CN112217605A

CN112217605A - Transmission strategy selection method, device and storage medium

Info

Publication number: CN112217605A
Application number: CN202010928114.8A
Authority: CN
Inventors: 王昱丹; 陈文娟; 陈文生
Original assignee: Wangsu Science and Technology Co Ltd
Current assignee: Wangsu Science and Technology Co Ltd
Priority date: 2020-09-07
Filing date: 2020-09-07
Publication date: 2021-01-12
Anticipated expiration: 2040-09-07
Also published as: CN112217605B

Abstract

The invention discloses a method, equipment and a storage medium for selecting a transmission strategy. In the invention, the mapping table of the IP section and the network system is constructed based on the principle that the IP section and the network system from the same base station correspond to each other, before selecting a transmission strategy, searching a preset network system corresponding to IP section information provided by a client in a preset IP section and network system mapping table, and then the searched network system is used as the target type of the network for TCP service transmission with the client terminal providing the IP section information, namely the target network standard, effectively ensures the accuracy of the determined target network standard, finally selects the adaptive target transmission strategy in a targeted manner according to the actually determined target network standard, so that the transmission strategy according to the transmission TCP transmission service can take account of the difference in transmission of different network systems, therefore, the influence of different network systems on the transmission quality is avoided, and the improvement of the total transmission quality of a TCP transmission layer is effectively promoted.

Description

Transmission strategy selection method, device and storage medium

Technical Field

The embodiment of the invention relates to the technical field of computer networks, in particular to a method, equipment and a storage medium for selecting a transmission strategy.

Background

At present, domestic network systems are divided into a wireless network WIFI and a MOBILE network MOBILE (including 2G, 3G, 4G and 5G), and at present, the WIFI and 4G are the most common. WIFI is mainly applicable to indoors, 4G is more applicable to outdoors, and to the user, WIFI's transmission quality mainly receives the distance between equipment and the router to and router performance influences, 4G's transmission quality receives the degree of congestion of basic station, equipment performance and the signal strength that equipment can receive influences. Under most conditions, WIFI internet surfing is stable, and transmission quality is relatively good 4G. For CDN manufacturers, the acceleration service makes a transmission policy in a targeted manner mainly according to the service characteristics and the evaluation index, for example, when the on-demand service evaluation index is the stuck rate, in order to reduce video playing stuck, the transmission policy needs to be configured to send packets at a faster and more stable speed, but the transmission policy causes the transmission service to be easily congested, and even causes packet loss; on the contrary, when the evaluation index is the packet loss rate, the transmission policy needs to be configured to reduce the packet loss as much as possible.

However, at present, a fixed transmission strategy is usually selected for the acceleration service of the same domain name, so that it is difficult to give consideration to the difference in transmission of different network systems in terms of strategy, and the influence of different network systems on quality is ignored, thereby resulting in that the transmission quality cannot be sufficiently improved.

Disclosure of Invention

Embodiments of the present invention provide a method, an apparatus, and a storage medium for selecting a transmission policy, which are used to solve the above technical problems.

To solve the above technical problem, an embodiment of the present invention provides a method for selecting a transmission policy, including the following steps:

acquiring IP section information of a client;

searching a preset network system corresponding to the IP section information in a preset IP section and network system mapping table;

taking the searched preset network standard as a target network standard;

and selecting one transmission strategy from preset transmission strategies as a target transmission strategy according to the target network standard so as to execute Transmission Control Protocol (TCP) transmission service based on the target transmission strategy.

An embodiment of the present invention further provides a device for selecting a transmission policy, including:

at least one processor; and the number of the first and second groups,

a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform a method of selecting a transmission strategy as described above.

Embodiments of the present invention also provide a computer-readable storage medium storing a computer program which, when executed by a processor, implements a method of selecting a transmission policy as described above.

Compared with the prior art, the embodiment of the invention constructs the mapping table of the IP section and the network system based on the principle that the IP section and the network system from the same base station correspond to each other, before selecting a transmission strategy, searching a preset network system corresponding to IP section information provided by a client in a preset IP section and network system mapping table, and then the searched network system is used as the target type of the network for TCP service transmission with the client terminal providing the IP section information, namely the target network standard, effectively ensures the accuracy of the determined target network standard, finally selects the adaptive target transmission strategy in a targeted manner according to the actually determined target network standard, so that the transmission strategy according to the transmission TCP transmission service can take account of the difference in transmission of different network systems, therefore, the influence of different network systems on the transmission quality is avoided, and the improvement of the total transmission quality of a TCP transmission layer is effectively promoted.

In addition, after the preset network system corresponding to the IP segment information is searched in the preset IP segment and network system mapping table, the method further includes: if a preset network system corresponding to the IP section information is found in a preset IP section and network system mapping table, executing the step of taking the found preset network system as a target network system; otherwise, acquiring characteristic information of a TCP transmission layer, analyzing and processing the characteristic information by using a preset network system determining model, determining a network system corresponding to the characteristic information, taking the determined network system group as the target network system, and executing the step of selecting one transmission strategy from preset transmission strategies as a target transmission strategy according to the target network system. When the preset network system corresponding to the IP section information provided by the client is not found in the preset IP section and network system mapping table, the characteristic information of the TCP transmission layer is analyzed and processed through a network system determination model obtained based on pre-training, and then the network system is determined, so that the defects of the IP section and network system mapping table are overcome.

In addition, after the analyzing and processing are performed on the feature information by using a preset network system determining model and a network system corresponding to the feature information is determined, the method further includes: and establishing a corresponding relation between the IP section information and the network system, and adding the corresponding relation into a mapping table of the IP section and the network system. Through updating the preset IP section and network system mapping table according to the actual situation, the information recorded in the IP section and network system mapping table can be richer, and more network systems suitable for different clients can be covered.

In addition, the training of the network system determination model includes: acquiring characteristic information of a TCP transmission layer corresponding to a known network system and a network system type mark of the known network system to obtain a training sample; constructing a network standard training model according to a preset machine learning algorithm; and performing iterative training on the network standard training model by using the training samples until a preset iteration termination condition is met, so as to obtain the network standard determination model.

In addition, the obtaining of the characteristic information of the TCP transport layer corresponding to the known network system includes: and acquiring the maximum message segment length received by a TCP transmission layer and the initial receiving window size of a client during the period from the beginning of TCP connection establishment to the beginning of packet sending, and acquiring the characteristic information of the TCP transmission layer corresponding to the known network system. The maximum message segment length irrelevant to the area coverage and the size of the initial receiving window of the client are selected as the characteristic information of the TCP transmission layer corresponding to the known network system, so that the finally trained network system determination model has certain generalization while considering the accuracy rate, namely one network system determination model can be suitable for TCP transmission services of the client and the server in different areas.

In addition, the obtaining of the maximum segment length received by the TCP transport layer and the initial receiving window size of the client during the period from the beginning of the TCP connection establishment to the beginning of the packet transmission includes: and acquiring the maximum message segment length and the initial receiving window size of the client carried in the first TCP request received by the TCP transmission layer during the period from the connection establishment of the TCP to the packet sending start, and acquiring the characteristic information of the TCP transmission layer corresponding to the known network system. For long connection, only the first TCP request has a value, so that when a network system determination model is trained, only the maximum segment length carried in the first TCP request and the size of the initial receiving window of the client are obtained as the characteristic information of the TCP transmission layer corresponding to the known network system, and the influence of multiple accesses of a single client IP can be avoided.

In addition, the acquiring the network standard type label of the known network standard includes: acquiring a log or URL reported by a client; and acquiring the network type mark of the known network type from the log or the URL.

In addition, the acquiring the network standard type label of the known network standard from the log or the URL includes: analyzing the log or the URL to obtain an analysis result; judging whether the analysis result comprises network type information or not; and if so, acquiring the network type mark of the known network type from the network type information.

In addition, after the determining whether the analysis result includes the network type information, the method further includes: if the analysis result does not include the network standard information, a test environment is established, manual marking is carried out on the known network standard based on the test environment packet capturing identification, and the network standard type mark of the known network standard is obtained.

In addition, the construction of the mapping table of the IP section and the network system comprises the following steps: analyzing and processing sample characteristic information in a sample characteristic set by using the network system determining model, and determining a network system corresponding to each sample characteristic information, wherein the sample characteristic information is pre-collected characteristic information of a TCP (transmission control protocol) transmission layer provided by clients of different IPs in the same IP section; counting the determined IP number corresponding to different network systems; selecting a network standard of which the IP number meets a preset condition as a preset network standard corresponding to the IP section; and establishing a corresponding relation between the IP section and the preset network system to obtain a mapping table of the IP section and the network system.

Drawings

One or more embodiments are illustrated by way of example in the accompanying drawings, which correspond to the figures in which like reference numerals refer to similar elements and which are not to scale unless otherwise specified.

Fig. 1 is a detailed flowchart of a selection method of a transmission policy according to a first embodiment of the present invention;

fig. 2 is a detailed flowchart of a method for selecting a transmission policy according to a second embodiment of the present invention;

fig. 3 is a schematic structural diagram of a selection means of a transmission strategy according to a third embodiment of the present invention;

fig. 4 is a schematic structural diagram of a transmission policy selecting device according to a fourth embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the embodiments of the present invention will be described in detail with reference to the accompanying drawings. However, it will be appreciated by those of ordinary skill in the art that in various embodiments of the invention, numerous technical details are set forth in order to provide a better understanding of the present application. However, the technical solution claimed in the present application can be implemented without these technical details and various changes and modifications based on the following embodiments. The following embodiments are divided for convenience of description, and should not constitute any limitation to the specific implementation manner of the present invention, and the embodiments may be mutually incorporated and referred to without contradiction.

For convenience of description, the present embodiment takes the selection device of the transmission policy as an example for the sake of description.

Implementation details of the selection method of the transmission policy of the present embodiment are described below, and the following description is provided only for convenience of understanding and is not necessary for implementing the present embodiment.

The specific flow of this embodiment is shown in fig. 1, and specifically includes the following steps:

step 101, acquiring IP section information of a client.

Specifically, the IP address is allocated by a Network Interface Controller (NIC) according to the Network size of the user unit and the recent development plan. Conceptually, each IP address can be composed of two parts, namely a network identity netid and a host identity hostid. In fact, IP addresses can be classified into A, B, C categories.

That is, in a general case, the acquired IP segment information may include class a IP segment information (IP _ a segment information), class B IP segment information (IP _ B segment information), and class C IP segment information (IP _ C segment information).

In general, IP _ a segment addresses are allocated to a small number of large-scale networks, IP _ B segment addresses are allocated to medium-scale networks, and IP _ C segment addresses are allocated to small-scale networks. Therefore, in actual use, whether the acquired IP segment information is IP _ a segment information, IP _ B segment information, or IP _ C segment information may be set according to the size of the network scale.

For convenience of description, in this embodiment, the IP segment information acquired from the client is taken as IP _ C segment information as an example, that is, the IP segment information used for determining the target network type is IP _ C segment information.

In addition, it is worth mentioning that, in this embodiment, the IP segment information is selected to determine the network type, because the IP segment information coming out from the same base station, whether it is the IP _ a segment information, the IP _ B segment information, or the IP _ C segment information, is fixed, and because in practical application, the number of base stations is limited, the target network type of the network where the client is currently located can be quickly determined by means of table lookup based on the limited number of fixed IP segment information.

And 102, searching a preset network system corresponding to the IP section information in a preset IP section and network system mapping table.

As can be seen from the above description, the IP segment information selected in this embodiment is IP _ C segment information, and thus the preset IP segment and the network system mapping table also need to be a mapping table established based on the IP _ C segment information and the corresponding network system.

Correspondingly, when the step 102 is executed, specifically, the preset network system corresponding to the IP _ C segment information acquired from the client in the step 101 is searched in the preset IP _ C segment information and network system mapping table.

It should be understood that the above examples are only examples for better understanding of the technical solution of the present embodiment, and are not to be taken as the only limitation to the present embodiment.

In addition, in order to facilitate understanding of the technical solution provided by this embodiment, a technical name of "network type" is first described herein.

The "network system" refers to the type of network. At present, network systems are generally divided into two categories, one is wireless network WIFI, and the other is MOBILE network MOBILE. The MOBILE can be divided into 2G, 3G, 4G and 5G. Since 5G is not yet fully popularized, and 2G and 3G have been substantially replaced by 4G, the MOBILE in the present embodiment is mainly 4G.

Accordingly, the preset network type found in step 102 is either WIFI, i.e. a wireless network type, or 4G, i.e. a mobile network type.

In addition, it should be understood that, in practical applications, in order to ensure the implementation of the above steps, the above IP segment and network system mapping table need to be constructed before the step 102 is executed.

For convenience of understanding, this embodiment provides a specific way for constructing a mapping table between an IP segment and a network system, which is as follows:

(1) and analyzing and processing the sample characteristic information in the sample characteristic set by using a preset network system determining model, and determining the network system corresponding to each sample characteristic information.

Specifically, the sample feature information is feature information of a TCP transport layer corresponding to a client request of different IPs in the same IP segment collected in advance from a server.

(2) And counting the determined IP number corresponding to different network systems.

For ease of understanding, the following description is made with reference to the IP segment and network format mapping statistical table given in table 1:

TABLE 1IP section and network system mapping statistical table

IP section	IP number of MOBILE	IP number of WIFI
			IP section 1	83	2
IP section 2	101	7
			IP section 3	91	6
IP section 4	2	151
			IP section 5	1	169
IP section 6	1	157

As can be seen from the contents recorded in table 1, for IP segment 1, 83 network systems corresponding to the IP are MOBILE, and 2 network systems corresponding to the IP are WIFI; for the IP section 2, 101 network systems corresponding to the IP are MOBILE, and 7 network systems corresponding to the IP are WIFI; for the IP section 3, the network systems corresponding to 91 IPs are MOBILE, and the network systems corresponding to 6 IPs are WIFI; for IP section 4, the network system corresponding to 2 IPs is MOBILE, and the network system corresponding to 151 IPs is WIFI; for the IP section 5, the network system corresponding to 1IP is MOBILE, and the network system corresponding to 169 IPs is WIFI; for the IP segment 6, the network system corresponding to 1IP is MOBILE, and the network system corresponding to 157 IPs is WIFI.

(3) And selecting the network system of which the IP number meets the preset condition as the preset network system corresponding to the IP section.

Specifically, the above-mentioned meeting the preset condition can be divided into the following two cases in practical application:

case 1: the two network systems, namely the MOBILE and the WIFI, correspond to the IP number, and at this time, the condition that the preset condition is met can be that the preset condition is larger than a certain preset threshold value, or the network system with the largest IP number is selected as the preset network system corresponding to the IP section.

Case 2: the two network systems of MOBILE and WIFI have only one corresponding IP number, and the meeting of the preset condition can be selecting the network system with the corresponding IP number as the preset network system corresponding to the IP section.

(4) And establishing a corresponding relation between the IP section and the preset network system to obtain a mapping table of the IP section and the network system.

For convenience of understanding, in this embodiment, the network system meeting the preset condition is set to select the network system with the largest number of IPs as the preset network system corresponding to the IP segment, and based on table 1, the finally obtained mapping table of the IP segment and the network system is shown in table 2.

TABLE 2IP section and network system mapping table

IP section	Default network system
		IP section 1	MOBILE
IP section 2	MOBILE
		IP section 3	MOBILE
IP section 4	WIFI
		IP section 5	WIFI
IP section 6	WIFI

In addition, it should be noted that, in practical applications, a manner of combining table 1 and table 2 into one table, such as the style shown in table 3, may also be adopted when constructing the IP segment and network system mapping table.

TABLE 3IP section and network system mapping table

IP section	IP number of MOBILE	IP number of WIFI	Default network system
				IP section 1	83	2	MOBILE
IP section 2	101	7	MOBILE
				IP section 3	91	6	MOBILE
IP section 4	2	151	WIFI
				IP section 5	1	169	WIFI
IP section 6	1	157	WIFI

And 103, taking the searched preset network system as a target network system.

And 104, selecting one transmission strategy from preset transmission strategies as a target transmission strategy according to the target network standard, and executing Transmission Control Protocol (TCP) transmission service based on the target transmission strategy.

It should be understood that the RTT of the WIFI type network system is more stable than the RTT of the MOBILE (this embodiment refers to 4G specifically) type network system, and the RTT values are usually smaller and less variable. Therefore, aiming at the blockage rate index, when the target network type is WIFI, the selected target transmission strategy can be used for detecting retransmission in time when packet loss occurs, so that blockage is reduced; for a MOBILE station with unstable RTT and easy degradation, which is likely to cause jamming when packet loss occurs in a short time, the RTT degradation needs to be controlled as much as possible in the selected target transmission strategy, so as to reduce the influence on transmission efficiency and further improve transmission quality.

The target transmission policy applicable to different target network systems may be preset by a technician according to an actual situation, which is not limited in this embodiment.

It is not difficult to find out through the above description that the selection method of the transmission policy provided in this embodiment constructs a mapping table of IP segments and network systems based on the principle that IP segments and network systems from the same base station correspond to each other, and before selecting the transmission policy, searches a preset network system corresponding to IP segment information provided by a client in the preset mapping table of IP segments and network systems, and further uses the found network system as a target type of a network for TCP service transmission with the client providing the IP segment information, i.e. a target network system, thereby effectively ensuring the accuracy of the determined target network system, and finally selects an adaptive target transmission policy in a targeted manner according to the actually determined target network system, so that the transmission policy according to which the TCP service is transmitted can take account of differences in transmission of different network systems, thereby avoiding the influence of different network systems on transmission quality, thereby effectively promoting the improvement of the overall transmission quality of the TCP transmission layer.

A second embodiment of the present invention relates to a method of selecting a transmission policy. The second embodiment is further improved on the basis of the first embodiment, and the main improvements are as follows: and when the preset network system corresponding to the obtained IP section information is not found in the preset IP section and network system mapping table, analyzing and processing the obtained characteristic information of the TCP transmission layer by using a preset network system determination model, and further determining a target network system.

As shown in fig. 2, a method for selecting a transmission policy according to a second embodiment includes the following steps:

step 201, acquiring the IP segment information of the client.

It is to be understood that step 201 in this embodiment is substantially the same as step 101 in the first embodiment, and will not be described herein again.

Step 202, judging whether a preset network system corresponding to the IP segment information is found in a preset IP segment and network system mapping table.

Specifically, if the preset network system corresponding to the IP segment information is found in the preset IP segment and network system mapping table, step 204 is directly executed; otherwise, step 203 is executed.

Step 203, obtaining characteristic information of a TCP transmission layer, analyzing and processing the characteristic information by using a preset network system determination model, determining a network system corresponding to the characteristic information, and taking the determined network system group as the target network system.

Specifically, the characteristic information acquired in this embodiment may be from a TCP connection or from a TCP request.

In addition, the characteristic information of the above-mentioned TCP Layer may include, in practical application, information such as a quadruplet, a URL, a server operator, a Maximum Segment length (MSS), a Time required for TCP to establish a connection (hereinafter referred to as a connection Time est), a Time sst from completion of establishing a connection to receipt of a Secure socket protocol (SSL) handshake packet, a Time set from receipt of an SSL handshake packet to completion of an SSL handshake, a Time reqt from completion of establishing a connection or completion of an SSL handshake to receipt of a request, an initial Round-Trip Time (RTT), and an initial receiving window of a client.

When a network system determination model is obtained through training, training test comparison shows that est, sst, set, reqt and initial RTT of different regions are related to region coverage, the prediction accuracy of a single region and the model trained by the characteristics on the region can reach 95%, but the prediction accuracy is extremely unstable when the model is applied to other regions. Therefore, in consideration of the generalization of the network system determination model, in this embodiment, the training feature value selects the MSS and the initial receiving window size of the client regardless of the area coverage.

That is to say, in this embodiment, the obtained feature information mainly includes the MSS and the initial receiving window size of the client.

Furthermore, it is worth mentioning that in practical applications, TCP connections have long connections and short connections. For the long connection, only the first TCP request carries the above-mentioned feature information, so that in order to avoid the influence of repeated statistics caused by multiple requests of a single client IP, the obtained feature information is specifically the feature information at the initial stage of packet sending.

In addition, it should be understood that, in this case, since the target network standard is determined by the preset network standard determination model based on the acquired feature information, in order to ensure the accuracy of the determined target network standard, the network standard determination model in this embodiment is obtained based on training including the feature information of the same type.

In order to facilitate understanding of the history of the network system determination model, a specific training method is provided in this embodiment, which is specifically as follows:

(1) and acquiring characteristic information of a TCP transmission layer corresponding to a known network system and a network system type mark of the known network system to obtain a training sample.

Specifically, the characteristic information of the TCP transport layer corresponding to the known network system obtained here also needs to at least include the maximum segment length and the size of the client initial receiving window.

In addition, the network standard type label is essentially identification information for identifying the type of the known network standard. In this embodiment, the wireless network type is identified by WIFI, and the MOBILE network type is identified by MOBILE. Namely, the network type is either WIFI or MOBILE.

Specifically, the method includes acquiring a maximum packet segment length and a client initial receiving window size received by the TCP transport layer during a period from when a connection is established to when a packet transmission is started (i.e., a packet transmission initial stage), and taking the received maximum packet segment length and the client initial receiving window size as the characteristic information of the TCP transport layer corresponding to the known network system.

Further, in practical applications, the long connection in the TCP connection carries the above-mentioned characteristic information only when the first TCP request is received, and the output of the subsequent requests is "-". Therefore, in order to avoid being affected by a single client IP, the obtained characteristic information is, specifically, the maximum segment length carried in the first TCP request received by the TCP transport layer and the size of the client initial receiving window during the period from the start of establishing the connection to the start of sending the packet.

Furthermore, it is worth mentioning that, in practical applications, the characteristic information of the TCP transport layer can be refined to various granularities (dimensions of characteristic acquisition), such as the request and connection mentioned above, and can be processed to the granularity of the machine. However, since the network system of the same connection is generally fixed, in this embodiment, in order to ensure that the acquired feature information is a request or a connection, the feature information of the TCP transport layer is acquired at the granularity of the request or the connection when acquiring the feature information of the TCP transport layer.

In addition, in this embodiment, the network type label of the known network type is obtained by obtaining a log or a URL reported by the client and then obtaining the log or the URL.

Further, regarding the operation of acquiring the network type label from the log or the URL, the specific operation is:

firstly, analyzing the log or the URL to obtain an analysis result; and then, judging whether the analysis result comprises network type information.

Correspondingly, if the network standard type mark of the known network standard is included, the network standard type mark of the known network standard is obtained from the network standard information; otherwise, a test environment is built, packet capturing identification is carried out based on the built test environment, then manual marking is carried out on the known network system according to the information identified by the packet capturing identification, namely, the network system is divided into a WIFI type or a MOBILE type according to judgment of technicians.

It should be noted that, in this embodiment, it is determined whether the network type identifier can be obtained by judging whether the analysis result includes the network type information, because the network type information at least includes identification information for identifying a network to which the network type belongs, such as WIFI or MOBILE, in a normal case.

In addition, it should be understood that, in practical applications, in order to ensure the accuracy of the network system determination model obtained by training, the training samples required by training need to be sufficient. Therefore, when the number of the obtained feature information and the number of the network standard type labels are large, in order to ensure the usability of the training samples, the feature information and the network standard type labels which correspond to each other need to be used as one training sample, that is, each training sample is composed of the feature information and the network standard type labels which correspond to each other.

(2) And constructing a network standard training model according to a preset machine learning algorithm.

Specifically, when the number of features in the obtained feature information is relatively small, a simple discriminant algorithm can be used to construct a network system training model, such as a decision tree, logistic regression, naive bayes, support vector machine SVM, and the like, which are not listed here.

(3) And performing iterative training on the network standard training model by using the training samples until a preset iteration termination condition is met, so as to obtain the network standard determination model.

It should be understood that, because the feature information in the training samples corresponds to the network standard, that is, the network standard in each training sample is used to identify the feature information in the training sample, and both the feature information and the network standard are from a known network standard. Therefore, after the training samples are sequentially input into the network standard training model and are analyzed and processed by the network standard training model, the output result is the network standard, and the network standard determination model can be obtained after continuous iterative training until the preset iteration termination condition is met.

It should be understood that, in practical applications, model iterative training does not necessarily ensure that each training sample input into the network format training model has the same network format as the corresponding known network format after being analyzed and processed by the network format training model. And if the network system obtained by all the training samples is the same as the known network system, the model is over-fitted. Therefore, in order to avoid the above problem, in this embodiment, an iteration termination condition is specifically preset, that is, when the preset iteration termination condition is met, the iteration training is stopped, and then the network format training model of the card is used as the finally required network format determination model.

In practical applications, a person skilled in the art may select different algorithms to set the iteration termination condition according to needs, for example, when a machine algorithm used for training a network system determination model is a decision tree algorithm, the preset iteration termination condition may be set according to leaf node purity.

In addition, it is worth mentioning that, in practical application, in order to ensure the accuracy of the network system determination model obtained by training, before the training sample is used to perform iterative training on the network system training model, the training sample may be randomly divided into a training set and a test set, that is, a part of data in the training sample is used as training data, and the other part is used as test data.

Correspondingly, the essence of the training process is that the network standard training model is iteratively trained by using training data, when training is finished, the current network standard training model is tested based on the divided test data before being used as the network standard determination model, and when any test data is input and the output result is the network standard corresponding to the input test data, the training of the current network standard training model can be considered to be finished and the network standard determination model can be used.

It should be understood that the above examples are only examples for better understanding of the technical solution of the present embodiment, and are not to be taken as the only limitation to the present embodiment. In practical application, a person skilled in the art may also use a ten-fold cross validation method to validate whether the network format training model trained by using the training samples can be used as the network format determination model.

By training the obtained network system determination model in the above manner, the accuracy and recall rate of the trained network system determination model can reach more than 80% under the condition of considering the generalization of the model (only the size of the MSS and the initial receiving window of the client is used as the feature information). If the number of models or the number of regions is not considered in practical application, and characteristic information such as est, sst, set, reqt, initial RTT and the like is introduced when the network standard is trained to determine the models, the accuracy and the recall rate of the network standard determination models in different regions can be improved to 95%.

In addition, it is worth mentioning that, in practical application, in order to ensure that the corresponding preset network system can be found in the mapping table of the IP segment and the network system after the IP segment information of the client is received, after the target network system is determined by using the preset network system determination model, the corresponding relationship between the IP segment information and the network system determined by the network system determination model can be established, and the established corresponding relationship is added to the mapping table of the IP segment and the network system, so that the information recorded in the mapping table of the IP segment and the network system can be richer, and more network systems suitable for different clients can be covered.

And 204, taking the searched preset network system as a target network system.

Step 205, according to the target network type, selecting one transmission policy from preset transmission policies as a target transmission policy, so as to execute a transmission control protocol TCP transmission service based on the target transmission policy.

It is to be understood that step 204 and step 205 in this embodiment are substantially the same as step 103 and step 104 in the first embodiment, and are not described herein again.

Therefore, in the method for selecting a transmission policy provided in this embodiment, when the preset network system corresponding to the IP segment information provided by the client is not found in the preset IP segment and network system mapping table, the characteristic information of the TCP transmission layer is analyzed and processed by the network system determination model obtained based on the pre-training, so as to determine the network system, thereby making up for the deficiency of the IP segment and network system mapping table.

It should be understood that the above steps of the various methods are divided for clarity, and the implementation may be combined into one step or split into a plurality of steps, and all that includes the same logical relationship is within the protection scope of the present patent; it is within the scope of the patent to add insignificant modifications to the algorithms or processes or to introduce insignificant design changes to the core design without changing the algorithms or processes.

A third embodiment of the present invention relates to a transmission policy selection apparatus, as shown in fig. 3, including: an obtaining module 301, a searching module 302 and a selecting module 303.

The acquiring module 301 is configured to acquire IP segment information of a client; a determining module 302, configured to search a preset network standard corresponding to the IP segment information in a preset IP segment and network standard mapping table, and use the searched preset network standard as a target network standard; a selecting module 303, configured to select one transmission policy from preset transmission policies as a target transmission policy according to the target network type, so as to execute a TCP transmission service based on the target transmission policy.

In addition, in another example, the searching module 302 is specifically configured to, when a preset network system corresponding to the IP segment information is found in a preset IP segment and network system mapping table, use the found preset network system as a target network system;

in addition, in another example, the selection device of the transmission policy further includes a network type determination module.

Specifically, the searching module 302 is specifically configured to, when a preset network system corresponding to the IP segment information is found in a preset IP segment and network system mapping table, take the found preset network system as a target network system; when the preset network system corresponding to the IP segment information is not found, the notification acquisition module 301 acquires the characteristic information of the TCP transmission layer, and notifies the network system determination module to analyze and process the characteristic information by using a preset network system determination model, determine the network system corresponding to the characteristic information, and use the determined network system group as the target network system.

In addition, in another example, the network type determining module is further configured to establish a correspondence between the IP segment information and the network type, and add the correspondence to the IP segment and network type mapping table.

In addition, in another example, the selection device of the transmission policy further includes a mapping table construction module.

Specifically, the mapping table constructing module is configured to analyze and process sample feature information in a sample feature set by using the network format determining model, and determine a network format corresponding to each sample feature information, where the sample feature information is pre-collected feature information of a TCP transport layer provided by clients of different IPs in the same IP segment; counting the determined IP number corresponding to different network systems; selecting a network standard of which the IP number meets a preset condition as a preset network standard corresponding to the IP section; and establishing a corresponding relation between the IP section and the preset network system to obtain a mapping table of the IP section and the network system.

In addition, in another example, the selection device of the transmission strategy further comprises a network system determination model training module.

Specifically, the network system determination model training module is used for acquiring characteristic information of a TCP transmission layer corresponding to a known network system and a network system type label of the known network system to obtain a training sample; constructing a network standard training model according to a preset machine learning algorithm; and performing iterative training on the network standard training model by using the training samples until a preset iteration termination condition is met, so as to obtain the network standard determination model.

In addition, in another example, the network system determination model training module is further configured to obtain a maximum segment length received by a TCP transport layer and an initial receiving window size of a client during a period from when a TCP establishes a connection to when a packet starts to be sent, so as to obtain characteristic information of the TCP transport layer corresponding to the known network system.

In addition, in another example, the network format determination model training module is further configured to obtain a maximum segment length and a client initial receiving window size carried in a first TCP request received by a TCP transport layer during a period from a time when a TCP establishes a connection to a time when a packet starts to be sent, so as to obtain characteristic information of the TCP transport layer corresponding to the known network format.

In addition, in another example, the network type determination model training module is further configured to obtain a log or a URL reported by the client; and acquiring the network type mark of the known network type from the log or the URL.

In addition, in another example, the network type determination model training module is further configured to analyze the log or the URL to obtain an analysis result; and judging whether the analysis result comprises network type information or not.

Correspondingly, if the network standard type mark of the known network standard is included, the network standard type mark of the known network standard is obtained from the network standard information; otherwise, a test environment is set up, manual marking is carried out on the known network standard based on the test environment packet capturing identification, and the network standard type mark of the known network standard is obtained.

It should be understood that the present embodiment is a device embodiment corresponding to the first or second embodiment, and the present embodiment can be implemented in cooperation with the first or second embodiment. The related technical details mentioned in the first or second embodiment are still valid in this embodiment, and are not described herein again to reduce repetition. Accordingly, the related-art details mentioned in the present embodiment can also be applied to the first or second embodiment.

It should be noted that, all the modules involved in this embodiment are logic modules, and in practical application, one logic unit may be one physical unit, may also be a part of one physical unit, and may also be implemented by a combination of multiple physical units. In addition, in order to highlight the innovative part of the present invention, a unit which is not so closely related to solve the technical problem proposed by the present invention is not introduced in the present embodiment, but this does not indicate that there is no other unit in the present embodiment.

A fourth embodiment of the present invention relates to a transmission policy selection apparatus, as shown in fig. 4, including at least one processor 401; and a memory 402 communicatively coupled to the at least one processor 401; the memory 402 stores instructions executable by the at least one processor 401, and the instructions are executed by the at least one processor 401, so that the at least one processor 401 can execute the method for selecting the transmission policy described in the first or second embodiment.

Where the memory 402 and the processor 401 are coupled by a bus, which may include any number of interconnected buses and bridges that couple one or more of the various circuits of the processor 401 and the memory 402 together. The bus may also connect various other circuits such as peripherals, voltage regulators, power management circuits, and the like, which are well known in the art, and therefore, will not be described any further herein. A bus interface provides an interface between the bus and the transceiver. The transceiver may be one element or a plurality of elements, such as a plurality of receivers and transmitters, providing a means for communicating with various other apparatus over a transmission medium. The data processed by the processor 401 may be transmitted over a wireless medium via an antenna, which may receive the data and transmit the data to the processor 401.

The processor 401 is responsible for managing the bus and general processing and may provide various functions including timing, peripheral interfaces, voltage regulation, power management, and other control functions. And memory 402 may be used to store data used by processor 401 in performing operations.

A fifth embodiment of the present invention relates to a computer-readable storage medium storing a computer program. The computer program, when executed by a processor, implements an embodiment of the method of selection of a transmission policy described above.

That is, as can be understood by those skilled in the art, all or part of the steps in the method for implementing the embodiments described above may be implemented by a program instructing related hardware, where the program is stored in a storage medium and includes several instructions to enable a device (which may be a single chip, a chip, or the like) or a processor (processor) to execute all or part of the steps of the method described in the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.

It will be understood by those of ordinary skill in the art that the foregoing embodiments are specific embodiments for practicing the invention, and that various changes in form and details may be made therein without departing from the spirit and scope of the invention in practice.

Claims

1. A method for selecting a transmission policy, comprising:

acquiring IP section information of a client;

taking the searched preset network standard as a target network standard;

2. The method for selecting a transmission policy according to claim 1, wherein after the preset network system corresponding to the IP segment information is searched in a preset mapping table of IP segments and network systems, the method further comprises:

if a preset network system corresponding to the IP section information is found in a preset IP section and network system mapping table, executing the step of taking the found preset network system as a target network system;

otherwise, acquiring characteristic information of a TCP transmission layer, analyzing and processing the characteristic information by using a preset network standard determining model, determining a network standard corresponding to the characteristic information, taking the determined network standard as the target network standard, and executing the step of selecting one transmission strategy from preset transmission strategies as a target transmission strategy according to the target network standard.

3. The method for selecting a transmission policy according to claim 2, wherein after the analyzing and processing the feature information by using a preset network system determining model and determining the network system corresponding to the feature information, the method further comprises:

and establishing a corresponding relation between the IP section information and the network system, and adding the corresponding relation into a mapping table of the IP section and the network system.

4. The method for selecting a transmission strategy according to claim 2, wherein the training of the network format determination model comprises:

acquiring characteristic information of a TCP transmission layer corresponding to a known network system and a network system type mark of the known network system to obtain a training sample;

constructing a network standard training model according to a preset machine learning algorithm;

and performing iterative training on the network standard training model by using the training samples until a preset iteration termination condition is met, so as to obtain the network standard determination model.

5. The method for selecting a transmission policy according to claim 4, wherein the obtaining of the characteristic information of the TCP transport layer corresponding to the known network type includes:

and acquiring the maximum message segment length received by a TCP transmission layer and the initial receiving window size of a client during the period from the beginning of TCP connection establishment to the beginning of packet sending, and acquiring the characteristic information of the TCP transmission layer corresponding to the known network system.

6. The method for selecting a transmission policy according to claim 5, wherein said obtaining the maximum segment length received by the TCP transport layer and the initial receiving window size of the client during a period from the beginning of establishing the TCP connection to the beginning of sending the packet, to obtain the characteristic information of the TCP transport layer corresponding to the known network type, comprises:

and acquiring the maximum message segment length and the initial receiving window size of the client carried in the first TCP request received by the TCP transmission layer during the period from the connection establishment of the TCP to the packet sending start, and acquiring the characteristic information of the TCP transmission layer corresponding to the known network system.

7. The method for selecting a transmission policy according to claim 4, wherein the obtaining the network standard type label of the known network standard comprises:

acquiring a log or URL reported by a client;

and acquiring the network type mark of the known network type from the log or the URL.

8. The method for selecting a transmission policy according to claim 7, wherein the obtaining the network standard type label of the known network standard from the log or the URL includes:

analyzing the log or the URL to obtain an analysis result;

judging whether the analysis result comprises network type information or not;

and if so, acquiring the network type mark of the known network type from the network type information.

9. The method for selecting a transmission policy according to claim 8, wherein after said determining whether the analysis result includes the network type information, the method further comprises:

if the analysis result does not include the network standard information, a test environment is established, manual marking is carried out on the known network standard based on the test environment packet capturing identification, and the network standard type mark of the known network standard is obtained.

10. The method for selecting a transmission policy according to claim 1, wherein the constructing of the mapping table between the IP segment and the network system comprises:

analyzing and processing sample characteristic information in a sample characteristic set by using the network system determining model, and determining a network system corresponding to each sample characteristic information, wherein the sample characteristic information is pre-collected characteristic information of a TCP (transmission control protocol) transmission layer provided by clients of different IPs in the same IP section;

counting the determined IP number corresponding to different network systems;

selecting a network standard of which the IP number meets a preset condition as a preset network standard corresponding to the IP section;

and establishing a corresponding relation between the IP section and the preset network system to obtain a mapping table of the IP section and the network system.

11. A selection device of a transmission policy, comprising:

at least one processor; and the number of the first and second groups,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform a method of selecting a transmission strategy according to any one of claims 1 to 10.

12. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out a method of selecting a transmission policy according to any one of claims 1 to 10.