invention provides a network congestion processing method and device, so as to implement dynamically adjustable selective packet loss and ensure normal operation of the system.

A network congestion processing method includes:

Obtaining the CPU usage of the current CPU of the soft interrupt, and determining whether to perform packet loss processing according to the current CPU usage;

When it is determined that the packet loss processing is performed, determining a current packet loss rate according to the current CPU usage rate;

Determining the number of data packets satisfying the packet loss condition according to a preset data packet selection criterion;

Selective packet loss according to the current packet loss rate and the number of data packets satisfying the packet loss condition.

Optionally, determining a current packet loss rate according to the current CPU usage, including:

Determining a current packet loss weight according to the current CPU usage;

Determining the current packet loss rate according to the current packet loss weight.

Optionally, determining a current packet loss weight according to the current CPU usage, including:

Determining whether the current CPU usage is greater than a first threshold;

If the current CPU usage is greater than the first threshold, determining that the current packet loss value is increased by one, and if the current CPU usage is less than or equal to the first threshold, determining whether the CPU usage rate is Less than the second threshold;

If the current CPU usage is less than the second threshold, further determining whether the current packet loss value is zero. If it is zero, determining that the current packet loss weight is unchanged, if not, then Determining that the current packet loss weight is decreased by one;

If the current CPU usage is greater than or equal to the second threshold, the new CPU usage is cyclically acquired after the preset time interval.

Optionally, determining the number of data packets that meet the packet loss condition according to the preset data packet selection criteria, including:

Obtaining packet information, where the packet information includes a protocol type of the data packet, a data packet length, a source port, and a destination port;

Determining whether the protocol type of the data packet is a preset protocol type, if it is not a preset protocol type, determining that the data packet does not satisfy a packet loss condition, and if it is a preset protocol type, performing a preset protocol type on the data packet Marking, and detecting whether the length of the data packet of the data packet is greater than a preset length;

If the data packet length of the data packet is less than or equal to the preset length, determining that the data packet does not meet the packet loss condition, and if the data packet length of the data packet is greater than the preset length, determining that the data packet is satisfied The number of packets for the packet condition is increased by one.

Optionally, before detecting whether the protocol type of the data packet is a preset protocol type, the method includes:

Detecting whether the data packet includes a preset protocol type flag, if yes, determining that the data packet does not satisfy the packet loss condition, and if not, detecting whether the protocol type of the data packet is a preset protocol Type of discussion.

Optionally, after determining the number of data packets that meet the packet loss condition, the method further includes:

Detecting whether the source port and/or the destination port of the data packet is a pre-release port. If the port is pre-released, it is determined that the data packet does not meet the packet loss condition. If the port is not pre-released, the data is determined. The packet satisfies the packet loss condition.

Optionally, performing selective packet loss according to the current packet loss rate and the number of data packets that meet the packet loss condition, including:

Determining the number of lost packets according to the current packet loss rate and the number of data packets satisfying the packet loss condition;

The data packet with the number of lost packets is randomly selected in the data packet satisfying the packet loss condition for packet loss processing.

A network congestion processing device includes:

The packet loss processing judging module is configured to obtain the CPU usage of the current CPU of the soft interrupt, and determine whether to perform packet loss processing according to the current CPU usage rate;

The packet loss rate determining module is configured to determine a current packet loss rate according to the current CPU usage rate when determining that the packet loss processing is performed;

a data packet quantity determining module, configured to determine, according to a preset data packet selection criterion, a number of data packets that meet the packet loss condition;

The packet loss control module is configured to perform selective packet loss according to the current packet loss rate and the number of data packets satisfying the packet loss condition.

Optionally, the packet loss rate determining module includes:

The packet loss weight determining unit is configured to determine a current packet loss weight according to the current CPU usage ratio;

The packet loss rate determining unit is configured to determine the current packet loss rate according to the current packet loss weight value.

Optionally, the packet loss weight determining unit includes:

a first threshold determining subunit, configured to determine whether the current CPU usage ratio is greater than a first threshold;

a first packet loss weight determining subunit, configured to determine that the current packet loss weight is increased by 1 if the current CPU usage is greater than the first threshold;

a second threshold determining subunit, configured to determine whether the CPU usage is less than a second threshold if the current CPU usage is less than or equal to the first threshold;

The second packet loss determining unit is configured to determine whether the current packet loss value is zero if the current CPU usage is less than the second threshold, and if the current value is zero, determine the current lost The weight of the packet is unchanged, if not zero, it is determined that the current packet loss value is decreased by 1;

The CPU usage cycle acquires the sub-unit, and if the current CPU usage is greater than or equal to the second threshold, the CPU usage is cyclically acquired after the preset time interval.

Optionally, the data packet quantity determining module includes:

a packet information obtaining unit, configured to acquire data packet information, where the data packet information includes a protocol type, a data packet length, a source port, and a destination port of the data packet;

a protocol type determining unit, configured to detect whether the protocol type of the data packet is a preset protocol type, and if it is not a preset protocol type, determining that the data packet does not satisfy a packet loss condition;

a packet length determining unit, configured to: if the protocol type of the data packet is a preset protocol type, perform a preset protocol type marking on the data packet, and detect whether the data packet length of the data packet is greater than a preset length;

a packet quantity determining unit, configured to determine that the data packet does not satisfy the packet loss condition if the data packet length of the data packet is less than or equal to the preset length, if the data packet length of the data packet is greater than the The preset length determines the number of packets that meet the packet loss condition plus one.

Optionally, the data packet quantity determining module includes:

a flag detecting unit, before detecting whether the protocol type of the data packet is a preset protocol type, detecting whether the data packet includes a preset protocol type flag, and if yes, determining that the data packet does not satisfy a packet loss condition, if If not, it is detected whether the protocol type of the data packet is a preset protocol type.

Optionally, the data packet quantity determining module further includes:

The port detecting unit is configured to: after determining the number of data packets satisfying the packet loss condition, detecting whether the source port and/or the destination port of the data packet is a pre-release port, and if the port is pre-released, determining the data packet If the packet loss condition is not met, if the port is not pre-released, it is determined that the data packet satisfies the packet loss condition.

Optionally, the packet loss control module includes:

a packet loss determining unit, configured to set the number of packets according to the current packet loss rate and the packet loss condition Determine the number of lost packets based on the number of packages;

The packet loss control unit is configured to randomly select the packet with the number of lost packets in a data packet satisfying the packet loss condition to perform packet loss processing.

Embodiments of the present invention provide a storage medium including a computer readable storage medium having computer executable instructions stored thereon, the computer executable instructions being executed by a computer processor to implement the network congestion processing method described above. The embodiment of the present invention solves the problem that the selective packet loss processing or the selective packet loss processing increases the CPU load when the network is congested by the soft interrupt in the prior art, and the dynamically adjustable selective packet loss is implemented to ensure the normal operation of the system. run.

DRAWINGS

FIG. 1 is a flowchart of a network congestion processing method according to an embodiment;

2 is a flowchart of a network congestion processing method according to an embodiment;

3 is a flowchart of a network congestion processing method according to an embodiment;

4 is a flowchart of a network congestion processing method according to another embodiment;

FIG. 5 is a schematic structural diagram of a network congestion processing apparatus according to an embodiment.

detailed description

FIG. 1 is a flowchart of a network congestion processing method according to an embodiment. This embodiment may be configured to have a large number of data flows in a router, and a soft terminal occupies a large number of CPUs to cause network congestion. The method may be configured by the network congestion provided by this embodiment. The processing device is implemented, and the device can be implemented in software and/or hardware, including:

In step 110, the CPU usage of the current CPU of the soft interrupt is obtained, and whether the packet loss processing is performed is determined according to the current CPU usage. If the packet loss processing is performed, step 120 is performed. If the packet loss processing is not performed, step 110 is executed after the preset time.

Among them, the interrupt refers to the situation that the system has to be processed by the CPU immediately during the running of the program. The CPU pauses the process of terminating the executing program and then processing the new situation. The interrupt includes soft interrupt and hard interrupt. The hard interrupt is generated by hardware, such as a disk, a network card, or a clock. The soft interrupt is generated by a currently running process. CPU usage refers to the number of running programs The CPU resources used, the higher the CPU usage, indicates that many programs are currently running. When the CPU usage reaches 100%, it affects the normal operation of the basic functions of the intelligent terminal.

In this embodiment, when a large number of data flows exist in the router, the CPU usage of the frequent network soft interrupt may reach 100%, which affects the normal operation of the basic functions of the intelligent terminal, and is solved by selective packet loss.

Optionally, the CPU usage of the current soft interrupt is detected. If the CPU usage of the current soft interrupt is greater than or equal to 91%, the packet loss processing is determined. If the CPU usage of the current soft interrupt is less than 91%, the determination is not performed. Packet loss processing.

In step 120, the current packet loss rate is determined according to the current CPU usage.

The packet loss rate refers to the ratio of the number of dropped packets to the number of transmitted packets. Optionally, the packet loss rate may be 10%, that is, one data packet is discarded in every 10 data packets currently sent.

In this embodiment, the current packet loss rate is determined according to the current CPU usage rate, and the packet loss rate can be dynamically adjusted. When the network is congested, the packet loss rate is increased, and the CPU resource release efficiency is improved. When the amount of data is small, the packet loss is reduced. Rate to avoid large packet loss.

In step 130, the number of data packets satisfying the packet loss condition is determined according to a preset data packet selection criterion.

The preset data packet selection criterion refers to a packet selection condition of a selective packet loss preset in the system. To ensure the basic functions of the intelligent terminal, optionally, the discarded data packet does not include an interactive small data packet. , commonly used data packets, such as HTTPS, HTTP or DNS.

Optionally, the data packet satisfying the packet loss condition is selected in the buffer data packet in the buffer area of the Ethernet receiving and transmitting data packet.

In step 140, selective packet loss is performed according to the current packet loss rate and the number of data packets satisfying the packet loss condition.

In this embodiment, the data packet that meets the packet loss condition is determined according to the current packet loss rate and the number of data packets satisfying the packet loss condition, and the packet loss processing is performed. In the case that the buffer size of the Ethernet receiving and transmitting data packets is not reduced, the data packet in the buffer area is selectively packetized, and the packet loss processing is performed to introduce a new CPU load.

In this embodiment, the selective packet loss is performed according to the current packet loss rate that can be dynamically adjusted and the number of data packets that meet the packet loss condition, which solves the problem that the network packet is not properly lost when the soft interruption causes the network congestion in the prior art. Packet processing increases the CPU load and implements dynamically adjustable selective loss. Package to ensure the normal operation of the system.

Based on the above embodiment, step 140 may also be:

The number of lost packets is determined according to the current packet loss rate and the number of data packets satisfying the packet loss condition;

Optionally, the current packet loss rate may be 10%, the number of data packets satisfying the packet loss condition may be 50, the number of lost packets may be determined to be 5, and the system randomly selects 5 of the 50 data packets satisfying the packet loss condition. The packet is packetized.

2 is a flowchart of a method for processing network congestion according to an embodiment. On the basis of the foregoing embodiment, the current packet loss rate is determined according to the current CPU usage ratio, and the current packet loss weight is determined according to the current CPU usage. The current packet loss rate is determined according to the current packet loss weight. include:

In step 210, the CPU usage of the current CPU of the soft interrupt is obtained, and whether the packet loss processing is performed is determined according to the current CPU usage. If the packet loss processing is performed, step 220 is performed. If the packet loss processing is not performed, step 210 is executed after the preset time.

In step 220, the current packet loss weight is determined according to the current CPU usage.

The packet loss weight refers to an identifier used to indicate the current CPU usage. Optionally, the current packet loss weight may be 0, 1, or 2.

In step 230, the current packet loss rate is determined according to the current packet loss weight.

In this embodiment, when the system determines the current packet loss weight, an instruction for generating a preset packet loss rate is generated, and the selective packet loss processing is performed according to the instruction. For example, the preset packet loss rate may be 10%. Optionally, if the current packet loss weight is 1, determining that the current packet loss rate is 10%, and if the current packet loss weight is 2, determining that the current packet loss rate is 20%.

In step 240, the number of data packets satisfying the packet loss condition is determined according to a preset data packet selection criterion.

In step 250, selective packet loss is performed according to the current packet loss rate and the number of data packets satisfying the packet loss condition.

FIG. 3 is a network congestion processing method according to an embodiment. On the basis of the foregoing embodiment, a method for determining a current packet loss weight according to a current CPU usage ratio, including:

In step 310, the CPU usage of the current CPU of the soft interrupt is obtained, and whether the packet loss processing is performed is determined according to the current CPU usage. If the packet loss processing is performed, step 320 is performed. If the packet loss processing is not performed, step 310 is executed after the preset time.

In step 320, it is determined whether the current CPU usage is greater than a first threshold. If the current CPU usage is greater than the first threshold, it is determined that the current packet loss value is incremented by one, and step 350 is performed. If the current CPU usage is less than or equal to the first threshold, step 330 is performed.

Wherein, the first threshold may be 90%.

In step 330, it is determined whether the CPU occupancy is less than a second threshold. If the current CPU usage is less than the second threshold, step 340 is performed. If the current CPU usage is greater than or equal to the second threshold, return to step 310 after the preset time interval.

Alternatively, the second threshold may be 68% and the preset time interval may be 3 minutes.

In step 340, it is determined whether the current packet loss weight is zero. If it is zero, it is determined that the current packet loss weight is unchanged. If it is not zero, it is determined that the current packet loss weight is decremented by 1, and step 350 is performed.

Optionally, the current CPU usage is obtained. If the current CPU usage is greater than the first threshold (for example, 90%), for example, the current CPU usage may be 92%, and the current packet loss weight is determined to be 1; The rate is less than the second threshold (for example, it may be 68%). For example, the current CPU usage may be 50%. Since the current packet loss weight is a positive integer greater than or equal to 0, if the current packet loss weight is zero, the current loss is determined. The weight of the packet is unchanged. If the current packet loss is greater than zero, the current packet loss is decremented by 1. If the current CPU usage is less than or equal to the first threshold and greater than or equal to the second threshold, for example. The current CPU usage can be 78%, and the current packet loss weight remains unchanged.

In step 350, the current packet loss rate is determined according to the current packet loss weight.

Optionally, if the current packet loss weight increases, the current current packet loss rate increases. If the current packet loss weight decreases, the corresponding current packet loss rate decreases, and the dynamic adjustment of the packet loss rate is implemented to avoid A fixed-size packet loss rate cannot quickly reduce the CPU usage of a soft interrupt when the network is heavily congested, or discard a large number of data packets when the CPU usage is not high, causing the system to retransmit the data packet.

Optionally, the CPU usage at different times is obtained according to the preset time interval, and the packet loss weight and the packet loss rate at different times are determined, and the dynamic adjustment of the packet loss rate is implemented.

In step 360, the number of data packets satisfying the packet loss condition is determined according to a preset data packet selection criterion.

In step 370, selective packet loss is performed according to the current packet loss rate and the number of data packets satisfying the packet loss condition.

In this embodiment, the current CPU usage ratio is compared with a preset threshold, and the current packet loss weight and packet loss rate are determined according to the comparison result. The CPU usage rate at different times is obtained at a preset time interval, and the packet loss rate is dynamically implemented. Adjustment.

FIG. 4 is a flowchart of a network congestion processing method according to another embodiment. On the basis of the foregoing embodiment, a method for determining a number of data packets that meets a packet loss condition is provided, including:

In step 410, the CPU usage of the current CPU of the soft interrupt is obtained, and whether the packet loss processing is performed is determined according to the current CPU usage. If the packet loss processing is performed, step 420 is performed. If the packet loss processing is not performed, step 410 is executed after the preset time.

In step 420, the current packet loss rate is determined according to the current CPU usage.

In step 430, the packet information is obtained, wherein the packet information includes a protocol type of the data packet, a packet length, a source port, and a destination port.

The protocol type of the data packet refers to a four-layer protocol type, including an application process, a transport layer, a network interconnection layer, and a host to a network layer; a packet length refers to the number of bytes included in the data packet; the source port refers to It is the local terminal packet sending port, and the destination port refers to the packet receiving port of the destination terminal.

Alternatively, the packet information can be obtained by an unpacking method. In the unpacking process, the second-layer packet type of the data packet is first obtained, and the unpacking mode is determined according to the second-layer packet type. Optionally, if the data packet is a Point-to-Point Protocol over Ethernet (PPPOE) type packet on the Ethernet, the packet header of the data packet needs to be stripped first.

The related pointer moves down 8 bytes to obtain packet information such as the protocol type, destination port, and source port of the data packet. After obtaining the information, the data packet needs to be restored in time. The data packet is an IP (Internet Protocol) protocol type packet, after stripping the packet header,

The related pointer can directly acquire the packet information without shifting.

In step 440, it is detected whether the protocol type of the data packet is a preset protocol type. If the packet If the protocol type is not the default protocol type, it is determined that the data packet does not meet the packet loss condition. If the protocol type of the data packet is the preset protocol type, step 460 is performed.

In this embodiment, the default protocol type may be a UDP (User Datagram Protocol) and a TCP (Transmission Control Protocol) of the transport layer. If the protocol type of the data packet does not belong to UDP or TCP, It is determined that the data packet does not satisfy the packet loss condition and is not a data packet within the packet loss scope.

In this embodiment, the packet loss process only discards the data packets of the preset protocol type (UDP and TCP), and ensures that the basic services of the terminal are not affected.

In step 450, the data packet is subjected to a preset protocol type flag, and it is detected whether the data packet length of the data packet is greater than a preset length. If the data packet length of the data packet is less than or equal to the preset length, it is determined that the data packet does not meet the packet loss condition. If the data packet length of the data packet is greater than the preset length, step 470 is performed.

The preset length may be 256 bytes, and the data packet length is less than or equal to 256 bytes, determining that the data packet does not satisfy the packet loss condition, and is not in the packet loss range; if the data packet length is greater than 256 bytes Then, it is determined that the data packet satisfies the packet loss condition and belongs to the data packet within the packet loss range. Optionally, the data packet that meets the packet loss condition may be a BT (Bit Torrent) download data packet.

In this embodiment, the data packet whose packet length is longer than the preset length is selected for discarding, which can speed up the adjustment efficiency of the CPU usage.

In step 460, it is determined that the number of packets satisfying the packet loss condition is increased by one. At the same time, returning to step 430, the next packet information is obtained.

In this embodiment, when it is determined that the data packet satisfies the packet loss condition, it is determined that the number of data packets satisfying the packet loss condition is increased by one. When it is determined that the data packet does not satisfy the packet loss condition, the number of data packets satisfying the packet loss condition is determined to be unchanged.

In step 470, selective packet loss is performed according to the current packet loss rate and the number of data packets satisfying the packet loss condition.

In this embodiment, the packet is selected according to the protocol type and the packet length of the data packet, and the data packet is selected in the discardable data packet according to the current packet loss rate and the number of data packets satisfying the packet loss condition, and the packet loss processing is performed. That is, when the network is guaranteed to operate normally, packet loss processing is performed, and selective packet loss is implemented.

Based on the above embodiment, before step 440, the method may further include:

Check if the packet contains a preset protocol type tag. If yes, it is determined that the data packet does not satisfy the packet loss condition. If not, step 450 is performed.

In this embodiment, when it is detected that the protocol type of the data packet belongs to the preset protocol type, the data packet is marked with a preset protocol type, indicating that the data packet is detected. Since the discarded data packet is randomly determined according to the packet loss rate, there is a data packet that is marked but not discarded, and the software interrupt processing function may make multiple calls to the same data packet. To avoid repeated detection of the same data packet, Before performing the preset packet selection standard detection on the data packet, it is detected whether the data packet contains the preset protocol type flag. If it exists, the data packet is not detected, the detection times are reduced, and the processing efficiency is improved.

Based on the foregoing embodiment, after the step 460, the method may further include:

Check if the source port and/or destination port of the packet is a pre-release port. If the port is pre-released, it is determined that the data packet does not meet the packet loss condition. If the port is not pre-released, it is determined that the data packet satisfies the packet loss condition.

In this embodiment, the pre-release port may be, for example, 80 ports, 443 ports, 53 ports, or 22 ports, etc., and the hash table is established according to the pre-release port, if the source port or the destination port of the data packet belongs to a hash. A pre-release port in the table determines that the data packet does not meet the packet loss condition. If the source port and the destination port of the data packet do not belong to the pre-release port in the hash table, it is determined that the data packet satisfies the packet loss condition.

In this embodiment, the data packet that the source port and/or the destination port belongs to the pre-release port is determined as a data packet that does not satisfy the packet loss condition, and the data packet loss is reduced.

FIG. 5 is a network congestion processing apparatus according to another embodiment 5, where the apparatus is configured to perform the network congestion processing method provided by the embodiment of the present invention, including:

The packet loss processing judging module 510 is configured to obtain the CPU usage of the current CPU of the soft interrupt, and determine whether to perform packet loss processing according to the current CPU usage rate;

The packet loss rate determining module 520 is configured to determine a current packet loss rate according to the current CPU usage rate when determining that the packet loss processing is performed;

The packet quantity determining module 530 is configured to determine, according to the preset data packet selection criterion, the number of data packets that meet the packet loss condition;

The packet loss control module 540 is configured to enter according to the current packet loss rate and the number of data packets satisfying the packet loss condition. Row selective packet loss.

Optionally, the packet loss rate determining module 520 includes:

The packet loss weight determining unit 521 is configured to determine a current packet loss weight according to the current CPU usage rate;

The packet loss rate determining unit 522 is configured to determine a current packet loss rate according to the current packet loss weight.

Optionally, the packet loss weight determining unit 521 includes:

The first threshold determining subunit 5211 is configured to determine whether the current CPU usage ratio is greater than a first threshold;

The first packet loss weight determining sub-unit 5212 is configured to determine that the current packet loss weight is increased by 1 if the current CPU usage is greater than the first threshold;

The second threshold determining sub-unit 5213 is configured to determine whether the CPU usage ratio is less than a second threshold if the current CPU usage is less than or equal to the first threshold;

The second packet loss determining unit 5214 is configured to determine whether the current packet loss value is zero if the current CPU usage is less than the second threshold, and if the value is zero, determine that the current packet loss value does not change. If it is not zero, it determines that the current packet loss weight is reduced by 1;

The CPU usage cycle acquisition sub-unit 5215 is configured to cycle to acquire a new CPU usage rate after the preset time interval if the current CPU usage is greater than or equal to the second threshold.

Optionally, the data packet quantity determining module 530 includes:

The packet information obtaining unit 531 is configured to acquire data packet information, where the data packet information includes a protocol type of the data packet, a data packet length, a source port, and a destination port;

The protocol type determining unit 532 is configured to detect whether the protocol type of the data packet is a preset protocol type, and if it is not the preset protocol type, determine that the data packet does not satisfy the packet loss condition;

The packet length determining unit 533 is configured to: if the protocol type of the data packet is a preset protocol type, perform a preset protocol type marking on the data packet, and detect whether the data packet length of the data packet is greater than a preset length;

The packet quantity determining unit 534 is configured to determine that the data packet does not satisfy the packet loss condition if the data packet length of the data packet is less than or equal to the preset length, and if the data packet length of the data packet is greater than the preset length, determine that the packet loss is satisfied. The number of packets for the condition is increased by one.

Optionally, the data packet quantity determining module 530 includes:

The flag detecting unit 535 checks before detecting whether the protocol type of the data packet is a preset protocol type. Whether the data packet contains a preset protocol type flag, if it is included, determines that the data packet does not satisfy the packet loss condition, and if not, detects whether the protocol type of the data packet is a preset protocol type.

Optionally, the data packet quantity determining module 530 further includes:

The port detecting unit 536 is configured to detect whether the source port and/or the destination port of the data packet are pre-release ports after determining the number of data packets satisfying the packet loss condition, and if the port is pre-released, determine that the data packet is not satisfied. The packet condition, if it is not pre-released, determines that the packet satisfies the packet loss condition.

Optionally, the packet loss control module 540 includes:

The packet loss determining unit 541 is configured to determine the number of lost packets according to the current packet loss rate and the number of data packets satisfying the packet loss condition;

The packet loss control unit 542 is configured to randomly select a packet of the number of lost packets in the data packet satisfying the packet loss condition to perform packet loss processing.

The network congestion processing apparatus provided by the embodiment of the present invention may perform the network congestion processing method provided by any embodiment of the present invention, and has a function module and a beneficial effect corresponding to the execution method.

A computer readable storage medium storing computer executable instructions arranged to perform the network congestion processing method described above.

Industrial applicability

The present disclosure solves the problem that the selective packet loss processing or the selective packet loss processing increases the CPU load when the network is congested due to the soft interruption in the prior art, and the dynamically adjustable selective packet loss is realized to ensure the normal operation of the system.

Claims

A network congestion processing method includes:

Obtaining the CPU usage of the current CPU of the soft interrupt, and determining whether to perform packet loss processing according to the current CPU usage;

When it is determined that the packet loss processing is performed, determining a current packet loss rate according to the current CPU usage rate;

Determining the number of data packets satisfying the packet loss condition according to a preset data packet selection criterion;

Selective packet loss according to the current packet loss rate and the number of data packets satisfying the packet loss condition.
The method of claim 1, wherein determining the current packet loss rate according to the current CPU usage comprises:

Determining a current packet loss weight according to the current CPU usage;

Determining the current packet loss rate according to the current packet loss weight.
The method of claim 2, wherein determining the current packet loss weight according to the current CPU usage comprises:

Determining whether the current CPU usage is greater than a first threshold;

If the current CPU usage is greater than the first threshold, determining that the current packet loss value is increased by one, and if the current CPU usage is less than or equal to the first threshold, determining whether the CPU usage rate is Less than the second threshold;

If the current CPU usage is less than the second threshold, determining whether the current packet loss value is zero. If it is zero, determining that the current packet loss weight is unchanged. If not, determining The current packet loss weight is decreased by one;

If the current CPU usage is greater than or equal to the second threshold, the new CPU usage is cyclically acquired after the preset time interval.
The method according to claim 1, wherein the number of data packets satisfying the packet loss condition is determined according to a preset data packet selection criterion, including:

Obtaining packet information, where the packet information includes a protocol type of the data packet, a data packet length, a source port, and a destination port;

Determining whether the protocol type of the data packet is a preset protocol type, if it is not a preset protocol type, determining that the data packet does not satisfy a packet loss condition, and if it is a preset protocol type, performing a preset protocol type on the data packet Marking, and detecting whether the length of the data packet of the data packet is greater than a preset length;

If the data packet length of the data packet is less than or equal to the preset length, determining that the data packet does not satisfy the packet loss condition, and if the data packet length of the data packet is greater than the preset length, determining to satisfy The number of packets with packet loss conditions is increased by one.
The method of claim 4, wherein before detecting whether the protocol type of the data packet is a preset protocol type, the method further comprises:

Detecting whether the data packet includes a preset protocol type identifier, if yes, determining that the data packet does not satisfy the packet loss condition, and if not, detecting whether the protocol type of the data packet is a preset protocol type.
The method according to claim 4, wherein after determining the number of data packets satisfying the packet loss condition, the method further comprises:

Detecting whether the source port and/or the destination port of the data packet is a pre-release port. If the port is pre-released, it is determined that the data packet does not meet the packet loss condition. If the port is not pre-released, the data is determined. The packet satisfies the packet loss condition.
The method according to claim 1, wherein the selective packet loss according to the current packet loss rate and the number of data packets satisfying the packet loss condition comprises:

Determining the number of lost packets according to the current packet loss rate and the number of data packets satisfying the packet loss condition;

The data packet with the number of lost packets is randomly selected in the data packet satisfying the packet loss condition for packet loss processing.
A network congestion processing device includes:

The packet loss processing judging module is configured to obtain the CPU usage of the current CPU of the soft interrupt, and determine whether to perform packet loss processing according to the current CPU usage rate;

The packet loss rate determining module is configured to determine a current packet loss rate according to the current CPU usage rate when determining that the packet loss processing is performed;

a data packet quantity determining module, configured to determine, according to a preset data packet selection criterion, a number of data packets that meet the packet loss condition;

The packet loss control module is configured to perform selective packet loss according to the current packet loss rate and the number of data packets satisfying the packet loss condition.
The apparatus according to claim 8, wherein the packet loss rate determining module comprises:

The packet loss weight determining unit is configured to determine a current packet loss weight according to the current CPU usage ratio;

The packet loss rate determining unit is configured to determine the current packet loss rate according to the current packet loss weight, where the packet loss rate determining unit includes:

a first threshold determining subunit, configured to determine whether the current CPU usage ratio is greater than a first threshold;

a first packet loss weight determining subunit, configured to: if the current CPU usage is greater than the first threshold a value, determining that the current packet loss weight is increased by one;

a second threshold determining subunit, configured to determine whether the CPU usage is less than a second threshold if the current CPU usage is less than or equal to the first threshold;

The second packet loss determining unit is configured to determine whether the current packet loss value is zero if the current CPU usage is less than the second threshold, and if the current value is zero, determine the current lost The weight of the packet is unchanged, if not, it is determined that the current weight loss value is decreased by 1;

The CPU usage cycle acquires the sub-unit, and if the current CPU usage is greater than or equal to the second threshold, the CPU usage is cyclically acquired after the preset time interval;

The packet loss control module includes:

The packet loss determining unit is configured to determine the number of lost packets according to the current packet loss rate and the number of data packets satisfying the packet loss condition;

The packet loss control unit is configured to randomly select the packet with the number of lost packets in a data packet satisfying the packet loss condition to perform packet loss processing.
The apparatus of claim 8, wherein the data packet quantity determining module comprises:

a packet information obtaining unit, configured to acquire data packet information, where the data packet information includes a protocol type, a data packet length, a source port, and a destination port of the data packet;

a protocol type determining unit, configured to detect whether the protocol type of the data packet is a preset protocol type, and if it is not a preset protocol type, determining that the data packet does not satisfy a packet loss condition;

a packet length determining unit, configured to: if the protocol type of the data packet is a preset protocol type, perform a preset protocol type marking on the data packet, and detect whether the data packet length of the data packet is greater than a preset length;

a packet quantity determining unit, configured to determine that the data packet does not satisfy the packet loss condition if the data packet length of the data packet is less than or equal to the preset length, if the data packet length of the data packet is greater than the The preset length determines the number of packets that meet the packet loss condition plus one;

a flag detecting unit, before detecting whether the protocol type of the data packet is a preset protocol type, detecting whether the data packet includes a preset protocol type flag, and if yes, determining that the data packet does not satisfy a packet loss condition, if If not, detecting whether the protocol type of the data packet is a preset protocol type;

The port detecting unit is configured to: after determining the number of data packets satisfying the packet loss condition, detecting whether the source port and/or the destination port of the data packet is a pre-release port, and if yes, determining that the data packet is not The packet loss condition is met, and if not, it is determined that the data packet satisfies the packet loss condition.
A computer readable storage medium storing computer executable instructions arranged to perform the method of claims 1-7.