Intelligent control method and system based on VMess
By using VMess intelligent control methods and systems on the remote control platform, mining and parsing traffic characteristics, the problem that the remote control platform cannot distinguish external data exceptions is solved, and the control security and data transmission security are improved.
Patent Information
- Application Number
- CN202411127499.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-16
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2044-08-16
AI Technical Summary
The existing remote control platform cannot independently make abnormal distinctions in external data, resulting in hidden dangers in the safety factor of the control process.
Using intelligent control methods and systems based on VMess, by receiving control decision requests, mining initial traffic characteristics, monitoring network traffic information, extracting features, activating traffic analysis module to process traffic characteristics, generating traffic analysis results, and determining whether to receive decision dependencies based on the results data.
Improve the security of equipment control and the efficiency of distinguishing abnormal data to ensure the security of data transmission.
Smart Images

Figure CN118659928B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field, and in particular to an intelligent control method and system based on VMess. Background Art
[0002] At present, when controlling equipment, the platform mainly collects data from the equipment as the basis for control analysis. With the development of network technology, data transmission has also shifted from offline to online. However, for traditional remote equipment control platforms, data from external sources can enter the intranet as long as they are verified, and data security cannot be guaranteed. The existing technology has a technical problem that the remote control platform cannot autonomously distinguish external data by abnormality, resulting in hidden dangers in the safety factor of the control process. Summary of the invention
[0003] The present application provides a VMess-based intelligent control method and system for solving the technical problem in the prior art that the remote control platform is unable to autonomously distinguish abnormal external data, resulting in hidden dangers in the safety factor of the control process.
[0004] In view of the above problems, the present application provides an intelligent control method and system based on VMess.
[0005] In a first aspect of the present application, a VMess-based intelligent control method is provided, wherein the method is applied to a VMess-based intelligent control system, wherein the system deploys a remote control platform, wherein the remote control platform and a plurality of distributed devices are connected in encrypted communication based on the VMess proxy protocol, including:
[0006] Receiving a control decision request, wherein the control decision request includes a decision task type and a request device number;
[0007] Performing frequency mining according to the decision task type and the request device number to obtain initial traffic characteristics, wherein the initial traffic characteristics include a first traffic scale and a first transmission rate;
[0008] Monitoring network traffic information of decision-dependent data of the control decision request, performing feature extraction, and obtaining a second traffic scale, a second transmission rate, and protocol header character features;
[0009] activating a traffic analysis module to process the first traffic scale, the first transmission rate, the second traffic scale, the second transmission rate, and the protocol header character features to generate a traffic analysis result;
[0010] When the traffic analysis result is failed, refusing to receive the decision-dependent data;
[0011] When the traffic analysis result is passed, the decision-dependent data is received to make a decision.
[0012] The second aspect of the present application provides an intelligent control system based on VMess, the system comprising:
[0013] A decision request receiving unit, used to receive a control decision request, wherein the control decision request includes a decision task type and a request device number;
[0014] An initial traffic feature acquisition unit, configured to perform frequency mining according to the decision task type and the request device number to obtain an initial traffic feature, wherein the initial traffic feature includes a first traffic scale and a first transmission rate;
[0015] A character feature acquisition unit, used for monitoring the network flow information of the decision-dependent data of the control decision request, performing feature extraction, and obtaining a second flow scale, a second transmission rate, and a protocol header character feature;
[0016] A flow analysis result generating unit, used for activating a flow analysis module, processing the first flow scale, the first transmission rate, the second flow scale, the second transmission rate and the protocol header character feature, and generating a flow analysis result;
[0017] A receiving rejection unit, configured to reject receiving the decision-dependent data when the flow analysis result is a failure;
[0018] The decision unit is used to receive the decision-dependent data and make a decision when the flow analysis result is passed.
[0019] One or more technical solutions provided in this application have at least the following technical effects or advantages:
[0020] The present application receives a control decision request, wherein the control decision request includes a decision task type and a requested device number, and then performs frequency mining based on the decision task type and the requested device number to obtain initial traffic features, wherein the initial traffic features include a first traffic scale and a first transmission rate, and then monitors the network traffic information of the decision-dependent data of the control decision request, performs feature extraction, obtains a second traffic scale, a second transmission rate, and a protocol header character feature, activates a traffic parsing module, processes the first traffic scale, the first transmission rate, the second traffic scale, the second transmission rate, and the protocol header character feature, generates a traffic parsing result, and then when the traffic parsing result is not passed, the decision-dependent data is rejected; when the traffic parsing result is passed, the decision-dependent data is received to make a decision. The technical effect of improving the safety of device control and the efficiency of distinguishing abnormal data is achieved. BRIEF DESCRIPTION OF THE DRAWINGS
[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0022] Figure 1 A flow chart of a VMess-based intelligent control method provided in an embodiment of the present application;
[0023] Figure 2 A schematic diagram of a process for obtaining initial flow characteristics in a VMess-based intelligent control method provided in an embodiment of the present application;
[0024] Figure 3 A schematic diagram of a flow chart of generating flow analysis results in a VMess-based intelligent control method provided in an embodiment of the present application;
[0025] Figure 4 A schematic diagram of the structure of a VMess-based intelligent control system provided in an embodiment of the present application.
[0026] Explanation of the reference numerals: decision request receiving unit 11 , initial traffic feature obtaining unit 12 , character feature obtaining unit 13 , traffic analysis result generating unit 14 , rejection receiving unit 15 , decision making unit 16 . DETAILED DESCRIPTION
[0027] The present application provides a VMess-based intelligent control method and system to solve the technical problem in the prior art that the remote control platform is unable to autonomously distinguish abnormal external data, resulting in hidden dangers in the safety factor of the control process.
[0028] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0029] It should be noted that the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or inherent to these processes, methods, products or devices. Embodiment 1
[0030] like Figure 1 As shown, the present application provides an intelligent control method based on VMess, which is applied to an intelligent control system based on VMess, wherein the system deploys a remote control platform, and the remote control platform and several distributed devices are connected in encrypted communication based on the VMess proxy protocol, including:
[0031] S100: receiving a control decision request, wherein the control decision request includes a decision task type and a request device number;
[0032] In a possible embodiment, the remote control platform is used to perform encrypted communication with several distributed devices in the area through the VMess proxy protocol, thereby performing intelligent control of several distributed devices. Among them, the VMess proxy protocol is used to encrypt and transmit data on an insecure network. The remote control platform accepts a control decision request. The control decision request is used to describe the situation in which several distributed devices need to be controlled, including a decision task type and a requested device number. The decision task type is used to determine the control requirements for device control adjustment based on the status of several distributed devices, and optional decision task types include device troubleshooting, device joint call, device remote control, etc. The requested device number is the device number that needs to be controlled in several distributed devices, and the requested device number corresponds to a distributed device and a decision task type. Receiving the control decision request provides data support for subsequent intelligent control.
[0033] S200: performing frequency mining according to the decision task type and the request device number to obtain initial traffic characteristics, wherein the initial traffic characteristics include a first traffic scale and a first transmission rate;
[0034] Further, such as Figure 2 As shown, frequency mining is performed according to the decision task type and the request device number to obtain initial traffic characteristics, wherein the initial traffic characteristics include a first traffic scale and a first transmission rate. Step S200 of the embodiment of the present application also includes:
[0035] Retrieving the decision task type through the dependent data attribute calibration table to obtain a task dependent data attribute set;
[0036] The request device number and the decision task type are retrieved through the control starting point calibration table to obtain the number of task control starting points;
[0037] Determining the decision-dependent data scale according to the number of task control starting points and the task-dependent data attribute set;
[0038] Frequent mining is performed based on the decision-dependent data scale to obtain the initial traffic characteristics, wherein the initial traffic characteristics include the first traffic scale and the first transmission rate.
[0039] Further, the decision-dependent data scale is used for frequency mining to obtain the initial traffic characteristics, wherein the initial traffic characteristics include the first traffic scale and the first transmission rate. Step S200 of the embodiment of the present application further includes:
[0040] Loading channel characteristics according to the requested device number and the remote control platform, wherein the channel characteristics include channel type, channel length and channel width;
[0041] Collecting a data transmission flow sample set based on the channel type, the channel length, the channel width and the decision-dependent data scale as constraints;
[0042] Traversing the data transmission flow sample set to perform feature analysis and generate a flow scale feature set and a transmission rate feature set;
[0043] The traffic scale feature set and the transmission rate feature set are traversed to perform central trend analysis to generate the first traffic scale and the first transmission rate.
[0044] In an embodiment of the present application, frequency mining is performed based on the decision task type and the request device number, and in-depth data analysis is performed on the task transmission conditions that meet the conditions, so as to obtain the initial traffic characteristics. Among them, the initial traffic characteristics reflect the data traffic conditions required when controlling tasks that meet the decision task type and the request device number. The initial traffic characteristics include a first traffic scale and a first transmission rate. The first traffic scale is used to measure the traffic size consumed by all data required for task control of transmission. The first transmission rate is used to describe the amount of data transmitted per unit time, which is the total amount of data required for task control of transmission compared to the transmission time.
[0045] In a possible embodiment, the dependent data attribute calibration table is used to calibrate the data attributes that different decision task types depend on, wherein the dependent data attribute calibration table has a mapping relationship between the decision task type and the dependent data attribute set. The decision task type is used as an index to search in the dependent data attribute calibration table to obtain the task dependent data attribute set. The control starting point calibration table is used to calibrate the source of the dependent data, and has a mapping relationship between the decision task type and the request device number and the corresponding task control quantity. The request device number and the decision task type are used as indexes to search in the control starting point calibration table to obtain the task control starting point quantity. The task control starting point quantity reflects the data transmission load size required when the request device performs control under the decision task type. Further, the task control starting point quantity and the task dependent data attribute set are used as the decision dependent data scale to determine the amount of data that needs to be analyzed when performing intelligent control.
[0046] In a possible embodiment, the decision-dependent data scale is used as an index to perform frequency mining in big data to obtain the initial traffic characteristics. Based on the requested device number and the remote control platform, the channel characteristics that meet the requested device deployment are loaded. Among them, the channel characteristics are used to describe the transmission channel conditions between the device of the requested device number and the remote control platform, including the channel type, channel length and channel width. Among them, the channel type is used to describe the medium or transmission method for transmitting signals during wireless communication, including coaxial cable channels, optical fiber channels, air propagation channels, etc. The channel length is used to describe the transmission distance of the channel between the device of the requested device number and the remote control platform. The channel width is used to describe the bandwidth of the communication channel when transmitting data.
[0047] In one embodiment, the channel type, channel length, channel width and the decision-dependent data scale are used as constraints to search in big data and obtain a data transmission flow sample set that meets the constraint requirements. Then, the data is analyzed based on the data transmission flow sample set to determine the initial flow characteristics of the data transmission situation that meets the decision task type and the request device number, thereby achieving the goal of conducting a practical and reliable analysis of the data transmission flow situation. The data transmission flow sample set is subjected to feature analysis using the flow scale and transmission rate as indexes, respectively, to obtain a flow scale feature set and a transmission rate feature set. Among them, the flow scale feature is the amount of flow required to transmit all data. The transmission rate feature is the amount of data transmitted per unit time, which is the ratio of the total amount of data required for intelligent control to the transmission duration.
[0048] In an embodiment of the present application, a central trend analysis is performed on the traffic scale feature set and the transmission rate feature set, respectively, to determine the data in the feature set that best represents the overall trend of the data in the feature set, thereby obtaining the first traffic scale and the first transmission rate. Optionally, the frequency of occurrence of the traffic scales in the traffic scale feature set is counted separately, and the statistical result is compared with the total frequency of occurrence of each traffic scale in the traffic scale feature set to obtain the weight value of each traffic scale. Furthermore, the traffic scale feature set is weighted according to the weight value of each traffic scale, and the result of the weighted calculation is used as the first traffic scale. Based on the same principle, the first transmission rate is obtained by analyzing the transmission rate feature set.
[0049] S300: monitoring the network traffic information of the decision-dependent data of the control decision request, performing feature extraction, and obtaining a second traffic scale, a second transmission rate, and a protocol header character feature;
[0050] In a possible embodiment, the actual network traffic consumed by the decision-dependent data of the control decision request is monitored, and the monitoring results are feature extracted with the traffic scale, transmission rate and protocol header as indexes to obtain the second traffic scale, the second transmission rate and the protocol header character features. The second traffic scale is the size of the network traffic consumed when the decision-dependent data is actually transmitted. The second transmission rate is the amount of data transmitted per unit time when the decision-dependent data is actually transmitted. The protocol header character features are used to describe the data packet attributes and types of the decision-dependent data.
[0051] S400: activating a traffic analysis module to process the first traffic scale, the first transmission rate, the second traffic scale, the second transmission rate, and the protocol header character features to generate a traffic analysis result;
[0052] Further, such as Figure 3 As shown, the traffic parsing module is activated, the first traffic scale, the first transmission rate, the second traffic scale, the second transmission rate and the protocol header character features are processed to generate a traffic parsing result. Step S400 of the embodiment of the present application also includes:
[0053] The traffic analysis module includes a feature comparison layer and a feature fusion layer, wherein the feature comparison layer includes a traffic scale comparison channel, a transmission rate comparison channel and a character feature comparison channel deployed in parallel;
[0054] Comparing the first flow scale with the second flow scale through the flow scale comparison channel to generate a first deviation vector;
[0055] Compare the first transmission rate and the second transmission rate through the transmission rate comparison channel to generate a second deviation vector;
[0056] Compare the protocol header character features and VMess header character features through the character feature comparison channel to generate a third deviation vector;
[0057] The first deviation vector, the second deviation vector, and the third deviation vector are analyzed through the feature fusion layer to generate the flow analysis result.
[0058] Further, the traffic parsing module is activated to process the first traffic scale, the first transmission rate, the second traffic scale, the second transmission rate and the protocol header character features to generate a traffic parsing result. Previously, step S400 of the embodiment of the present application also includes:
[0059] Construct the first deviation vector comparison function:
[0060] ;
[0061] in, represents the first deviation vector or the second deviation vector, Characterizing a first traffic scale or a first transmission rate, Characterizing a second traffic scale or a second transmission rate, Characterize the scaling parameter, Characterize directional parameters;
[0062] constructing the traffic scale comparison channel or the transmission rate comparison channel according to the first deviation vector comparison function;
[0063] Construct the second deviation vector comparison function:
[0064] ;
[0065] ;
[0066] ;
[0067] in, is the third deviation vector, Characterize the protocol header character features, Characterization Characters at different positions, Characterize the VMess head character features, Characterization Characters at different positions, Characterize the same position Characters and The characters are the same as those in the header, count() represents the counting function, i represents the character sequence number, M represents the total number of characters in the protocol header character feature, and N represents the total number of characters in the VMess header character feature;
[0068] constructing the character feature comparison channel according to the second deviation vector comparison function;
[0069] Collect a flow feature sample data set and train the feature fusion layer.
[0070] In the embodiment of the present application, the flow analysis module is a functional module for performing a refined analysis of the flow situation of the control decision request. After the flow analysis module is activated, the first flow scale, the first transmission rate, the second flow scale, the second transmission rate and the protocol header character features are input into the flow analysis module for processing to obtain the flow analysis result. The flow analysis result reflects
[0071] In one embodiment, the traffic analysis module includes a feature comparison layer and a feature fusion layer. The feature comparison layer is used to compare traffic from three dimensions, including a traffic scale comparison channel, a transmission rate comparison channel, and a character feature comparison channel deployed in parallel. The parallel deployment of the three comparison channels achieves the goal of efficiently performing traffic comparison and improving comparison efficiency. The feature fusion layer is a network layer used to analyze whether the traffic contains the VMess proxy protocol from three dimensions. The first traffic scale and the second traffic scale are input into the first deviation vector comparison function in the traffic scale comparison channel for comparison, and a first deviation vector is obtained. The first deviation vector reflects the degree of deviation between the actual consumption traffic and the consumption traffic determined according to the mining data. The first transmission rate and the second transmission rate are input into the first deviation vector comparison function in the transmission rate comparison channel for data comparison, and a second deviation vector is obtained. The second deviation vector reflects the degree of deviation between the actual data transmission speed and the data transmission speed determined according to the mining data. Furthermore, the protocol header character features and the VMess header character features are input into the character feature comparison channel for comparison, and the third deviation vector is obtained. The third deviation vector is used to describe the degree of difference between characters. Then, the first deviation vector, the second deviation vector and the third deviation vector are input into the feature fusion layer for analysis, and the traffic is comprehensively analyzed from three dimensions to see whether the VMess proxy protocol is included in the traffic to obtain the traffic parsing result. Among them, the traffic parsing result includes pass and fail. When the traffic parsing result is pass, it indicates that the traffic transmitted by the control decision request includes the VMess proxy protocol. When the traffic parsing result is fail, it indicates that the traffic transmitted by the control decision request does not include the VMess proxy protocol.
[0072] In one embodiment, the traffic scale comparison channel and the transmission rate comparison channel are constructed based on the first deviation vector comparison function, and then the character feature comparison channel is constructed based on the second deviation vector comparison function. The traffic feature sample data set is obtained by searching in big data using the traffic feature as an index. The feature fusion layer is trained based on the traffic feature sample data set. The goal of constructing multiple comparison channels and feature fusion layers is achieved.
[0073] Furthermore, a flow feature sample data set is collected to train the feature fusion layer. Step S400 of the embodiment of the present application further includes:
[0074] Any traffic feature sample data of the traffic feature sample data set includes a traffic scale record value, a transmission rate record value, a protocol header character record value, and a VMess proxy protocol inclusion tag, wherein when the VMess proxy protocol inclusion tag is equal to 1, the VMess proxy protocol is included, and when the VMess proxy protocol inclusion tag is equal to 0, the VMess proxy protocol is not included;
[0075] According to the feature comparison layer, traverse the traffic feature sample data set to obtain a deviation vector identification data set;
[0076] The VMess proxy protocol includes labels as supervision and the deviation vector identification data set is used as input to configure the neural network and generate the feature fusion layer.
[0077] Further, the VMess proxy protocol includes a label as supervision, the deviation vector identification data set is used as input, the neural network is configured to generate the feature fusion layer, and step S400 of the embodiment of the present application further includes:
[0078] The VMess proxy protocol includes a label as supervision and the deviation vector identification data set as input, configuring a neural network to generate a first feature fusion channel;
[0079] The support vector machine is configured with the label included in the VMess proxy protocol as supervision and the deviation vector identification data set as input to generate a second feature fusion channel;
[0080] The VMess proxy protocol includes a label as supervision, and the deviation vector identification data set is used as input to configure the random forest to generate a third feature fusion channel;
[0081] Construct the output decision function:
[0082] ;
[0083] in, Represents the output value of the feature fusion layer, Characterize the output value of the first feature fusion channel, the second feature fusion channel, or the third feature fusion channel;
[0084] The feature fusion layer is constructed according to the first feature fusion channel, the second feature fusion channel, the third feature fusion channel and the output decision function.
[0085] In a possible embodiment, any traffic feature sample data of the traffic feature sample data set includes a traffic scale record value, a transmission rate record value, a protocol header character record value, and a VMess proxy protocol inclusion label, wherein when the VMess proxy protocol inclusion label is equal to 1, the VMess proxy protocol is included, and when the VMess proxy protocol inclusion label is equal to 0, the VMess proxy protocol is not included. The VMess proxy protocol inclusion label is used to identify whether the VMess proxy protocol is included in the traffic. By using the feature comparison layer, feature comparison is performed on the traffic scale record value, transmission rate record value, and protocol header character record value in the traffic feature sample data set, respectively, so as to obtain the deviation vector identification data set. Among them, the deviation vector identification data set describes the deviation of the traffic from three dimensions. Then, with the VMess proxy protocol inclusion label as supervision and the deviation vector identification data set as input, the neural network is configured to generate the feature fusion layer. The goal of building an intelligent network layer for comprehensive traffic analysis is achieved.
[0086] In one embodiment, the VMess proxy protocol contains labels as supervision, the deviation vector identification data set as input, and the network layer constructed based on the convolutional neural network is supervised and trained until the output reaches convergence, generating a first feature fusion channel. The VMess proxy protocol contains labels as supervision, the deviation vector identification data set as input, and the support vector machine is configured to generate a second feature fusion channel. Optionally, a person skilled in the art sets weights for different categories of data sets in the deviation vector identification data set, and then inputs the deviation vector identification data set into the support vector machine constructed according to the set weights, and the VMess proxy protocol contains labels as supervision to generate the second feature fusion channel. Then, the deviation vector identification data set is used as input, and the VMess proxy protocol contains labels as supervision data to train the network layer in the random forest until the output reaches convergence, generating the third feature fusion channel. The output decision parameter is used to comprehensively analyze and judge the output values of the first feature fusion channel, the second feature fusion channel, and the third feature fusion channel. Furthermore, the first feature fusion channel, the second feature fusion channel, and the third feature fusion channel are fully connected in parallel, and then the output layers of the first feature fusion channel, the second feature fusion channel, and the third feature fusion channel are connected to the output decision function to generate the feature fusion layer.
[0087] S500: When the traffic analysis result is failed, the decision-dependent data is rejected;
[0088] S600: When the traffic analysis result is passed, receiving the decision dependency data to make a decision.
[0089] In an embodiment of the present application, when the flow parsing result is not passed, it indicates that the decision-dependent data does not contain the VMess proxy protocol, and data security cannot be guaranteed. At this time, the decision-dependent data is rejected. Then, when the flow parsing result is passed, it indicates that the decision-dependent data includes the VMess proxy protocol. At this time, the data is safe and the decision-dependent data can be received to make a decision. The technical effect of improving the reliability of intelligent control and the efficiency of security analysis is achieved.
[0090] In summary, the embodiments of the present application have at least the following technical effects:
[0091] The present application receives a control decision request, wherein the control decision request includes a decision task type and a request device number, and then performs frequency mining according to the decision task type and the request device number to obtain the initial traffic characteristics, thereby achieving the goal of determining the traffic situation of the decision-dependent data according to the control decision request, and then monitoring the network traffic information of the decision-dependent data of the control decision request, performing feature extraction, obtaining the second traffic scale, the second transmission rate and the protocol header character characteristics, thereby achieving the goal of multi-dimensional feature recognition of the traffic situation, and then activating the traffic parsing module, performing a comprehensive analysis of the first traffic scale, the first transmission rate, the second traffic scale, the second transmission rate and the protocol header character characteristics, generating a traffic parsing result, and then when the traffic parsing result is not passed, the decision-dependent data is rejected; when the traffic parsing result is passed, the decision-dependent data is received for decision making. The technical effect of improving the efficiency of distinguishing abnormal data, improving the degree of control intelligence and control safety is achieved. Embodiment 2
[0092] Based on the same inventive concept as the VMess-based intelligent control method in the aforementioned embodiment, Figure 4 As shown, the present application provides an intelligent control system based on VMess, and the system and method embodiments in the present application are based on the same inventive concept. The system includes:
[0093] A decision request receiving unit 11 is used to receive a control decision request, wherein the control decision request includes a decision task type and a request device number;
[0094] An initial traffic feature obtaining unit 12 is used to perform frequency mining according to the decision task type and the request device number to obtain an initial traffic feature, wherein the initial traffic feature includes a first traffic scale and a first transmission rate;
[0095] A character feature acquisition unit 13, used for monitoring the network flow information of the decision-dependent data of the control decision request, performing feature extraction, and obtaining a second flow scale, a second transmission rate, and a protocol header character feature;
[0096] A traffic analysis result generating unit 14 is used to activate a traffic analysis module, process the first traffic scale, the first transmission rate, the second traffic scale, the second transmission rate and the protocol header character feature, and generate a traffic analysis result;
[0097] A rejection receiving unit 15, configured to reject receiving the decision-dependent data when the flow analysis result is failed;
[0098] The decision unit 16 is configured to receive the decision-dependent data and make a decision when the traffic analysis result is passed.
[0099] Furthermore, the initial flow characteristic obtaining unit 12 is used to perform the following steps:
[0100] Retrieving the decision task type through the dependent data attribute calibration table to obtain a task dependent data attribute set;
[0101] The request device number and the decision task type are retrieved through the control starting point calibration table to obtain the number of task control starting points;
[0102] Determining the decision-dependent data scale according to the number of task control starting points and the task-dependent data attribute set;
[0103] Frequent mining is performed based on the decision-dependent data scale to obtain the initial traffic characteristics, wherein the initial traffic characteristics include the first traffic scale and the first transmission rate.
[0104] Furthermore, the initial flow characteristic obtaining unit 12 is used to perform the following steps:
[0105] Loading channel characteristics according to the requested device number and the remote control platform, wherein the channel characteristics include channel type, channel length and channel width;
[0106] Collecting a data transmission flow sample set based on the channel type, the channel length, the channel width and the decision-dependent data scale as constraints;
[0107] Traversing the data transmission flow sample set to perform feature analysis and generate a flow scale feature set and a transmission rate feature set;
[0108] The traffic scale feature set and the transmission rate feature set are traversed to perform central trend analysis to generate the first traffic scale and the first transmission rate.
[0109] Furthermore, the traffic analysis result generating unit 14 is used to perform the following steps:
[0110] The traffic analysis module includes a feature comparison layer and a feature fusion layer, wherein the feature comparison layer includes a traffic scale comparison channel, a transmission rate comparison channel and a character feature comparison channel deployed in parallel;
[0111] Comparing the first flow scale with the second flow scale through the flow scale comparison channel to generate a first deviation vector;
[0112] Compare the first transmission rate and the second transmission rate through the transmission rate comparison channel to generate a second deviation vector;
[0113] Compare the protocol header character features and VMess header character features through the character feature comparison channel to generate a third deviation vector;
[0114] The first deviation vector, the second deviation vector, and the third deviation vector are analyzed through the feature fusion layer to generate the flow analysis result.
[0115] Furthermore, the traffic analysis result generating unit 14 is used to perform the following steps:
[0116] Construct the first deviation vector comparison function:
[0117] ;
[0118] in, represents the first deviation vector or the second deviation vector, Characterizing a first traffic scale or a first transmission rate, Characterizing a second traffic scale or a second transmission rate, Characterize the scaling parameter, Characterize directional parameters;
[0119] constructing the traffic scale comparison channel or the transmission rate comparison channel according to the first deviation vector comparison function;
[0120] Construct the second deviation vector comparison function:
[0121] ;
[0122] ;
[0123] ;
[0124] in, is the third deviation vector, Characterize the protocol header character features, Characterization Characters at different positions, Characterize the VMess head character features, Characterization Characters at different positions, Characterize the same position Characters and The characters are the same as those in the header, count() represents the counting function, i represents the character sequence number, M represents the total number of characters in the protocol header character feature, and N represents the total number of characters in the VMess header character feature;
[0125] constructing the character feature comparison channel according to the second deviation vector comparison function;
[0126] Collect a flow feature sample data set and train the feature fusion layer.
[0127] Furthermore, the traffic analysis result generating unit 14 is used to perform the following steps:
[0128] Any traffic feature sample data of the traffic feature sample data set includes a traffic scale record value, a transmission rate record value, a protocol header character record value, and a VMess proxy protocol inclusion tag, wherein when the VMess proxy protocol inclusion tag is equal to 1, the VMess proxy protocol is included, and when the VMess proxy protocol inclusion tag is equal to 0, the VMess proxy protocol is not included;
[0129] According to the feature comparison layer, traverse the traffic feature sample data set to obtain a deviation vector identification data set;
[0130] The VMess proxy protocol includes labels as supervision and the deviation vector identification data set is used as input to configure the neural network and generate the feature fusion layer.
[0131] Furthermore, the traffic analysis result generating unit 14 is used to perform the following steps:
[0132] The VMess proxy protocol includes a label as supervision and the deviation vector identification data set as input, configuring a neural network to generate a first feature fusion channel;
[0133] The support vector machine is configured with the label included in the VMess proxy protocol as supervision and the deviation vector identification data set as input to generate a second feature fusion channel;
[0134] The VMess proxy protocol includes a label as supervision, and the deviation vector identification data set is used as input to configure the random forest to generate a third feature fusion channel;
[0135] Construct the output decision function:
[0136] ;
[0137] in, Represents the output value of the feature fusion layer, Characterize the output value of the first feature fusion channel, the second feature fusion channel, or the third feature fusion channel;
[0138] The feature fusion layer is constructed according to the first feature fusion channel, the second feature fusion channel, the third feature fusion channel and the output decision function.
[0139] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. Other embodiments are within the scope of the attached claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0140] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.
[0141] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.
Claims
1. Intelligent control method based on VMess, It is characterized in that Applied to a VMess-based intelligent control system, the system deploys a remote control platform, the remote control platform and several distributed devices are connected in encrypted communication based on the VMess proxy protocol, including: Receiving a control decision request, wherein the control decision request includes a decision task type and a request device number; Performing frequency mining according to the decision task type and the request device number to obtain initial traffic characteristics, wherein the initial traffic characteristics include a first traffic scale and a first transmission rate; Monitoring network traffic information of decision-dependent data of the control decision request, performing feature extraction, and obtaining a second traffic scale, a second transmission rate, and protocol header character features; activating a traffic analysis module to process the first traffic scale, the first transmission rate, the second traffic scale, the second transmission rate, and the protocol header character features to generate a traffic analysis result; When the traffic analysis result is failed, refusing to receive the decision-dependent data; When the traffic analysis result is passed, receiving the decision-dependent data to make a decision; The flow analysis module is activated to process the first flow scale, the first transmission rate, the second flow scale, the second transmission rate and the protocol header character feature to generate a flow analysis result, including: The traffic analysis module includes a feature comparison layer and a feature fusion layer, wherein the feature comparison layer includes a traffic scale comparison channel, a transmission rate comparison channel and a character feature comparison channel deployed in parallel; Comparing the first flow scale with the second flow scale through the flow scale comparison channel to generate a first deviation vector; Compare the first transmission rate and the second transmission rate through the transmission rate comparison channel to generate a second deviation vector; Compare the protocol header character features and VMess header character features through the character feature comparison channel to generate a third deviation vector; The first deviation vector, the second deviation vector and the third deviation vector are analyzed by the feature fusion layer to generate the flow analysis result; Further, activating a traffic analysis module, processing the first traffic scale, the first transmission rate, the second traffic scale, the second transmission rate and the protocol header character features, and generating a traffic analysis result, previously comprising: Construct the first deviation vector comparison function: ; in, represents the first deviation vector or the second deviation vector, Characterizing a first traffic scale or a first transmission rate, Characterizing a second traffic scale or a second transmission rate, Characterize the scaling parameters, Characterize directional parameters; constructing the traffic scale comparison channel or the transmission rate comparison channel according to the first deviation vector comparison function; Construct the second deviation vector comparison function: ; ; ; in, is the third deviation vector, Characterize the protocol header character features, Characterization Characters at different positions, Characterize the VMess head character features, Characterization Characters at different positions, Characterize the same position Characters and The characters are the same as those in the header, count() represents the counting function, i represents the character sequence number, M represents the total number of characters in the protocol header character feature, and N represents the total number of characters in the VMess header character feature; constructing the character feature comparison channel according to the second deviation vector comparison function; Collecting a flow feature sample data set and training the feature fusion layer; Further, a flow feature sample data set is collected to train the feature fusion layer, including: Any traffic feature sample data of the traffic feature sample data set includes a traffic scale record value, a transmission rate record value, a protocol header character record value, and a VMess proxy protocol inclusion tag, wherein when the VMess proxy protocol inclusion tag is equal to 1, the VMess proxy protocol is included, and when the VMess proxy protocol inclusion tag is equal to 0, the VMess proxy protocol is not included; According to the feature comparison layer, traverse the traffic feature sample data set to obtain a deviation vector identification data set; Using the VMess proxy protocol containing labels as supervision and the deviation vector identification data set as input, the neural network is configured to generate the feature fusion layer; Further, the VMess proxy protocol includes a label as supervision, and the deviation vector identification data set is used as input to configure the neural network to generate the feature fusion layer, including: The VMess proxy protocol includes a label as supervision and the deviation vector identification data set as input, configuring a neural network to generate a first feature fusion channel; The support vector machine is configured with the label included in the VMess proxy protocol as supervision and the deviation vector identification data set as input to generate a second feature fusion channel; The VMess proxy protocol includes a label as supervision, and the deviation vector identification data set is used as input to configure the random forest to generate a third feature fusion channel; Construct the output decision function: ; in, Represents the output value of the feature fusion layer, Characterize the output value of the first feature fusion channel, the second feature fusion channel, or the third feature fusion channel; The feature fusion layer is constructed according to the first feature fusion channel, the second feature fusion channel, the third feature fusion channel and the output decision function.
2. The method according to claim 1, It is characterized in that Performing frequency mining according to the decision task type and the request device number to obtain initial traffic characteristics, wherein the initial traffic characteristics include a first traffic scale and a first transmission rate, including: Retrieving the decision task type through the dependent data attribute calibration table to obtain a task dependent data attribute set; The request device number and the decision task type are retrieved through the control starting point calibration table to obtain the number of task control starting points; Determining the decision-dependent data scale according to the number of task control starting points and the task-dependent data attribute set; Frequent mining is performed based on the decision-dependent data scale to obtain the initial traffic characteristics, wherein the initial traffic characteristics include the first traffic scale and the first transmission rate.
3. The method according to claim 2, It is characterized in that Performing frequency mining on the decision-dependent data scale to obtain the initial traffic characteristics, wherein the initial traffic characteristics include the first traffic scale and the first transmission rate, including: Loading channel characteristics according to the requested device number and the remote control platform, wherein the channel characteristics include channel type, channel length and channel width; Collecting a data transmission flow sample set based on the channel type, the channel length, the channel width and the decision-dependent data scale as constraints; Traversing the data transmission flow sample set to perform feature analysis and generate a flow scale feature set and a transmission rate feature set; The traffic scale feature set and the transmission rate feature set are traversed to perform central trend analysis to generate the first traffic scale and the first transmission rate.
4. Intelligent control system based on VMess, It is characterized in that The system comprises: A decision request receiving unit, used to receive a control decision request, wherein the control decision request includes a decision task type and a request device number; An initial traffic feature acquisition unit, configured to perform frequency mining according to the decision task type and the request device number to obtain an initial traffic feature, wherein the initial traffic feature includes a first traffic scale and a first transmission rate; A character feature acquisition unit, used for monitoring the network flow information of the decision-dependent data of the control decision request, performing feature extraction, and obtaining a second flow scale, a second transmission rate, and a protocol header character feature; A flow analysis result generating unit, used for activating a flow analysis module, processing the first flow scale, the first transmission rate, the second flow scale, the second transmission rate and the protocol header character feature, and generating a flow analysis result; A receiving rejection unit, configured to reject receiving the decision-dependent data when the flow analysis result is a failure; A decision-making unit, configured to receive the decision-dependent data and make a decision when the traffic analysis result is passed; The flow analysis result generating unit is further used to perform the following steps: The traffic analysis module includes a feature comparison layer and a feature fusion layer, wherein the feature comparison layer includes a traffic scale comparison channel, a transmission rate comparison channel and a character feature comparison channel deployed in parallel; Comparing the first flow scale with the second flow scale through the flow scale comparison channel to generate a first deviation vector; Compare the first transmission rate and the second transmission rate through the transmission rate comparison channel to generate a second deviation vector; Compare the protocol header character features and VMess header character features through the character feature comparison channel to generate a third deviation vector; The first deviation vector, the second deviation vector and the third deviation vector are analyzed by the feature fusion layer to generate the flow analysis result; Construct the first deviation vector comparison function: ; in, represents the first deviation vector or the second deviation vector, Characterizing a first traffic scale or a first transmission rate, Characterizing a second traffic scale or a second transmission rate, Characterize the scaling parameters, Characterize directional parameters; constructing the traffic scale comparison channel or the transmission rate comparison channel according to the first deviation vector comparison function; Construct the second deviation vector comparison function: ; ; ; in, is the third deviation vector, Characterize the protocol header character features, Characterization Characters at different positions, Characterize the VMess head character features, Characterization Characters at different positions, Characterize the same position Characters and The characters are the same as those in the header, count() represents the counting function, i represents the character sequence number, M represents the total number of characters in the protocol header character feature, and N represents the total number of characters in the VMess header character feature; constructing the character feature comparison channel according to the second deviation vector comparison function; Collecting a flow feature sample data set and training the feature fusion layer; Any traffic feature sample data of the traffic feature sample data set includes a traffic scale record value, a transmission rate record value, a protocol header character record value, and a VMess proxy protocol inclusion tag, wherein when the VMess proxy protocol inclusion tag is equal to 1, the VMess proxy protocol is included, and when the VMess proxy protocol inclusion tag is equal to 0, the VMess proxy protocol is not included; According to the feature comparison layer, traverse the traffic feature sample data set to obtain a deviation vector identification data set; Using the VMess proxy protocol containing labels as supervision and the deviation vector identification data set as input, the neural network is configured to generate the feature fusion layer; The VMess proxy protocol includes a label as supervision and the deviation vector identification data set as input, configuring a neural network to generate a first feature fusion channel; The support vector machine is configured with the label included in the VMess proxy protocol as supervision and the deviation vector identification data set as input to generate a second feature fusion channel; The VMess proxy protocol includes a label as supervision, and the deviation vector identification data set is used as input to configure the random forest to generate a third feature fusion channel; Construct the output decision function: ; in, Represents the output value of the feature fusion layer, Characterize the output value of the first feature fusion channel, the second feature fusion channel, or the third feature fusion channel; The feature fusion layer is constructed according to the first feature fusion channel, the second feature fusion channel, the third feature fusion channel and the output decision function.
Citation Information
Patent Citations
Anonymous network traffic identification method and device based on traffic reconstruction and inheritance learning
CN114615093A
Network abnormal data transmission behavior clustering identification method based on gravitation model
CN117633560A
