WO2017114231A1

WO2017114231A1 - Packet sending method, tcp proxy, and tcp client

Info

Publication number: WO2017114231A1
Application number: PCT/CN2016/111079
Authority: WO
Inventors: 王文涛; 朱坤; 秦韵
Original assignee: 华为技术有限公司
Priority date: 2015-12-30
Filing date: 2016-12-20
Publication date: 2017-07-06
Also published as: CN106936730B; CN106936730A

Abstract

Disclosed in embodiments of the present invention are a packet sending method, a TCP proxy, and a TCP client, for use in effectively avoiding entrance of an excessively-great quantity of packets into a network in a instantaneous time, and in improving the TCP throughput. The method in an embodiment of the present invention comprises: sending M packets to a TCP client; receiving N TCP response packets of the M packets, sent by the TCP client; calculating a sample transmission bandwidth of a TCP connection between a TCP proxy and the TCP client according to each sample TCP response packet, the sample TCP response packet being at least one of the N TCP response packets; determining an estimated-bandwidth sample set according to the according to the sample transmission bandwidth, the estimated-bandwidth sample set comprising the sample transmission bandwidth calculated according to each sample TCP response packet; determining an estimated bandwidth according to the estimated-bandwidth sample set; determining a target bandwidth according to the estimated bandwidth; and determining the quantity of available tokens according to the target bandwidth, and controlling the sending of a packet according to the quantity of the available tokens.

Description

Message sending method, TCP proxy and TCP client

Technical field

The present invention relates to the field of communications, and in particular, to a packet sending method, a TCP proxy, and a TCP client.

Background technique

Network congestion refers to the excessive number of packets arriving at a certain part of the communication network, which makes the part of the network unable to process the packets, which may cause the performance of the network or the entire network to degrade. In severe cases, the network may even be caused. The communications business has come to a standstill, a phenomenon known as congestion. Congestion is a state of continuous overloaded network where users' demand for network resources (including link bandwidth, storage space, and processor processing power) exceeds the inherent capacity.

In the prior art, an application based on a Transmission Control Protocol (English name: TCP) controls the number of packets entering the network through a TCP sliding window mechanism. When the network is congested, the TCP congestion control adjusts the transmission window to adapt to the network quality according to the additive-increase multiplicative-decrease (English abbreviation: AIMD) algorithm.

Since the participants of the TCP sliding window mechanism are only the transmitting end and the receiving end of a single TCP connection, it only considers the receiving capability of the receiving end, and does not take into account the transmission capability of the network, and it is prone to occur that the network instantaneously enters too many messages and appears. Traffic bursts. For example, when a network intermediate device, such as a router, is busy or congested, the packet sending or receiving capability based on the sliding window exceeds the network transmission capability, and the network cannot process too many packets, that is, there is no packet. The transmission rate is controlled. At this time, the packets in the window are discarded by the network device, and the network transmission quality is deteriorated, eventually causing the TCP throughput to drop.

Summary of the invention

The embodiment of the invention provides a packet sending method, a TCP proxy and a TCP client, which are used to effectively prevent the network from entering too many packets in an instant and improve the TCP throughput.

In view of the above, the first aspect of the present invention provides a packet sending method, including:

After the TCP proxy sends M packets to the TCP client, it receives N TCP response packets of M messages fed back by the TCP client. After receiving the N TCP response packets, the TCP proxy can receive the TCP response packets. The sample TCP response message is determined in the paper, and the sample transmission bandwidth of the TCP connection between the TCP proxy and the TCP client is calculated according to each sample TCP response message.

The TCP proxy determines the estimated bandwidth sample set according to the calculated sample transmission bandwidth, determines the estimated bandwidth according to the estimated bandwidth sample set, determines the target bandwidth according to the estimated bandwidth, determines the available token token number according to the target bandwidth, and finally controls the packet according to the available token number. send.

In the embodiment of the present invention, whether the packet is sent out is controlled by the TCP proxy, and the specific expression is: whether the packet is sent by the number of tokens, and the token number is determined by the target bandwidth, so Effectively avoid the network from entering too many messages in an instant and improve TCP throughput.

Preferably, in the embodiment of the present invention, the TCP proxy determines the number of available token tokens according to the target bandwidth, and finally controls the sending of the packet according to the number of available tokens. The following are possible implementation manners, as follows:

In an implementation manner, the TCP proxy determines the number of available token tokens according to the target bandwidth, and finally controls the sending of the packet according to the number of available tokens, including:

The TCP proxy can set the token growth rate according to the target bandwidth, and then determine whether the number of tokens obtained according to the token growth rate reaches a preset value. If yes, the packet is sent to the TCP client.

In this implementation manner, the TCP proxy controls the sending of the packet by determining the number of tokens to increase according to the target bandwidth, and determining whether the number of tokens obtained according to the token growth rate reaches a preset value. The TCP proxy sends the message to the TCP client.

That is, in the embodiment of the present invention, a specific implementation manner for the TCP proxy to control the sending of the packet is provided, which is used to effectively prevent the network from entering too many packets in an instant and improve the TCP throughput.

In a possible design, the message sending method described above may further include:

The TCP proxy can set a sending timer and start the sending timer. When the sending timer reaches the preset duration, it performs the step of determining whether the number of tokens obtained according to the token number increasing rate reaches a preset value.

In the embodiment of the present invention, when the TCP proxy determines the token obtained according to the growth rate of the number of tokens When the number does not reach the preset value, the message that needs to be sent is cached in the TCP proxy side socket buffer waiting to be sent. It needs to wait for the instruction to trigger the sending of the message, that is, the TCP proxy needs to wait. When the secondary packet has been sent successfully, the TCP client sends an acknowledgement message (English name: Acknowledgement, English abbreviation: ACK). At this time, the TCP proxy will perform the determination again to determine whether the number of tokens that grow according to the number of tokens reaches the preset. The numerical step, however, due to network failure or other reasons of communication, often causes the ACK round trip time (English name: round-trip time, English abbreviation: RTT) to be too long, in this case, even if the number of tokens has reached the pre- If the value is set, the data in the buffer will not be sent out for a long time. In this embodiment of the present invention, the TCP proxy sets a sending timer as long as the sending timer reaches the preset. When the duration is long, a step of determining whether the number of tokens obtained according to the growth rate of the token reaches a preset value is performed, and when When the number of tokens reaches the preset value, the TCP proxy sends a message to the TCP client. That is, it can effectively avoid the situation that the data in the socket socket buffer cannot be sent out for a long time.

In a possible design, the message sending method may further include:

The TCP proxy can set the maximum depth of the token resource queue and the token resource queue. The token resource queue is used to store the token obtained by the TCP proxy according to the number of tokens. The maximum depth is used to indicate that the token resource queue can be stored according to the number of tokens. The maximum number of tokens obtained.

In the embodiment of the present invention, the token resource queue is used to store the token obtained by the TCP proxy according to the number of tokens, and the maximum depth is used to indicate the maximum number of tokens that can be stored in the token resource queue according to the token growth rate. In practical applications, on the one hand, if there is no packet to be sent by the TCP proxy in a certain period of time, at this time, more tokens are generated according to the number of tokens, and the maximum depth of the token resource queue is constrained. The token growth cannot exceed a certain threshold. On the other hand, the maximum depth of the token resource queue plays a buffering role. That is, when the TCP proxy does not need to send packets in a period of time, the token resource queue can A certain number of tokens are saved. In this case, if a large number of packets need to be sent, the number of packets allowed to be sent can be maximized, and the relative rate of sending packets is relatively uniform.

In one possible design, the TCP proxy calculates the TCP generation based on each sample TCP response message. The sample transmission bandwidth of the TCP connection between the TCP client and the TCP client is specifically:

For each sample TCP response message, the TCP proxy can calculate the sample transmission bandwidth of the TCP connection according to the sending time of the packet corresponding to the sample TCP response message, the receiving time of the sample TCP response message, and the length of the corresponding packet. .

In one possible design, the TCP proxy may also include before determining the target bandwidth based on the estimated bandwidth:

The TCP proxy can determine the service type of the TCP connection between the TCP clients, and then obtain the policy configuration information, where the policy configuration information includes the correspondence between the adjustment ratio of the target bandwidth and the service type, and the policy configuration information is provided by the user through the TCP proxy. The product configuration interface is set, and then the adjustment ratio of the target bandwidth can be determined according to the service type and the policy configuration information, and then the target bandwidth is determined according to the estimated bandwidth and the policy configuration information;

At this point, the TCP proxy determines the target bandwidth based on the estimated bandwidth and policy configuration information:

The TCP proxy calculates the target bandwidth according to the following formula:

Target bandwidth = adjustment ratio of target bandwidth × estimated bandwidth.

In the embodiment of the present invention, the TCP proxy can receive the policy configuration information input by the user through the TCP proxy. The policy configuration information can include the correspondence between the adjustment ratio of the target bandwidth and the service type. Further, the bandwidth of the target bandwidth The adjustment ratio may specifically be a floating upper limit value of the target bandwidth or a floating lower limit value of the target bandwidth or a percentage adjustment of the target bandwidth. That is, in the embodiment of the present invention, the target bandwidth is determined not only according to the estimated bandwidth determined by the TCP proxy, but also the target bandwidth is determined according to the requirement of the TCP connection service type between the TCP proxy and the TCP client. In an actual application, the corresponding policy configuration information may be manually selected according to the actual service type requirement, such as a video service or a picture service. The TCP proxy can make the set number of tokens grow more reasonable and more accurate through the policy configuration information input by the user and the target bandwidth determined by the bandwidth estimation. In practical applications, the network can prevent the network from entering too many packets in an instant. Improve TCP throughput.

In another implementation manner, the TCP proxy controls the sending of the packet, including:

The TCP proxy can send the target bandwidth to the TCP client, so that the TCP client sets the token growth rate according to the target bandwidth, and determines whether the number of tokens obtained according to the token number growth rate reaches a preset value, and if so, sends a report. Text.

In this implementation manner, the TCP proxy sends the target bandwidth to the TCP client, so that the TCP client sets the token growth rate according to the target bandwidth, and determines whether the number of tokens obtained according to the token number growth rate reaches a preset value. The TCP client sends the message. That is, another specific implementation manner is provided for the TCP proxy to control the sending of the packet, which is used to effectively prevent the network from entering too many packets in an instant and improve the TCP throughput.

In a second aspect, an embodiment of the present invention provides a packet sending method, including:

The TCP client can receive the target bandwidth sent by the TCP proxy, wherein the target bandwidth is determined by the TCP proxy according to the estimated bandwidth, and the estimated bandwidth is determined by the TCP proxy according to the estimated bandwidth sample set, and the estimated bandwidth sample set is determined by the TCP proxy according to the sample transmission bandwidth. The sample transmission bandwidth is obtained by the TCP proxy calculating the TCP connection between the TCP proxy and the TCP client according to each sample TCP response message, and the sample TCP response packet is at least one of N TCP response messages, and N TCP response reports. The TCP client receives the M packets sent by the TCP proxy and feeds back the packets to the TCP proxy. Then the TCP client sets the token growth rate according to the target bandwidth, and finally determines whether the number of tokens obtained according to the token growth rate is The preset value is reached. If it is reached, the TCP client sends the message.

In a possible design, the message sending method further includes:

The TCP client can set a sending timer and start the sending timer. When the sending timer reaches the preset duration, the TCP client determines whether the current number of tokens reaches the preset value.

When the TCP client determines that the number of tokens obtained according to the token growth rate does not reach the preset value, the packet to be sent is buffered in the TCP client socket socket buffer waiting to be sent, and needs to wait for the trigger to be sent. The arrival of the message instruction, that is, the TCP client needs to wait for the ACK that is sent back when the last message has been sent successfully. At this time, the TCP client will perform the determination again to determine whether the number of tokens that grow according to the number of tokens reaches the pre-up. The step of setting the value, however, due to network failure or other reasons, the communication often causes the ACK round trip time RTT to be too long. In this case, even if the number of tokens has reached the preset value, the TCP client side socket socket buffer In the embodiment of the present invention, the TCP client sets a sending timer, as long as the sending timer reaches the preset duration, the data in the area may not be sent out for a long time. Will perform a token that is determined by the rate of increase in the number of tokens. The step of whether the number reaches the preset value, and when the number of tokens reaches the preset value, the TCP client sends a message. That is to say, it can effectively avoid the situation that the data in the buffer cannot be sent out for a long time.

In a possible design, the message sending method may further include:

The TCP client can set the maximum depth of the token resource queue and the token resource queue. The token resource queue is used to store the token obtained according to the growth rate of the token. The maximum depth is used to indicate that the token resource queue can be stored according to the token growth rate. The maximum number of tokens.

In the embodiment of the present invention, the token resource queue is used to store the token obtained according to the growth rate of the number of tokens, and the maximum depth is used to indicate the maximum number of tokens that can be stored in the token resource queue according to the growth rate of the number of tokens. In practical applications, on the one hand, if there is no packet to be sent by the TCP client in a certain period of time, at this time, according to the number of tokens, the rate will generate more tokens, and the maximum depth constraint of the token resource queue. The token growth cannot exceed a certain threshold. On the other hand, the maximum depth of the token resource queue plays a buffering role, that is, when the TCP client does not need to send a message during a period of time, the token resource The queue can store a certain number of tokens. In this case, if a large number of packets need to be sent, the maximum number of packets allowed to be sent can be increased, and the relative rate of sending packets is relatively uniform.

In a possible design, the target bandwidth is obtained by the TCP proxy according to the estimated bandwidth and the policy configuration information, where the policy configuration information may specifically include a correspondence between the adjustment ratio of the target bandwidth and the service type, and the policy configuration information may be provided by using a TCP proxy. The product configuration interface is set. Further, the adjustment ratio of the target bandwidth may specifically be a floating upper limit value of the target bandwidth or a floating lower limit value of the target bandwidth or a percentage adjustment of the target bandwidth.

That is, in the embodiment of the present invention, the target bandwidth is determined not only according to the estimated bandwidth determined by the TCP proxy, but also the target bandwidth is determined according to the requirement of the TCP connection service type between the TCP proxy and the TCP client. The rate of the number of tokens set by the TCP client can be made more reasonable and more accurate. In practical applications, the network can be prevented from entering too many packets in an instant and the TCP throughput can be improved.

In a third aspect, an embodiment of the present invention provides a TCP proxy, where the TCP proxy has an implementation. The function of TCP proxy behavior in the method design. The functions may be implemented by hardware or by corresponding software implemented by hardware. The hardware or software includes one or more modules corresponding to the functions described above. The modules can be software and/or hardware.

In one possible design, the TCP proxy includes:

a sending module, configured to send M packets to the TCP client;

a receiving module, configured to receive N TCP response messages of M messages sent by the TCP client;

a processing module, configured to calculate, according to each sample TCP response message received by the receiving module, a sample transmission bandwidth of a TCP connection between the TCP proxy and the TCP client, where the sample TCP response packet is at least one of N TCP response packets And determining an estimated bandwidth sample set according to the sample transmission bandwidth, the estimated bandwidth sample set includes a sample transmission bandwidth calculated according to each sample TCP response message, determining an estimated bandwidth according to the estimated bandwidth sample set, and determining a target bandwidth according to the estimated bandwidth, according to the target bandwidth. Determine the number of available tokens and control the transmission of packets based on the number of available tokens.

In a possible design, the structure of the TCP proxy includes a transmitter, a receiver, and a processor, and the transmitter is configured to send M messages to the TCP client, and the receiver is configured to receive the TCP client. The N TCP response messages of the M messages sent by the terminal, the processor, configured to calculate a sample transmission bandwidth of the TCP connection between the TCP proxy and the TCP client according to each sample TCP response message received by the receiver The sample TCP response message is at least one of N TCP response messages, and the estimated bandwidth sample set is determined according to the sample transmission bandwidth, and the estimated bandwidth sample set includes the sample transmission bandwidth calculated according to each sample TCP response message, according to the estimation. The bandwidth sample set determines the estimated bandwidth, determines the target bandwidth according to the estimated bandwidth, determines the number of available token tokens according to the target bandwidth, and controls the sending of the packet according to the number of available tokens.

In a fourth aspect, an embodiment of the present invention provides a TCP client, which has a function of implementing TCP client behavior in the foregoing method design. The functions may be implemented by hardware or by corresponding software implemented by hardware. The hardware or software includes one or more modules corresponding to the functions described above.

In one possible design, the TCP client includes:

a receiving module, configured to receive a target bandwidth sent by the TCP proxy, where the target bandwidth is determined by the TCP proxy according to the estimated bandwidth, and the estimated bandwidth is determined by the TCP proxy according to the estimated bandwidth sample set, and the estimated bandwidth sample set is determined by the TCP proxy according to the sample transmission bandwidth, and the sample transmission is performed. Bandwidth is based on TCP proxy per The sample TCP response message is obtained by calculating the TCP connection between the TCP proxy and the TCP client, and the sample TCP response packet is at least one of N TCP response messages, and the N TCP response messages are TCP clients receiving the TCP. M packets sent by the proxy and then fed back to the TCP proxy packet;

a processing module, configured to set a token growth rate according to the target bandwidth, and determine whether the number of tokens obtained according to the token growth rate reaches a preset value;

And a sending module, configured to send a message if the processing module determines that the number of tokens obtained according to the number of tokens increases to a preset value.

In a possible design, the structure of the TCP client includes a receiver, a processor, and a transmitter, and the receiver is configured to receive a target bandwidth sent by the TCP proxy, and the target bandwidth is determined by the TCP proxy according to the estimated bandwidth, and the bandwidth is estimated. Determined by the TCP proxy according to the estimated bandwidth sample set, the estimated bandwidth sample set is determined by the TCP proxy according to the sample transmission bandwidth, and the sample transmission bandwidth is calculated by the TCP proxy according to each sample TCP response message to calculate the TCP connection between the TCP proxy and the TCP client. The sample TCP response message is at least one of the N TCP response messages, and the N TCP response messages are packets that the TCP client feeds back to the TCP proxy after receiving the M messages sent by the TCP proxy, and the processing is performed. And determining, according to the target bandwidth setting token growth rate, whether the number of tokens obtained according to the token number increasing rate reaches a preset value, and the transmitter is configured to: if the processor determines the token obtained according to the token number increasing rate When the number reaches the preset value, the message is sent.

In a fifth aspect, an embodiment of the present invention provides a computer storage medium, where the program storage code is stored in the computer storage medium, and the program code is used to indicate that the method of the first aspect or the second aspect is performed.

Compared with the prior art, it can be seen from the foregoing technical solution that, in the embodiment of the present invention, the TCP proxy sends M packets to the TCP client, and receives N TCP response packets of the M packets sent by the TCP client. The sample transmission bandwidth of the TCP connection between the TCP proxy and the TCP client is calculated according to each sample TCP response message, and the sample TCP response packet is at least one of N TCP response packets, and the estimated bandwidth is determined according to the sample transmission bandwidth. The sample set, the estimated bandwidth sample set includes a sample transmission bandwidth calculated according to each sample TCP response message, an estimated bandwidth is determined according to the estimated bandwidth sample set, a target bandwidth is determined according to the estimated bandwidth, and finally, the available token token number is determined according to the target bandwidth, The sending of the message is controlled according to the number of available tokens. Compared to the prior art, when sending a newspaper In this case, the packet transmission is limited by the token number, that is, the packet transmission rate is controlled, which effectively prevents the network from entering too many packets in an instant and improves the TCP throughput.

DRAWINGS

1 is a schematic structural diagram of a system according to an embodiment of the present invention;

2 is a schematic diagram of an embodiment of a method for sending a message according to an embodiment of the present invention;

FIG. 3 is a schematic structural diagram of a system in an embodiment of a packet sending method according to an embodiment of the present disclosure;

FIG. 4 is a schematic diagram of another embodiment of a packet sending method according to an embodiment of the present disclosure;

FIG. 5 is a schematic diagram of an embodiment of a TCP proxy according to an embodiment of the present invention; FIG.

FIG. 6 is a schematic diagram of another embodiment of a TCP proxy according to an embodiment of the present invention; FIG.

FIG. 7 is a schematic diagram of another embodiment of a TCP proxy according to an embodiment of the present invention; FIG.

FIG. 8 is a schematic diagram of an embodiment of a TCP client according to an embodiment of the present invention; FIG.

FIG. 9 is a schematic diagram of another embodiment of a TCP client according to an embodiment of the present invention.

detailed description

The embodiment of the invention provides a packet sending method, a TCP proxy and a client, which are used to effectively prevent the network from entering too many packets in an instant and improve the TCP throughput.

Referring to FIG. 1, in the embodiment of the present invention, a TCP client establishes a TCP connection with a TCP server through an SCG gateway.

The TCP client may be a device that provides voice and/or data connectivity to the user, such as a handheld device with wireless connectivity, or other processing device connected to the wireless modem.

The TCP proxy can be deployed on the SCG gateway or can be an independent network element in the network. The network element can be deployed between the TCP client and the SCG gateway, or between the SCG gateway and the TCP server.

The packets sent by the TCP server to the TCP client are sent to the TCP proxy first, and then sent to the TCP client by the TCP proxy.

In order to prevent the network from entering too many packets in an instant, the TCP proxy needs to send and control according to the current bandwidth of the network before sending the packet to the TCP client, which may be based on the target. The bandwidth determines the number of token tokens available, and then controls the transmission of the message based on the number of available tokens.

In the embodiment of the present invention, the TCP proxy performs transmission control in a plurality of manners, which are respectively described below:

1. The TCP proxy determines the growth rate of the number of tokens according to the target bandwidth and determines whether the number of tokens obtained according to the growth rate of the token reaches a preset value. If yes, the packet is sent to the TCP client:

Referring to FIG. 2, an embodiment of a packet sending method according to an embodiment of the present invention includes:

201. The TCP proxy sends M packets to the TCP client.

In the embodiment of the present invention, the TCP proxy can send M packets to the TCP client, wherein after the TCP proxy sends M packets to the TCP client, in the embodiment of the present invention, the TCP proxy can also receive the TCP client to send. N TCP response messages of the above M messages.

It should be noted that, in this embodiment, M is a positive integer greater than or equal to 1, and N is also a positive integer greater than or equal to 1, and M is greater than or equal to N, and specific values are not limited herein.

It should be noted that the TCP proxy may also be a processing unit deployed on a TCP server with a specific function, or may be a network device that is deployed in a data exchange process between the TCP server and the TCP client. For example, in the gateway of the service control gateway (English name: service control gateway, English abbreviation: SCG), it may also be an independent network device in the network, which is not limited herein.

For ease of understanding, in the embodiment of the present invention, the TCP proxy is deployed in the SCG gateway as an example, and the interaction process between the SCG gateway and the TCP client is as shown in FIG. 3:

Specifically, the TCP client is used as an example to describe the TCP client. The TCP server receives the get packet sent by the TCP client through the TCP proxy. After receiving the get packet, the TCP proxy performs protocol parsing on the get packet to identify the TCP client. The requested service type, and forwards the packet sent by the TCP server according to the service type to the TCP client. After receiving the packet, the TCP client may send a TCP response packet to the TCP proxy.

202. The TCP proxy calculates a sample transmission bandwidth of the TCP connection between the TCP proxy and the TCP client according to each sample TCP response packet.

In the embodiment of the present invention, the TCP proxy can calculate the TCP generation according to each sample TCP response message. The sample transmission bandwidth of the TCP connection between the TCP client and the TCP client is actually the sample transmission bandwidth of the TCP connection between the SCG gateway and the TCP client.

It should be noted that the sample TCP response packet may be one or more of the N response packets, which is not limited herein.

Specifically, when the TCP proxy receives the N response packets sent by the TCP client, the TCP proxy may send, according to the sending time of the packet corresponding to the sample TCP response packet, the TCP response message. The sample transmission bandwidth of the TCP connection is calculated by the reception time of the sample TCP response message and the length of the message corresponding to the sample TCP response message.

203. The TCP proxy determines an estimated bandwidth sample set according to the sample transmission bandwidth.

In the embodiment of the present invention, after the TCP proxy calculates the sample transmission bandwidth of the TCP connection between the TCP proxy and the TCP client according to each sample TCP response packet, the estimated bandwidth sample set may be determined according to the calculated sample transmission bandwidth, where The estimated bandwidth sample set includes a sample transmission bandwidth calculated by the TCP proxy according to each sample TCP response message.

It should be noted that the estimated bandwidth sample set may also include a sample transmission bandwidth calculated by the TCP proxy according to the at least one sample TCP response packet, which is not limited herein.

204. The TCP proxy determines an estimated bandwidth according to the estimated bandwidth sample set.

In the embodiment of the present invention, the TCP proxy may determine the estimated bandwidth according to the obtained estimated bandwidth sample set. Specifically, the TCP proxy may perform simple calculation on the sample transmission bandwidth calculated by the sample TCP response packet included in the estimated bandwidth sample set, for example, The average bandwidth is calculated as the estimated bandwidth. It can be understood that the estimated bandwidth can be obtained by other calculation methods, such as weighted average calculation, which is not limited herein.

205. The TCP proxy determines the target bandwidth according to the estimated bandwidth.

In the embodiment of the present invention, after determining the estimated bandwidth according to the estimated bandwidth sample set, the TCP proxy may determine the target bandwidth according to the estimated bandwidth. It should be noted that, before determining the target bandwidth according to the estimated bandwidth, the TCP proxy may also obtain policy configuration information, where the policy configuration information includes a correspondence between the adjustment ratio of the target bandwidth and the service type, and the policy configuration information is provided by the user through the TCP proxy. The configuration interface is set, and the adjustment ratio of the target bandwidth is determined according to the type of service identified by the TCP proxy.

In this embodiment, the specific manner in which the TCP proxy determines the target bandwidth according to the estimated bandwidth may be:

The TCP proxy determines the target bandwidth based on the estimated bandwidth and policy configuration information;

At this time, the TCP proxy determines the target bandwidth according to the estimated bandwidth and the policy configuration information, and specifically includes:

Target bandwidth = adjustment ratio of target bandwidth × estimated bandwidth.

It should be noted that the adjustment ratio of the target bandwidth may be a floating upper limit value of the target bandwidth or a floating lower limit value of the target bandwidth or a percentage adjustment of the target bandwidth, where the policy configuration information may be provided through a configuration interface provided by the TCP proxy. Input. For example, according to the actual service type requirement, it can be set through the product configuration interface provided by the TCP proxy. The configuration interface has a floating upper limit option of the target bandwidth or a floating lower limit option of the target bandwidth or a target percentage adjustment option for the user input. The floating upper limit of the target bandwidth, the floating lower limit of the target bandwidth, and the adjustment percentage of the target bandwidth. Here, the target bandwidth adjustment ratio is taken as the target floating upper limit value, and the TCP proxy determines according to the estimated bandwidth and the policy configuration information. The target bandwidth can be:

Target bandwidth = floating upper limit of the target bandwidth × estimated bandwidth.

It should be noted that the calculation method of the target bandwidth described by the above formula is only one of the calculation methods. In addition, according to the target bandwidth adjustment ratio and the estimated bandwidth, there may be other methods for calculating the target bandwidth, which is not limited herein. .

206. The TCP proxy sets a growth rate of the number of tokens according to the target bandwidth.

In the embodiment of the present invention, after the TCP proxy determines the target bandwidth, the token number growth rate may be set according to the target bandwidth.

207. The TCP proxy sets the maximum depth of the token resource queue and the token resource queue.

In the embodiment of the present invention, the TCP proxy may set a token resource queue, which is used to store a token obtained by the TCP proxy according to the number of tokens, and a maximum depth of the token resource queue, which is used to indicate that the token resource queue can be stored according to the number of tokens. The maximum number of tokens obtained by the growth rate.

In the embodiment of the present invention, if no packet needs to be sent by the TCP proxy in a certain period of time, at this time, more tokens are generated according to the token growth rate, and the maximum depth of the token resource queue constrains the token. The growth cannot exceed a certain threshold. On the other hand, when there is no packet to be sent by the TCP proxy in a period of time, the token resource queue can store a certain number of tokens. In this case, if a large number of packets need to be sent suddenly, The biggest increase is allowed to be sent The number of packets eventually reaches the uniform rate of sending packets.

It should be noted that there is no sequence relationship between step 207 and step 201 to step 206, which is not limited herein.

208, the TCP proxy sets the sending timer and starts the sending timer, when the sending timer reaches the preset duration, step 209 is performed;

In the embodiment of the present invention, the TCP proxy may set a sending timer and start the timer. When the sending timer reaches the preset duration, step 209 is performed.

In the embodiment of the present invention, when the TCP agent determines that the number of tokens obtained according to the number of tokens has not reached the preset value, the packets to be sent are buffered in the socket buffer of the TCP proxy side and are waiting to be sent. Waiting for the instruction to trigger the sending of the message, that is, the TCP proxy needs to wait for the ACK of the TCP client to feed back when the last sent message has been sent successfully. At this time, the TCP proxy will perform the determination again according to the number of tokens. The step of obtaining the number of tokens reaches the preset value. However, due to network failure or other communication reasons, the RTT of the ACK is often too long. In this case, even if the number of tokens has reached the preset value, the socket buffer In the embodiment of the present invention, the TCP proxy sets a sending timer, as long as the sending timer reaches the preset duration, when the data in the area is not sent out for a long time. The step of determining whether the number of tokens obtained according to the growth rate of the token reaches a preset value is performed, that is, it can be effectively Avoid the situation data socket buffer for a long time can not be sent out.

It can be understood that step 208 can be executed after step 206, which is not limited herein.

209, the TCP proxy determines whether the number of tokens currently obtained according to the number of token growth rate reaches a preset value, if yes, step 210 is performed;

In the embodiment of the present invention, the TCP proxy stores the packet that needs to be forwarded to the TCP client to the TCP proxy socket buffer for waiting to be sent. At this time, the TCP proxy determines whether the number of tokens in the token resource queue reaches the preset value. If it is determined that it is reached, then step 210 is performed. In addition, it can be understood that the specific value of the preset value is not limited herein.

210. The TCP proxy sends a packet to the TCP client.

In the embodiment of the present invention, when the TCP proxy determines that the number of tokens in the token resource queue reaches a preset value, the TCP proxy sends a packet to the TCP client.

For ease of understanding, the following is an example of a preset value of 3 tokens:

The TCP proxy stores 8 packets that need to be forwarded to the TCP client to the TCP proxy socket buffer for waiting to be sent. At this time, if the TCP proxy detects that the number of tokens in the token resource queue is five, the TCP proxy is from the socket. Three packets are fetched from the buffer, and the three packets are sent to the TCP client. It should be noted that the TCP proxy can also take out four packets from the buffer, or send other packets to the TCP client. , specifically here is not limited. It should be noted that, after the above-mentioned transmitted message is successfully sent to the TCP client, according to the Transmission Control Protocol/Internet Protocol (English: Voice Control Protocol/Internet Protocol, TCP/IP) TCP/IP protocol The TCP client sends an ACK to the TCP proxy. When the TCP proxy receives the ACK, it re-executes the step of determining whether the number of tokens in the token resource queue reaches a preset value. If the preset is reached, and the socket buffer is still If the packet is not sent, the packet in the socket buffer is sent. The details are not mentioned here.

If the TCP proxy detects that the number of tokens in the token resource queue is less than three, the above eight packets continue to be stored in the socket buffer and are waiting to be sent.

In addition, it should be noted that whenever the TCP proxy needs to forward the packet forwarded to the TCP client to the TCP proxy side socket buffer, it will determine whether the number of tokens in the token resource queue reaches the preset value.

In the embodiment of the present invention, the TCP proxy sends M packets to the TCP client, receives N TCP response packets of M messages sent by the TCP client, and calculates TCP proxy and TCP client according to each sample TCP response packet. The sample transmission bandwidth of the TCP connection between the terminals, the sample TCP response message is at least one of N TCP response messages, and the estimated bandwidth sample set is determined according to the sample transmission bandwidth, and the estimated bandwidth sample set includes a TCP response report according to each sample. Calculate the sample transmission bandwidth, determine the estimated bandwidth according to the estimated bandwidth sample set, determine the target bandwidth according to the estimated bandwidth, set the token growth rate according to the target bandwidth, and determine whether the number of tokens obtained according to the token number growth rate reaches a preset value. If it is reached, it sends a message to the TCP client. Compared with the prior art, when a packet is sent, the packet transmission is limited to the number of tokens, that is, the packet transmission rate is controlled, thereby effectively preventing the network from entering too many packets in an instant and improving the TCP throughput. .

2. The TCP proxy sends the target bandwidth to the TCP client, so that the TCP client sets the token growth rate according to the target bandwidth, and determines that the token growth rate is obtained according to the token number. Whether the number of obtained tokens reaches a preset value, and if so, the message is sent:

Referring to FIG. 4, another embodiment of a packet sending method according to an embodiment of the present invention includes:

Steps 401 to 405 are the same as steps 201 to 205 in the foregoing embodiment, and details are not described herein again.

406. The TCP proxy sends the target bandwidth to the TCP client.

In the embodiment of the present invention, after the TCP proxy determines the target bandwidth according to the estimated bandwidth, the target bandwidth may be sent to the TCP client, and the TCP client may receive the target bandwidth.

407. The TCP client sets a growth rate of the number of tokens according to the target bandwidth.

In the embodiment of the present invention, the TCP client may set a token resource queue, which is used to store the token obtained according to the growth rate of the token number, and may also set the maximum depth of the token resource queue, which is used to indicate that the token resource queue can be stored according to the number of tokens. The maximum number of tokens obtained by the rate.

In the embodiment of the present invention, if there is no packet to be sent by the TCP client in a certain period of time, more tokens are generated according to the number of tokens, and the maximum depth of the token resource queue constrains the token growth. It is impossible to exceed a certain threshold. On the other hand, when there is no packet to be sent by the TCP proxy in a period of time, the token resource queue can store a certain number of tokens. In this case, if a large number of packets need to be sent suddenly, the maximum can be The increase allows the number of packets to be sent, and finally achieves a relatively uniform rate of sending packets.

It should be noted that there is no sequence relationship between the step 408 and the step 401 to the step 407, which is not limited herein.

409, the TCP client sets the sending timer and starts the sending timer, when the sending timer reaches the preset duration, step 410 is performed;

In the embodiment of the present invention, the TCP client may set a sending timer and start the timer. When the sending timer reaches the preset duration, step 410 is performed.

In the embodiment of the present invention, when the TCP client determines that the number of tokens obtained according to the number of tokens has not reached the preset value, the packets to be sent are buffered in the socket buffer of the TCP client side and are waiting to be sent. At this time, it is necessary to wait for the arrival of the instruction for triggering the transmission of the message, that is, the TCP client waits for the ACK that is sent back when the last transmitted message has been sent successfully, and the TCP client will perform the determination again according to the number of tokens. Whether the number of tokens obtained reaches the pre- The step of setting the value, however, due to network failure or other communication reasons, the RTT of the ACK is often too long. In this case, even if the number of tokens has reached the preset value, the data in the socket buffer will be long. In the embodiment of the present invention, the TCP client sets a sending timer. As long as the sending timer reaches the preset duration, the current number of tokens is determined. The step of increasing the number of tokens for increasing the rate reaches a preset value, that is, the situation in which the data in the socket buffer cannot be sent for a long time can be effectively avoided.

It can be understood that step 409 can be executed after step 407, which is not limited herein.

410, the TCP client determines whether the number of tokens obtained according to the number of token growth rate reaches a preset value, if yes, step 411;

The TCP client stores the packet to be sent to the TCP client side socket buffer through the socket socket, and waits for the TCP client to determine whether the number of tokens in the token resource queue reaches the preset value. If it is determined that it is reached, step 411 is performed. It can be understood that the specific value of the preset value is not limited herein.

411. The TCP client sends a packet.

In the embodiment of the present invention, when the TCP client determines that the number of tokens currently increasing according to the number of tokens reaches a preset value, the TCP client sends a packet.

The TCP client stores the eight packets that need to be forwarded to the TCP client to the TCP client side socket buffer for transmission by the TCP client. In this case, if the TCP client detects the number of tokens in the token resource queue is 5 The TCP client extracts three packets from the socket buffer and sends the three packets to the TCP proxy. It should be noted that the TCP client can also take out four packets from the buffer, or other The number of packets is sent to the TCP proxy, which is not limited here. It should be noted that, after the above-mentioned transmitted message is successfully sent, according to the TCP/IP protocol, the TCP proxy feeds back an ACK to the TCP client, and when the TCP client receives the ACK, it re-executes the determination of the token resource queue. If the number of the tokens reaches the preset value, if the preset value is reached, and the packet in the socket buffer is not sent, the packet in the socket buffer is sent, which is not described here.

If the TCP client detects that the number of tokens in the token resource queue is less than three, the above 8 The message continues to be stored in the socket buffer waiting to be sent.

In addition, it should be noted that whenever the TCP client stores the packet to be sent to the TCP proxy through the socket socket to the TCP client side socket buffer, the TCP client determines the token in the token resource queue. The step of whether the number reaches the preset value.

Sending M packets to the TCP client through the TCP proxy, receiving N TCP response packets of the M packets sent by the TCP client, and calculating the TCP between the TCP proxy and the TCP client according to each sample TCP response packet. The sample transmission bandwidth of the connection, the sample TCP response message is at least one of N TCP response messages, and the estimated bandwidth sample set is determined according to the sample transmission bandwidth, and the estimated bandwidth sample set includes the sample calculated according to each sample TCP response message. The transmission bandwidth is determined according to the estimated bandwidth sample set, and the target bandwidth is sent to the TCP client, so that the TCP client sets the token growth rate according to the target bandwidth, and determines whether the number of tokens obtained according to the token growth rate reaches the pre-predetermined rate. Set the value, if it is reached, send a message. Compared with the prior art, when a packet is sent, the packet transmission is limited to the number of tokens, that is, the packet transmission rate is controlled, thereby effectively preventing the network from entering too many packets in an instant and improving the TCP throughput. .

A packet sending method in the embodiment of the present invention has been described above. The following describes the TCP proxy in the embodiment of the present invention.

Referring to FIG. 5, an embodiment of the TCP proxy of the present invention includes: a sending module 501, a receiving module 502, and a processing module 503.

The sending module 501 is configured to send M messages to the TCP client.

The receiving module 502 is configured to receive N TCP response messages of M messages sent by the TCP client.

The processing module 503 is configured to calculate, according to each sample TCP response message received by the receiving module 502, a sample transmission bandwidth of a TCP connection between the TCP proxy and the TCP client, where the sample TCP response packet is in the N TCP response packets. At least one, determining an estimated bandwidth sample set according to the sample transmission bandwidth, the estimated bandwidth sample set includes a sample transmission bandwidth calculated according to each sample TCP response message, determining an estimated bandwidth according to the estimated bandwidth sample set, and determining a target bandwidth according to the estimated bandwidth, according to The target bandwidth sets the rate of increase of the number of tokens, and determines whether the number of tokens obtained according to the growth rate of the number of tokens reaches a preset value;

The sending module 501 is further configured to: if the processing module 503 determines that the rate is increased according to the number of tokens If the number of tokens reaches a preset value, the packet is sent to the TCP client;

In combination with the foregoing embodiment, the processing module 503 is further configured to:

Setting a sending timer and starting the sending timer;

When the sending timer reaches the preset duration, it is determined whether the number of tokens obtained according to the token number increasing rate reaches a preset value;

Set the token resource queue and the maximum depth of the token resource queue. The token resource queue is used to store the token obtained according to the growth rate of the token. The maximum depth is used to indicate the maximum number of tokens that can be stored in the token resource queue according to the token growth rate.

For each sample TCP response message, the sample transmission of the TCP connection is calculated according to the transmission time of the packet corresponding to the sample TCP response message, the reception time of the sample TCP response message, and the length of the packet corresponding to the sample TCP response message. bandwidth;

Obtaining policy configuration information, where the policy configuration information includes a correspondence between a target bandwidth adjustment ratio and a service type;

The target bandwidth is determined according to the estimated bandwidth, the service type corresponding to the TCP connection, and the policy configuration information.

The TCP proxy in the embodiment of the present invention is described above from the perspective of a modular functional entity. The following describes the TCP proxy in the embodiment of the present invention from the perspective of hardware processing. Referring to FIG. 6, a TCP provided by the embodiment of the present invention is provided. Schematic diagram of the proxy structure, the TCP proxy 600 may be relatively different due to different configurations or performances, such as may include one or more physical device processors (eg, one or more processors) corresponding to the processing module 503 of the above embodiment. The physical device transmitter 602 corresponding to the sending module 501 in the above embodiment, the physical device receiver 603 and the memory 604 corresponding to the receiving module 502, one or more data 608 or a storage medium 605 storing the program code 609 (for example, one or A storage device in Shanghai). Among them, the memory 604 and the storage medium 605 may be short-term storage or persistent storage. The program stored on storage medium 605 can include one or more modules (not shown), each of which can include a series of instruction operations in a TCP proxy. Still further, central processor 601 can be configured to communicate with storage medium 605 to perform a series of instruction operations in storage medium 605 on TCP proxy 600.

TCP proxy 600 may also include one or more power sources 606, one or more operating systems 607.

The steps performed by the TCP proxy in this embodiment may be based on the schematic diagram of the TCP proxy structure shown in FIG. 6. For details, refer to the corresponding process in the foregoing embodiment, and details are not described herein again.

Referring to FIG. 7, another embodiment of a TCP proxy according to an embodiment of the present invention includes: a sending module 701 receiving a module 702 and a processing module 703.

The sending module 701 is configured to send M messages to the TCP client.

The receiving module 702 is configured to receive N TCP response packets of the M packets sent by the TCP client.

The processing module 703 is configured to calculate, according to each sample TCP response message received by the receiving module 702, a sample transmission bandwidth of a TCP connection between the TCP proxy and the TCP client, where the sample TCP response packet is Determining, by the at least one of the N TCP response messages, an estimated bandwidth sample set according to the sample transmission bandwidth, where the estimated bandwidth sample set includes a sample transmission bandwidth calculated according to each of the sample TCP response messages, according to The estimated bandwidth sample set determines an estimated bandwidth, and the target bandwidth is determined according to the estimated bandwidth;

The sending module 701 is further configured to send the target bandwidth to the TCP client, so that the TCP client sets the token growth rate according to the target bandwidth, and determines whether the number of tokens obtained according to the token number increase rate reaches a preset value, and if so, Then send a message.

For the same reason, the corresponding physical device in this embodiment can refer to the structural diagram shown in FIG. 6. For the corresponding steps, refer to the foregoing embodiment, and details are not described herein again.

The TCP proxy in the embodiment of the present invention is described above. The following describes the TCP client in the embodiment of the present invention. Referring to FIG. 8, an embodiment of the TCP client in the embodiment of the present invention includes: a receiving module 801, and a processing module. 802 and a transmitting module 803.

The receiving module 801 is configured to receive a target bandwidth sent by the TCP proxy, where the target bandwidth is determined by the TCP proxy according to the estimated bandwidth, and the estimated bandwidth is determined by the TCP proxy according to the estimated bandwidth sample set, and the estimated bandwidth sample set is determined by the TCP proxy according to the sample transmission bandwidth, and the sample is determined. The transmission bandwidth is obtained by the TCP proxy to calculate the TCP connection between the TCP proxy and the TCP client according to each sample TCP response message. The sample TCP response packet is at least one of N TCP response packets, and N TCP response packets. The message that the TCP client receives back to the TCP proxy after receiving the M packets sent by the TCP proxy;

The processing module 802 is configured to set a token growth rate according to the target bandwidth, and determine whether the number of tokens obtained according to the token growth rate reaches a preset value;

The sending module 803 is configured to send a message if the processing module 802 determines that the number of tokens obtained according to the number of tokens increases to a preset value.

In combination with the foregoing embodiment, the processing module 802 is further configured to:

Set the send timer and start the send timer;

The TCP client in the embodiment of the present invention is described above from the perspective of a modular functional entity. The TCP client in the embodiment of the present invention is described from the perspective of hardware processing, as shown in FIG. 9 , for convenience of description, only For a part related to the embodiment of the present invention, the specific technical details are not disclosed. Please refer to the method part corresponding to the embodiment of the present invention. Referring to FIG. 9, a schematic diagram of a TCP client structure is provided in the embodiment of the present invention:

The TCP client includes: a physical device corresponding to the receiving module 801 and the sending module 803 of the foregoing embodiment, and a physical device processor 902 corresponding to the receiver 901 and the processing module 802. The TCP client further includes a memory 904. For storing program code, when the program code is executed by the processor 902, the methods of the above-described embodiments of the present invention can be implemented. It will be understood by those skilled in the art that the TCP client structure shown in FIG. 9 does not constitute a limitation to the TCP client, and may include more or less components than those illustrated, or combine some components, or different component arrangements. .

The steps performed by the TCP client in this embodiment may be based on the structure of the TCP client shown in FIG. 9. For details, refer to the corresponding process in the foregoing embodiment, and details are not described herein again.

It should be noted that the processor involved in the foregoing device embodiment may be a central processing unit (English name: central processing unit, English abbreviation: CPU), a network processor (English full name: network processor, English abbreviation: NP) or A combination of CPU and NP. The processor may further include a hardware chip, which may be an application-specific integrated circuit (English name: ASIC), a programmable logic device (English full name: programmable logic device, English abbreviation: PLD) or combination. In addition, PLD can be complex programmable logic Device (English full name: complex programmable logic device, English abbreviation: CPLD), field programmable logic gate array (English full name: field-programmable gate array, English abbreviation: FPGA), general array logic (English full name: generic array logic, English The abbreviation: GAL) or any combination thereof is not limited in the present invention.

In the several embodiments provided by the present application, it should be understood that the disclosed systems, modules, and methods may be implemented in other manners. For example, the device embodiments described above are merely illustrative. For example, the division of the modules is only a logical function division. In actual implementation, there may be another division manner, for example, multiple units or components may be combined or Can be integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interface, device or unit, and may be in an electrical, mechanical or other form.

The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

The integrated modules, when implemented in the form of software functional units and sold or used as separate products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention, which is essential or contributes to the prior art, or all or part of the technical solution, may be embodied in the form of a software product stored in a storage medium. A number of instructions are included to cause a computer device (which may be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: a U disk, a mobile hard disk, a read only memory (English full name: Read-Only Memory, English abbreviation: ROM), a random access memory (English full name: Random Access Memory English: RAM), a disk Or a variety of media such as optical discs that can store program code.

The above embodiments are only used to illustrate the technical solutions of the present invention, and are not limited thereto; Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or equivalently replace some of the technical features. Modifications or substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the invention.

Claims

A packet sending method, characterized in that the method is performed by a TCP proxy, the method comprising:

Send M messages to the TCP client;

Receiving N TCP response messages of the M packets sent by the TCP client;

Calculating, according to each sample TCP response message, a sample transmission bandwidth of the TCP connection between the TCP proxy and the TCP client, where the sample TCP response packet is at least one of the N TCP response packets;

Determining an estimated bandwidth sample set according to the sample transmission bandwidth, where the estimated bandwidth sample set includes a sample transmission bandwidth calculated according to each of the sample TCP response messages;

Determining an estimated bandwidth according to the estimated bandwidth sample set;

Determining a target bandwidth according to the estimated bandwidth;

Determining the number of available token tokens according to the target bandwidth, and controlling transmission of the packet according to the available token number.
The packet sending method according to claim 1, wherein the determining the number of available token tokens according to the target bandwidth, and controlling the sending of the packet according to the available token number comprises:

Setting a token growth rate according to the target bandwidth;

Determining whether the number of tokens obtained according to the rate of increase of the number of tokens reaches a preset value, and if so, sending the message to the TCP client.
The packet sending method according to claim 1, wherein the determining the number of available token tokens according to the target bandwidth, and controlling the sending of the packet according to the available token number comprises:

Sending the target bandwidth to the TCP client, so that the TCP client sets a token growth rate according to the target bandwidth, and determines whether the number of tokens obtained according to the token number increase rate reaches a preset value. If yes, the message is sent.
The method for transmitting a message according to claim 2, wherein the method further comprises:

Setting a sending timer and starting the sending timer;

When the sending timer reaches the preset duration, the step of determining whether the number of tokens obtained according to the token number increasing rate reaches a preset value is performed.
The method for transmitting a message according to claim 4, wherein the method further comprises:

Setting a token resource queue and a maximum depth of the token resource queue, the token resource The source queue is configured to store a token obtained according to the rate of increase of the number of tokens, where the maximum depth is used to indicate the maximum number of tokens that can be stored in the token resource queue according to the growth rate of the number of tokens.
The packet transmission method according to claim 5, wherein the calculating a sample transmission bandwidth of the TCP connection between the TCP proxy and the TCP client according to each sample TCP response message comprises:

For each sample TCP response message, according to the sending time of the packet corresponding to the sample TCP response message, the receiving time of the sample TCP response message, and the length of the packet corresponding to the sample TCP response message. Calculating the sample transmission bandwidth of the TCP connection.
The packet sending method according to claim 6, wherein the determining, by the TCP proxy, the target bandwidth according to the estimated bandwidth, further comprises:

Obtaining policy configuration information, where the policy configuration information includes a correspondence between the adjustment ratio of the target bandwidth and a service type;

The determining the target bandwidth according to the estimated bandwidth includes:

Determining the target bandwidth according to the estimated bandwidth, a service type corresponding to the TCP connection, and the policy configuration information.
A message sending method, comprising:

The TCP client receives a target bandwidth sent by the TCP proxy, the target bandwidth being determined by the TCP proxy based on the estimated bandwidth, the estimated bandwidth being determined by the TCP proxy according to the estimated bandwidth sample set, the estimated bandwidth sample set being The TCP proxy determines, according to the sample transmission bandwidth, that the TCP transmission bandwidth is obtained by the TCP proxy according to each sample TCP response message, and the TCP connection between the TCP proxy and the TCP client is obtained, and the sample TCP response report is obtained. The message is at least one of the N TCP response messages, and the N TCP response messages are the messages that the TCP client feeds back to the TCP proxy after receiving the M messages sent by the TCP proxy;

Setting, by the TCP client, a token growth rate according to the target bandwidth;

The TCP client determines whether the number of tokens obtained according to the growth rate of the number of tokens reaches a preset value, and if so, sends a message.
The method for transmitting a message according to claim 8, wherein the method further comprises:

Setting, by the TCP client, a sending timer and starting the sending timer;

When the sending timer reaches the preset duration, the step of determining whether the number of tokens obtained according to the token number increasing rate reaches a preset value is performed.
The method for transmitting a message according to any one of claims 8 to 9, wherein the method further comprises:

The TCP client sets a token resource queue and a maximum depth of the token resource queue, the token resource queue is used to store the token, and the maximum depth is used to indicate that the token resource queue can be stored by the token. The maximum number.
A TCP proxy, comprising:

a sending module, configured to send M packets to the TCP client;

a receiving module, configured to receive N TCP response messages of the M packets sent by the TCP client;

a processing module, configured to calculate, according to each sample TCP response message received by the receiving module, a sample transmission bandwidth of a TCP connection between the TCP proxy and the TCP client, where the sample TCP response packet is Determining, according to the sample transmission bandwidth, an estimated bandwidth sample set, where the estimated bandwidth sample set includes a sample transmission bandwidth calculated according to each of the sample TCP response messages, according to the at least one of the N TCP response messages, The estimated bandwidth sample set determines an estimated bandwidth, determines a target bandwidth according to the estimated bandwidth, determines an available token token number according to the target bandwidth, and controls transmission of the packet according to the available token number.
The TCP proxy according to claim 11, wherein the processing module is configured to:

Setting a token growth rate according to the target bandwidth;

Determining whether the number of tokens obtained according to the growth rate of the number of tokens reaches a preset value;

The sending module is configured to send the message to the TCP client if the processing module determines that the number of tokens obtained according to the rate of increase of the token reaches a preset value.
The TCP proxy according to claim 11, wherein the sending module is configured to:

Sending the target bandwidth to the TCP client, so that the TCP client sets a token growth rate according to the target bandwidth, and determines whether the number of tokens obtained according to the token number increase rate reaches a preset value. If yes, the message is sent.
The TCP proxy according to claim 12, wherein the processing module is configured to:

Setting a sending timer and starting the sending timer;

When the sending timer reaches the preset duration, it is determined whether the number of tokens obtained according to the token number increasing rate reaches a preset value.
The TCP proxy according to claim 14, wherein the processing module is configured to:

Setting a token resource queue and a maximum depth of the token resource queue, where the token resource queue is used to store a token obtained according to the growth rate of the token number, where the maximum depth is used to indicate that the token resource queue can be stored according to the The maximum number of tokens obtained by increasing the number of tokens.
The TCP proxy according to claim 15, wherein said processing module is configured to:

For each sample TCP response message, according to the sending time of the packet corresponding to the sample TCP response message, the receiving time of the sample TCP response message, and the length of the packet corresponding to the sample TCP response message. Calculating the sample transmission bandwidth of the TCP connection.
The TCP proxy according to claim 16, wherein the processing module is configured to:

Obtaining policy configuration information, where the policy configuration information includes a correspondence between the adjustment ratio of the target bandwidth and a service type;

Determining the target bandwidth according to the estimated bandwidth, a service type corresponding to the TCP connection, and the policy configuration information.
A TCP client, comprising:

a receiving module, configured to receive a target bandwidth sent by a TCP proxy, where the target bandwidth is determined by the TCP proxy according to an estimated bandwidth, where the estimated bandwidth is determined by the TCP proxy according to an estimated bandwidth sample set, where the estimated bandwidth sample set is determined by Determining, by the TCP proxy, that the sample transmission bandwidth is obtained by the TCP proxy according to each sample TCP response message, and calculating a TCP connection between the TCP proxy and the TCP client, where the sample TCP is obtained. The response message is at least one of the N TCP response messages, and the N TCP response messages are sent back to the TCP proxy by the TCP client after receiving the M messages sent by the TCP proxy. Text

a processing module, configured to set a token growth rate according to the target bandwidth, and determine whether the number of tokens obtained according to the growth rate of the number of tokens reaches a preset value;

a sending module, if the processing module determines that the rate is increased according to the number of tokens The number of tokens reaches a preset value, and the message is sent.
The TCP client according to claim 18, wherein the processing module is configured to:

Setting a sending timer and starting the sending timer;

When the sending timer reaches the preset duration, it is determined whether the number of tokens obtained according to the token number increasing rate reaches a preset value.
The TCP client according to claim 18 or 19, wherein the processing module is configured to:

Setting a token resource queue and a maximum depth of the token resource queue, where the token resource queue is used to store a token obtained according to the growth rate of the token number, where the maximum depth is used to indicate that the token resource queue can be stored according to the The maximum number of tokens obtained by increasing the number of tokens.