WO2022012429A1 - 用于实现终端验证的方法、装置、系统、设备及存储介质 - Google Patents

用于实现终端验证的方法、装置、系统、设备及存储介质 Download PDF

Info

Publication number
WO2022012429A1
WO2022012429A1 PCT/CN2021/105494 CN2021105494W WO2022012429A1 WO 2022012429 A1 WO2022012429 A1 WO 2022012429A1 CN 2021105494 W CN2021105494 W CN 2021105494W WO 2022012429 A1 WO2022012429 A1 WO 2022012429A1
Authority
WO
WIPO (PCT)
Prior art keywords
transmission
terminal
target
data stream
downlink
Prior art date
Application number
PCT/CN2021/105494
Other languages
English (en)
French (fr)
Inventor
薛莉
徐威旺
孙超
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from CN202011198953.5A external-priority patent/CN114006714A/zh
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Priority to JP2023502830A priority Critical patent/JP2023533354A/ja
Priority to EP21842651.8A priority patent/EP4171095A4/en
Priority to CA3186107A priority patent/CA3186107A1/en
Publication of WO2022012429A1 publication Critical patent/WO2022012429A1/zh
Priority to US18/154,263 priority patent/US20230171264A1/en

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W12/00Security arrangements; Authentication; Protecting privacy or anonymity
    • H04W12/12Detection or prevention of fraud
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L63/00Network architectures or network communication protocols for network security
    • H04L63/12Applying verification of the received information
    • H04L63/126Applying verification of the received information the source of the received data
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L5/00Arrangements affording multiple use of the transmission path
    • H04L5/003Arrangements for allocating sub-channels of the transmission path
    • H04L5/0044Arrangements for allocating sub-channels of the transmission path allocation of payload
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L63/00Network architectures or network communication protocols for network security
    • H04L63/08Network architectures or network communication protocols for network security for authentication of entities

Definitions

  • the present application relates to the field of communication technologies, and in particular, to a method, apparatus, system, device and storage medium for realizing terminal authentication.
  • IOT terminals In industries such as smart parks, higher education, manufacturing, and finance, it is common for some abnormal terminals to illegally access servers. There are many types of abnormal terminals. Taking counterfeit terminals as an example, more and more Internet of things (IOT) terminals support network access functions, compared to smart terminals (such as computers, tablets, mobile phones, etc.) , IOT terminals have weak security protection functions and are easily counterfeited. In order to avoid potential security risks caused by counterfeit terminals, network devices generally verify whether the IOT terminals are counterfeit terminals by verifying the IOT terminals they access.
  • IOT Internet of things
  • the terminal verification process is generally as follows: after receiving a data stream output by an IOT terminal, the network device extracts the Internet Protocol (IP) address of the IOT terminal from the packets in the data stream, and queries the IP address. Address library, if the IP address of the IOT terminal is stored in the IP address library, the network device determines that the IOT terminal is not a counterfeit terminal, and the verification of the IOT terminal passes, if the IP address of the IOT terminal is not stored in the IP address library, Then the network device determines that the IOT terminal is a counterfeit terminal, and the IOT terminal fails the verification.
  • IP Internet Protocol
  • the IP address of the IOT terminal can be easily counterfeited, if the counterfeit terminal outputs a data stream to the network device by counterfeiting the IP address in the IP address database, then the IP address extracted by the network device from the data stream output by the counterfeit terminal is The IP address in the IP address database, the network device will pass the verification of the counterfeit terminal. It can be seen that the network device cannot accurately verify the counterfeit terminal through the above-mentioned terminal verification process, that is, the accuracy of the terminal verification is low.
  • the present application provides a method, apparatus, system, device and storage medium for realizing terminal verification, which can improve the accuracy of terminal verification.
  • the technical solution is as follows:
  • a method for realizing terminal verification comprising:
  • first transmission characteristic of the first terminal reconstructing the first transmission characteristic based on the first transmission characteristic to obtain a second transmission characteristic; if the difference between the first transmission characteristic and the second transmission characteristic is The difference between the two is greater than or equal to the target difference, and it is determined that the first terminal fails the verification; wherein, the first transmission characteristic is the overall transmission characteristic of at least one first data stream transmitted by the first terminal, and the The second transmission characteristic is the reconstructed first transmission characteristic.
  • reconstruction refers to reconstruction, and the reconstructing the first transmission characteristic based on the first transmission characteristic means: on the basis of the first transmission characteristic, through a preset algorithm Then, a second transmission feature is reconstructed, and the reconstructed second transmission feature is as consistent as possible with the first transmission feature.
  • the preset algorithm is used to reconstruct the normal transmission characteristics as much as possible, and the preset algorithm includes dimension reduction encoding and dimension increase decoding, wherein the dimension reduction encoding is an encoding method for reducing the dimension of transmission features, and the dimension increase encoding is an increase in dimension encoding.
  • the performing the first transmission feature based on the first transmission feature includes: performing dimension reduction encoding on the first transmission feature, and performing dimension-up decoding on the dimension-reduced encoded first transmission feature.
  • the preset algorithm is expressed by the following target model.
  • This method verifies the terminal by reconstructing the transmission characteristics of the terminal. For example, if the difference between the reconstructed transmission characteristics and the transmission characteristics of the terminal is large, it means that the transmission characteristics of the terminal are abnormal, and the terminal is abnormal. terminal, then it is determined that the terminal has not passed the verification. Because the terminal has specific normal transmission characteristics, the normal transmission characteristics are not easy to be counterfeited. Therefore, this method can accurately verify various abnormal terminals, improve the accuracy of terminal verification, and counterfeit terminals. It is a kind of abnormal terminal. Therefore, this method can also accurately verify the counterfeit terminal, instead of simply verifying the IP address of the terminal, so as to prevent the counterfeit terminal from passing the verification.
  • the transmission characteristic of a normal terminal is also the normal transmission characteristic of the terminal
  • the transmission characteristic of the abnormal terminal is also the abnormal transmission characteristic of the terminal.
  • the first transmission characteristic includes an uplink transmission characteristic
  • the uplink transmission characteristic is an overall transmission characteristic of at least one uplink data stream in the at least one first data stream.
  • the upstream transmission feature includes at least one of an upstream message feature and an upstream feature
  • the upstream message feature is an overall feature of an upstream message in the at least one upstream data stream
  • the upstream characteristics are statistical characteristics of the at least one upstream data stream.
  • the characteristics of the uplink packets include the average transmission interval of the uplink packets, the average value of the uplink load, the total uplink load size, the number of uplink packets, the number of uplink target packets, and the uplink target packets.
  • At least one of a ratio and a load fluctuation value of an uplink packet the average transmission interval of the uplink packet is the average transmission interval of the uplink packet in a time window
  • the average value of the uplink load is the average transmission interval of the uplink packet in the time window
  • the average size of the load of the target packets in the at least one upstream data flow, the total upstream load size is the total size of the load of the target packets in the at least one upstream data flow in the time window
  • the The number of upstream packets is the number of upstream packets in the at least one upstream data stream within the time window
  • the number of upstream target packets is the target number in the at least one upstream data stream within the time window.
  • the number of packets, the proportion of the uplink target packets is the proportion of the target packets in the at least one upstream data flow in the time window, and the uplink packet load fluctuation value is used to indicate that in the Fluctuation of the size of the target packet in the at least one upstream data stream within the time window;
  • the upstream characteristics include the upstream terminal port fluctuation value, the total number of upstream data streams, the number of upstream target data streams, the number of upstream data streams under each data stream type in the at least one data stream type, and at least one type of transmission. At least one of the number of upstream data streams under each transmission protocol type in the protocol type, and the upstream terminal port fluctuation value is used to indicate the at least one upstream data stream in the first terminal within the time window.
  • the fluctuation situation of the output port, the upstream target data flow is the upstream data flow whose corresponding server input port belongs to the target port range.
  • the uplink packet feature further includes at least one of a first receiving window fluctuation value and an average value of a first receiving window size, where the first receiving window fluctuation value is used to indicate that the Fluctuation of the size of the receiving window carried by the uplink packet within the time window.
  • the fluctuation value of the first receiving window is a standard deviation of the receiving window size carried by the uplink packet within the time window.
  • the first transmission characteristic further includes at least one of the total number of the at least one first data stream and a downlink transmission characteristic
  • the downlink transmission characteristic is the at least one first data stream.
  • the downlink transmission characteristic includes at least one of a downlink packet characteristic and a downlink characteristic
  • the downlink packet characteristic is an overall characteristic of a downlink packet in the at least one downlink data stream
  • the downstream characteristics are statistical characteristics of the at least one downstream data stream.
  • the characteristics of the downlink packets include the average transmission interval of downlink packets, the average value of downlink loads, the size of the total downlink load, the number of downlink packets, the number of downlink target packets, and the downlink target packets.
  • the average transmission interval of the downlink packets is the average transmission interval of the downlink packets in a time window
  • the average downlink load is the average transmission interval of the downlink packets in the time window the average size of the load of the target packets in the at least one downstream data stream
  • the total downstream load size is the total size of the load of the target packets in the at least one downstream data stream within the time window
  • the The number of downlink packets is the number of downlink packets in the at least one downlink data stream in the time window
  • the number of downlink target packets is the target number in the at least one downlink data stream in the time window.
  • the number of packets, the proportion of the downlink target packets is the proportion of the target packets in the at least one downstream data stream within the time window, and the downlink packet load fluctuation value is used to indicate that in the time window Fluctuation of the size of the target packet in the at least one downstream data stream within the time window;
  • the downstream characteristics include a downlink terminal port fluctuation value, the total number of downstream data streams, the number of downstream target data streams, the number of downstream data streams under each data stream type in at least one data stream type, and at least one type of transmission. At least one of the number of downlink data streams under each transmission protocol type in the protocol type, and the downlink terminal port fluctuation value is used to indicate the at least one downlink data stream in the first terminal within the time window.
  • the fluctuation situation of the input port, the downlink target data stream is the downlink data stream whose corresponding server output port belongs to the target port range.
  • the downlink packet feature further includes at least one of a second receiving window fluctuation value and an average value of a second receiving window size, where the second receiving window fluctuation value is used to indicate the downlink Fluctuation of the receive window size carried by the packet.
  • the fluctuation value of the second receiving window is a standard deviation of the receiving window size carried by the downlink packet within the time window.
  • the acquiring the first transmission characteristic of the first terminal includes:
  • the stream transmission characteristic of a first data stream includes at least one of transmission information, data stream type, destination port type, and packet characteristics of the first data stream, and the transmission information is used
  • the target port type is the port type of the port in the server that transmits the first data stream
  • the message feature is the feature of the message in the first data stream.
  • the transmission information includes at least one element in a five-tuple of the first data stream
  • the message characteristics include at least one of the sum of message transmission intervals, the load size, the sum of squares of the loads, the number of target messages, the total number of messages, the fluctuation value of the receiving window, the total size of the window, and the sum of squares of the window.
  • the sum of packet transmission intervals is the total duration of transmission intervals between packets in the first data stream within a time window
  • the load size is the target packet in the first data stream within the time window
  • the total size of the load is the square sum of the load size of the target packet
  • the number of the target packet is the total number of the target packet in the packet within the time window
  • the total number of packets is the total number of packets in the time window
  • the receiving window fluctuation value is used to indicate the fluctuation of the receiving window size carried by the packet in the time window.
  • the total size of the window is the sum of the sizes of the receiving windows carried by the message within the time window
  • the sum of squares of the windows is the sum of the squares of the sizes of the sliding windows.
  • the transmission information further includes at least one of a direction identifier and an identifier of the time window, where the direction identifier is used to indicate a transmission direction of the first data stream.
  • the reconstructing the first transmission characteristic based on the first transmission characteristic to obtain the second transmission characteristic includes:
  • the first transmission feature is input into a target model, and the target model reconstructs the first transmission feature based on the inputted first transmission feature, and outputs the second transmission feature.
  • the method before the inputting the first transmission feature into the target model, the method further includes:
  • the third transmission characteristic is an overall transmission characteristic of at least one data stream transmitted by a second terminal.
  • the method before the inputting the first transmission feature into the target model, the method further includes:
  • the target model reconstructs each fourth transmission feature based on the multiple input fourth transmission features, and outputs multiple and obtaining the target difference degree based on the plurality of fifth transmission characteristics and the plurality of fourth transmission characteristics;
  • the target terminal type is the terminal type of the first terminal, the plurality of fourth transmission characteristics and the plurality of fifth transmission characteristics are in one-to-one correspondence, and one fourth transmission characteristic is at least one transmission characteristic of a target terminal.
  • the obtaining the target difference degree based on the plurality of fifth transmission characteristics and the plurality of fourth transmission characteristics includes:
  • the target difference degree is obtained based on a degree of difference between at least one fifth transmission characteristic among the plurality of fifth transmission characteristics and a corresponding fourth transmission characteristic.
  • the method before the inputting the first transmission feature into the target model, the method further includes:
  • the target model Inputting multiple sixth transmission features of multiple target terminals under the target terminal type into the target model, the target model reconstructs each sixth transmission feature based on the multiple input sixth transmission features, and outputs multiple a seventh transmission characteristic; based on the degree of difference between the plurality of seventh transmission characteristics and the corresponding sixth transmission characteristic, it is determined that the target model has passed the verification;
  • the target terminal type is the terminal type of the first terminal, the plurality of sixth transmission characteristics are in one-to-one correspondence with the plurality of seventh transmission characteristics, and one sixth transmission characteristic is at least one transmission characteristic of a target terminal.
  • the method before the inputting the first transmission feature into the target model, the method further includes:
  • Acquire transmission information of at least one second data stream transmitted by the at least one second terminal associate and store the transmission information with the terminal type of the second terminal; acquire the transmission information associated with the terminal type based on the transmission information associated with the terminal type. multiple transmission characteristics of the at least one second terminal;
  • the transmission information is used to indicate the transmission attribute of the data stream, and one transmission characteristic is the overall transmission characteristic of at least one data stream transmitted by the terminal.
  • the method before the inputting the first transmission feature into the target model, the method further includes:
  • the target model is received from a control device.
  • the execution body of the method is a control device or a network device.
  • a method for realizing terminal verification comprising:
  • a third transmission characteristic is the overall transmission characteristic of at least one data stream transmitted by a second terminal
  • the target model is used to reconstruct the transmission characteristic of the verified terminal of the target terminal type, so as to verify the verification of the verified terminal.
  • the terminal performs verification
  • the transmission characteristic is the overall transmission characteristic of at least one data stream transmitted by the verified terminal.
  • the method further includes:
  • the target model is sent to a network device.
  • the method before the multiple third transmission features are used as the input and output of the initial model, before the training, the method further includes:
  • the plurality of fourth transmission characteristics are in one-to-one correspondence with the plurality of fifth transmission characteristics, and one fourth transmission characteristic is an overall transmission characteristic of at least one data stream transmitted by a target terminal.
  • the obtaining the target difference degree based on the plurality of fifth transmission characteristics and the plurality of fourth transmission characteristics includes:
  • the target difference degree is obtained based on a degree of difference between at least one fifth transmission characteristic among the plurality of fifth transmission characteristics and a corresponding fourth transmission characteristic.
  • the method before the multiple third transmission features are used as the input and output of the initial model, before the training, the method further includes:
  • the plurality of sixth transmission characteristics are in one-to-one correspondence with the plurality of seventh transmission characteristics, and one sixth transmission characteristic is an overall transmission characteristic of at least one data stream transmitted by a target terminal.
  • the method before the acquiring multiple third transmission characteristics of at least one second terminal of the target terminal type, the method further includes:
  • Acquire transmission information of at least one second data stream transmitted by the at least one second terminal associate and store the transmission information with the terminal type of the second terminal; acquire the transmission information associated with the terminal type based on the transmission information associated with the terminal type. multiple transmission characteristics of the at least one second terminal;
  • the transmission information is used to indicate the transmission attribute of the data stream, and one transmission characteristic is the overall transmission characteristic of at least one data stream transmitted by the terminal.
  • a system for realizing terminal verification includes a control device and a network device;
  • the control device is used for:
  • one third transmission characteristic is an overall transmission characteristic of at least one data stream transmitted by one second terminal
  • the network equipment is used for:
  • the first transmission characteristic is an overall transmission characteristic of at least one first data stream transmitted by the first terminal
  • the target model Inputting the first transmission feature into the target model, the target model reconstructs the first transmission feature based on the inputted first transmission feature, and outputs a second transmission feature, where the second transmission feature is the reconstructed first transmission feature;
  • the degree of difference between the first transmission characteristic and the second transmission characteristic is greater than or equal to the target degree of difference, it is determined that the first terminal fails the verification.
  • control device is further used for:
  • the plurality of fourth transmission characteristics are in one-to-one correspondence with the plurality of fifth transmission characteristics, and one fourth transmission characteristic is an overall transmission characteristic of at least one data stream transmitted by a target terminal.
  • control device is further used for:
  • the target difference degree is obtained based on a degree of difference between at least one fifth transmission characteristic among the plurality of fifth transmission characteristics and a corresponding fourth transmission characteristic.
  • control device is further used for:
  • the plurality of sixth transmission characteristics are in one-to-one correspondence with the plurality of seventh transmission characteristics, and one sixth transmission characteristic is an overall transmission characteristic of at least one data stream transmitted by a target terminal.
  • control device is further used for:
  • Acquire transmission information of at least one second data stream transmitted by the at least one second terminal associate and store the transmission information with the terminal type of the second terminal; acquire the transmission information associated with the terminal type based on the transmission information associated with the terminal type. multiple transmission characteristics of the at least one second terminal;
  • the transmission information is used to indicate the transmission attribute of the data stream, and one transmission characteristic is the overall transmission characteristic of at least one data stream transmitted by the terminal.
  • the first transmission characteristic includes an uplink transmission characteristic
  • the uplink transmission characteristic is an overall transmission characteristic of at least one uplink data stream in the at least one first data stream.
  • the upstream transmission feature includes at least one of an upstream message feature and an upstream feature
  • the upstream message feature is an overall feature of an upstream message in the at least one upstream data stream
  • the upstream characteristics are statistical characteristics of the at least one upstream data stream.
  • the characteristics of the uplink packets include the average transmission interval of the uplink packets, the average value of the uplink load, the total uplink load size, the number of uplink packets, the number of uplink target packets, and the uplink target packets.
  • At least one of a ratio and a load fluctuation value of an uplink packet the average transmission interval of the uplink packet is the average transmission interval of the uplink packet in a time window
  • the average value of the uplink load is the average transmission interval of the uplink packet in the time window
  • the average size of the load of the target packets in the at least one upstream data flow, the total upstream load size is the total size of the load of the target packets in the at least one upstream data flow in the time window
  • the The number of upstream packets is the number of upstream packets in the at least one upstream data stream within the time window
  • the number of upstream target packets is the target number in the at least one upstream data stream within the time window.
  • the number of packets, the proportion of the uplink target packets is the proportion of the target packets in the at least one upstream data flow in the time window, and the uplink packet load fluctuation value is used to indicate that in the Fluctuation of the size of the target packet in the at least one upstream data stream within the time window;
  • the upstream characteristics include the upstream terminal port fluctuation value, the total number of upstream data streams, the number of upstream target data streams, the number of upstream data streams under each data stream type in the at least one data stream type, and at least one type of transmission. At least one of the number of upstream data streams under each transmission protocol type in the protocol type, and the upstream terminal port fluctuation value is used to indicate the at least one upstream data stream in the first terminal within the time window.
  • the fluctuation situation of the output port, the upstream target data flow is the upstream data flow whose corresponding server input port belongs to the target port range.
  • the uplink packet feature further includes at least one of a first receiving window fluctuation value and an average value of a first receiving window size, where the first receiving window fluctuation value is used to indicate that the Fluctuation of the size of the receiving window carried by the uplink packet within the time window.
  • the first transmission characteristic further includes at least one of the total number of the at least one first data stream and a downlink transmission characteristic
  • the downlink transmission characteristic is the at least one first data stream.
  • the downlink transmission characteristic includes at least one of a downlink packet characteristic and a downlink characteristic
  • the downlink packet characteristic is an overall characteristic of a downlink packet in the at least one downlink data stream
  • the downstream characteristics are statistical characteristics of the at least one downstream data stream.
  • the characteristics of the downlink packets include the average transmission interval of downlink packets, the average value of downlink loads, the size of the total downlink load, the number of downlink packets, the number of downlink target packets, and the downlink target packets.
  • the average transmission interval of the downlink packets is the average transmission interval of the downlink packets in a time window
  • the average downlink load is the average transmission interval of the downlink packets in the time window the average size of the load of the target packets in the at least one downstream data stream
  • the total downstream load size is the total size of the load of the target packets in the at least one downstream data stream within the time window
  • the The number of downlink packets is the number of downlink packets in the at least one downlink data stream in the time window
  • the number of downlink target packets is the target number in the at least one downlink data stream in the time window.
  • the number of packets, the proportion of the downlink target packets is the proportion of the target packets in the at least one downstream data stream within the time window, and the downlink packet load fluctuation value is used to indicate that in the time window Fluctuation of the size of the target packet in the at least one downstream data stream within the time window;
  • the downstream characteristics include a downlink terminal port fluctuation value, the total number of downstream data streams, the number of downstream target data streams, the number of downstream data streams under each data stream type in at least one data stream type, and at least one type of transmission. At least one of the number of downlink data streams under each transmission protocol type in the protocol type, and the downlink terminal port fluctuation value is used to indicate the at least one downlink data stream in the first terminal within the time window.
  • the fluctuation situation of the input port, the downlink target data flow is the downlink data flow whose corresponding server output port belongs to the target port range.
  • the downlink packet feature further includes at least one of a second receiving window fluctuation value and an average value of a second receiving window size, where the second receiving window fluctuation value is used to indicate the downlink Fluctuation of the receive window size carried by the packet.
  • the network device is further used for:
  • the first transmission characteristics are obtained based on the streaming characteristics of the at least one first data stream.
  • the stream transmission characteristic of a first data stream includes at least one of transmission information, data stream type, destination port type, and packet characteristics of the first data stream, and the transmission information is used
  • the target port type is the port type of the port in the server that transmits the first data stream
  • the message feature is the feature of the message in the first data stream.
  • the transmission information includes at least one element in a five-tuple of the first data stream
  • the message characteristics include at least one of the sum of message transmission intervals, the load size, the sum of squares of the loads, the number of target messages, the total number of messages, the fluctuation value of the receiving window, the total size of the window, and the sum of squares of the window.
  • the sum of packet transmission intervals is the total duration of transmission intervals between packets in the first data stream within a time window
  • the load size is the target packet in the first data stream within the time window
  • the total size of the load, the square sum of the load is the square sum of the load size of the target packet
  • the number of target packets is the total number of target packets in the packets within the time window
  • the total number of packets is the total number of packets in the time window
  • the receiving window fluctuation value is used to indicate the fluctuation of the receiving window size carried by the packet in the time window
  • the total size of the window is the sum of the sizes of the receiving windows carried by the message within the time window
  • the sum of squares of the windows is the sum of the squares of the sizes of the sliding windows.
  • the transmission information further includes at least one of a direction identifier and an identifier of the time window, where the direction identifier is used to indicate a transmission direction of the first data stream.
  • an apparatus for realizing terminal authentication which is used for executing the above-mentioned method for realizing terminal authentication.
  • the apparatus for implementing terminal verification includes a functional module for implementing the method for implementing terminal verification provided in the above-mentioned first aspect or any optional manner of the above-mentioned first aspect.
  • an apparatus for realizing terminal authentication which is used for executing the above-mentioned method for realizing terminal authentication.
  • the apparatus for implementing terminal verification includes a functional module for implementing the method for implementing terminal verification provided in the above second aspect or any optional manner of the above second aspect.
  • an electronic device in a sixth aspect, includes a processor and a memory, and the memory stores at least one piece of program code, the program code is loaded and executed by the processor to realize the above-mentioned first aspect or the above-mentioned first aspect.
  • the action performed by the method is performed by the method.
  • a computer-readable storage medium is provided, and at least one piece of program code is stored in the storage medium, and the program code is loaded and executed by a processor to realize the above-mentioned first aspect or any one of the above-mentioned first aspects.
  • a computer program product or computer program includes program code
  • the program code is stored in a computer-readable storage medium
  • the processor of the electronic device reads from the computer-readable storage medium Get the program code
  • the processor executes the program code, so that the electronic device executes the first aspect or the method for realizing terminal verification provided in any optional manner of the first aspect, or executes the second aspect or The method for realizing terminal authentication provided in any optional manner of the above second aspect.
  • FIG. 1 is a schematic diagram of a system for realizing terminal verification provided by an embodiment of the present application
  • FIG. 2 is a schematic structural diagram of an electronic device provided by an embodiment of the present application.
  • FIG. 3 is a flowchart of a terminal type identification provided by an embodiment of the present application.
  • FIG. 4 is a schematic diagram of a storage terminal type provided by an embodiment of the present application.
  • FIG. 5 is a flowchart of a method for obtaining a model provided by an embodiment of the present application.
  • FIG. 6 is a schematic diagram of starting training provided by an embodiment of the present application.
  • FIG. 7 is a schematic diagram of a transmission feature acquisition provided by an embodiment of the present application.
  • FIG. 8 is a flowchart of a method for implementing terminal verification provided by an embodiment of the present application.
  • FIG. 9 is a schematic diagram of a method for realizing terminal verification provided by an embodiment of the present application.
  • FIG. 10 is a schematic structural diagram of an apparatus for realizing terminal verification provided by an embodiment of the present application.
  • FIG. 11 is a schematic structural diagram of an apparatus for realizing terminal verification provided by an embodiment of the present application.
  • FIG. 1 is a schematic diagram of a system for implementing terminal verification provided by an embodiment of the present application.
  • the system 100 includes multiple terminals 101 , multiple network devices 102 , multiple servers 103 , and a control device 104 .
  • a terminal 101 is used to send a data stream to the server 103 through the network device 102 .
  • the terminal 101 outputs the data stream to the network device 102, and the network device 102 forwards the data stream output by the terminal 101 to the server 103.
  • the terminal 101 is also the device that outputs the data stream.
  • the server 103 is a device for inputting data streams.
  • the terminal 101 is also used to receive data streams from the server 103 through the network device 102 .
  • the server 103 outputs the data stream to the network device 102, and the network device 102 forwards the data stream output by the server 103 to the terminal 101.
  • the terminal 101 is the device that inputs the data stream
  • the server 103 is the output device. device for data flow.
  • the device that outputs the data stream is denoted as the "first device”, and the device that inputs the data stream is denoted as the "second device”.
  • the data stream output by the terminal 101 is denoted as "upstream data stream”, at this time the terminal 101 is the first device, and the server 103 is the second device; the data stream input by the terminal 101 is denoted as "downstream data stream” Data flow”
  • the terminal 101 is the second device, and the server 103 is the first device.
  • the upstream data stream and the downstream data stream of the terminal 101 are both data streams transmitted by the terminal 101 .
  • the terminal 101 is an IOT terminal, such as a camera, a sound, a printer, an IP phone, an automatic teller machine (automatic teller machine, ATM) or an inquiry machine, and the like.
  • the network device 102 In addition to forwarding the data stream, the network device 102 is also used to collect the transmission characteristics of the terminal 101, and report the collected transmission characteristics of the terminal 101 to the control device 104, and the control device 104 based on the transmission characteristics of the terminal 101, the The terminal 101 performs verification to determine whether the terminal 101 is a normal terminal or an abnormal terminal.
  • a normal terminal is a terminal with normal behavior, and a terminal that legally interacts with the server according to preset rules is deemed to have normal behavior. For example, if the terminal accesses the server to perform the preset service within the time period specified by the preset rule, or if the terminal legally interacts with the server within the scope of authority specified by the preset rule, the behavior of the terminal is normal.
  • the terminal is a normal terminal.
  • An abnormal terminal is a terminal with abnormal behavior. The terminal does not interact with the server according to the preset rules, or the interaction between the terminal and the server is illegal, and the behavior is regarded as abnormal.
  • the terminal impersonates a terminal with the authority specified by the preset rules, and illegally interacts with the server, or if the terminal illegally interacts with the server outside the authority specified by the preset rules, the behavior of the terminal is abnormal, the The terminal is an abnormal terminal.
  • the illegal interaction between the terminal and the server includes the terminal illegally accessing specific content in the server, or the terminal accessing the server outside the time period specified by the preset rules, or the terminal being hacked as a springboard, maliciously attacking the network, or the terminal conducting illegal business, etc. various situations.
  • the transmission characteristics collected by the network device 102 are actual transmission characteristics of the terminal 101 and can reflect the behavior of the terminal 101 .
  • the control device 104 has the ability to reconstruct the actual transmission characteristics of the terminal 101, and the control device 104 reconstructs the actual transmission characteristics of the terminal 101 to obtain the reconstructed transmission characteristics, where the reconstructed transmission characteristics are The reconstructed actual transmission characteristic; if the difference between the actual transmission characteristic and the reconstructed transmission characteristic is less than the target difference, it means that the difference between the actual transmission characteristic and the reconstructed transmission characteristic is small, and the control device 104 determines that the terminal behavior embodied by the actual transmission characteristic is normal, the actual transmission characteristic is a normal transmission characteristic, and determines that the terminal 101 is a normal terminal, then the control device 104 passes the verification of the terminal 101; The difference degree between the reconstructed transmission characteristics is greater than or equal to the target difference degree, indicating that the difference between the actual transmission characteristics and the reconstructed transmission characteristics is relatively large, and the control device 104 determines that the terminal behavior embodied by the
  • the control device 104 reconstructs the actual transmission characteristics based on the target model. Before reconstructing the actual transmission feature based on the target model, the control device 104 obtains the target model through training.
  • the network device 102 collects multiple transmission characteristics of at least one normal terminal 101 under the same terminal type, and sends the multiple transmission characteristics to the control device 104, and the control device 104 uses the multiple transmission characteristics as an initial model.
  • the input and output are trained to obtain the target model, so that the subsequent control device 104 reconstructs the actual transmission characteristics of the terminal 101 under the terminal type based on the target model.
  • the control device 104 can also display the training progress, so that the user can learn the training progress.
  • the network device 102 described above is responsible for collecting transmission characteristics, and the control device 104 trains a target model based on the transmission characteristics collected by the network device 102, and verifies the terminal 101 based on the target model.
  • the network device 102 is responsible for collecting transmission characteristics, and the control device 104 trains at least one target model based on the transmission characteristics collected by the network device 102, each target model corresponds to a terminal type, and Send the at least one target model to the network device 102, and the network device 102, after collecting the transmission characteristics of any terminal 101, determines the any terminal 101 in the at least one target model based on the terminal type of the any terminal 101 The network device 102 reconstructs the transmission characteristics of any terminal 101 based on the determined target model, so as to verify the any terminal 101.
  • the network device 102 forwards the data stream forwarded by the network device 102 to the control device 104, and the control device 104 collects the transmission characteristics of the terminal 101 based on the data stream forwarded by the network device 102, and performs training, and The terminal 101 is verified based on the trained target model.
  • the network device 102 includes one of a firewall device, a router, and a switch.
  • the network device 102 and the control device 104 described above are two separate electronic devices, and in another possible implementation manner, the network device 102 and the control device 104 are the same electronic device, and the electronic device also has the network device 102 and control device 104 functions.
  • FIG. 2 is a schematic structural diagram of an electronic device provided by an embodiment of the present application.
  • the electronic device 200 is provided as the above-mentioned network device and/or control device.
  • the electronic device 200 may vary greatly due to different configurations or performances. It includes one or more processors 201 and one or more memories 202, wherein the processor 201 is a central processing unit (central processing units, CPU) or other types of processors, and the memory 202 stores at least one One piece of program code, where the at least one piece of program code is loaded and executed by the processor 201 to implement the steps performed by the network device and/or the control device provided by the various method embodiments described below.
  • the electronic device 200 also has components such as a wired or wireless network interface, a keyboard, and an input and output interface for input and output.
  • the electronic device 200 also includes other components for realizing device functions, which are not described here. Repeat.
  • a computer-readable storage medium such as a memory including program codes, and the program codes can be executed by a processor in a terminal to complete the method for realizing terminal authentication in the following embodiments .
  • the computer-readable storage medium is a non-transitory computer-readable storage medium, such as read-only memory (ROM), random access memory (RAM), compact disc read-only memory, CD-ROM), magnetic tapes, floppy disks, and optical data storage devices, etc.
  • the control device can identify the type of the terminal in the terminal verification system to determine the terminal type of the terminal in the terminal verification system.
  • type identification refer to the flowchart of a terminal type identification provided by an embodiment of the present application as shown in FIG. 3 .
  • the control device acquires a type identification instruction, where the type identification instruction is used to indicate the terminal type of the identification terminal.
  • the type identification instruction includes an identification mark, and the identification mark is used to indicate the terminal type of the identification terminal.
  • the type identification instruction is triggered by a user operation.
  • the control device detects that the user performs an operation on the control device for triggering the type identification instruction, the control device is triggered to acquire the type identification instruction.
  • the control devices in the terminal verification system can also be divided into multi-level control devices, namely a central control device and a plurality of regional control devices, wherein the central control device is used to manage multiple regions Control devices, each zone control device is used to manage network devices in at least one network zone.
  • the central control device acquires the type identification instruction based on the user operation, it can also send the acquired type identification instruction to each regional control device, so that each regional control device can receive the type identification instruction from the central control device.
  • the control device in this step 301 is an area control device.
  • each time the control device obtains the type identification instruction the control device performs a process of terminal type identification (that is, the following steps 302-308).
  • the control device receives the type identification instruction only once, and periodically performs the process of terminal type identification after receiving the type identification instruction. This embodiment of the present application does not specifically limit the timing of terminal type identification.
  • the control device sends an information acquisition instruction to at least one network device, where the information acquisition instruction is used to instruct the network device to upload the transmission information of the data stream.
  • the at least one network device is a network device managed by the control device.
  • the transmission information of a data stream is used to indicate the transmission attributes of the data stream, such as transmission address, transmission port, transmission protocol, transmission direction, and transmission time, which are respectively a transmission attribute of the data stream.
  • the information acquisition instruction includes an attribute identifier of at least one transmission attribute, so as to instruct the network device to upload at least one transmission attribute of the data stream based on the attribute identifier of the at least one transmission attribute.
  • the any network device receives the information acquisition instruction.
  • the any network device acquires at least one transmission information of at least one data stream based on the information acquisition instruction.
  • the at least one data stream is a data stream transmitted by the network device within a time window, the duration of the time window is the target duration, and the time window is the detection time for any network device to detect the at least one data stream.
  • the at least one data flow includes at least one upstream data flow, and optionally, the at least one data flow further includes at least one downstream data flow.
  • the transmission information of a data stream includes at least one element in the quintuple of the first data stream, and the quintuple of the first data stream includes source address information, source port identification, destination address information, and destination port identification of the data stream. , Transmission protocol type.
  • the source address information is used to indicate the network address of the first device that outputs the data stream, where the first device is a device that outputs the data stream to the network device, such as a server or a terminal, and the source address information includes the first device's IP address.
  • the source port identifier is used to indicate an output port in the first device for outputting the data stream.
  • the destination address information is used to indicate the IP address of the second device that inputs the data stream, where the second device is the device that inputs the data stream, that is, the device that receives the data stream forwarded by the network device, such as a server or terminal, the destination address information includes the network address of the second device.
  • the destination port identifier is used to indicate an input port in the second device for inputting the data stream.
  • the network address of a device includes at least one of an IP address of the device and a media access control (media access control, MAC) address, and the network address of the first device is also a source network address, such as a source IP address. address, source MAC address.
  • the network address of the second device is also the destination network address, for example, the destination IP address and the destination MAC address. It should be noted that, in the embodiments of the present application, the network address of the device is an IP address as an example for description.
  • the transport protocol type is used to indicate the transport protocol that the data stream follows, such as transmission control protocol (TCP), user datagram protocol (UDP), or internet control message protocol (internet control message protocol). , ICMP).
  • TCP transmission control protocol
  • UDP user datagram protocol
  • IP internet control message protocol
  • ICMP internet control message protocol
  • the transmission information further includes at least one of a direction identifier and an identifier of the time window, the direction identifier includes an uplink identifier or a downlink identifier, and if the direction identifier is an uplink identifier, it indicates the data stream It is an upstream data flow; if the direction identification is a downstream identification, it indicates that the data flow is a downstream data flow.
  • the any network device includes a plurality of ports, which are respectively at least one first input port, at least one first output port, at least one second input port and at least one second output port, wherein the first input port is used for inputting slave
  • the data stream output by the terminal is used to input the upstream data stream; the first output port is used to output the upstream data stream to the server; the second input port is used to input the data stream output from the server, that is, used for input Downstream data stream; the second output port is used for outputting the downstream data stream to the terminal.
  • the any network device acquires the transmission information of the at least one data stream based on the data stream transmitted by the input port or the output port in the any network device.
  • the process of acquiring the transmission information of the at least one data stream by the any network device based on the input port in the any network device and the information acquisition instruction is implemented by the following steps 3041-3043 .
  • Step 3041 For any input port in the at least one first input port or at least one second input port, the any network device obtains the metadata of the multiple packets input by the any input port within a time window .
  • a time window of target duration is set in any network device, and the any network device detects the data stream transmitted by the any network device within the time window to obtain the data stream transmitted by the any network device.
  • Streaming characteristics within the time window wherein the transmission information of the data stream belongs to the streaming characteristics of the data stream.
  • the time window is the time period of the target duration after the any network device receives the information acquisition instruction. For example, if the any network device performs this step 3041 within the target duration after receiving the information acquisition instruction, then, The target duration after the any network device receives the information acquisition instruction is also the time window.
  • the multiple packets are packets in the data flow input in the any input port within the time window.
  • the metadata of a packet includes flow identification information, source MAC address and destination MAC address of the data flow to which the packet belongs.
  • the flow identification information of the data flow is a quintuple, including the Source IP address, source port identifier, destination IP address, destination port identifier, and transmission protocol type.
  • the any network device obtains multiple packets from the data stream input by the any input port, and parses the multiple packets to obtain metadata of the multiple packets.
  • Step 3042 The any network device deduplicates the metadata of the multiple packets to obtain the metadata of at least one packet, where the metadata of the at least one packet is different from each other, and the metadata of the at least one packet is different from each other.
  • Each packet belongs to a different data flow.
  • One or more data streams may be input to any input port within the time window, and some of the multiple packets may belong to the same data stream.
  • any network device may The metadata of multiple packets is deduplicated to obtain the metadata of at least one packet.
  • the multiple packets are respectively packets 1-3, wherein the metadata of the packet 1 is the same as the metadata of the packet 2, and is different from the metadata of the packet 3, and the metadata of the packet 1 and the packet 2 are both belong to data flow 1 and packet 3 belongs to data flow 2, any network device retains the metadata of packet 1 and the metadata of packet 3 by deduplicating the metadata of packets 1-3.
  • Step 3043 For any packet in the at least one packet, the any network device generates transmission information of the data stream to which the any packet belongs based on the information acquisition instruction.
  • the any network device generates transmission information of the data stream to which the any packet belongs based on at least one attribute identifier included in the information acquisition instruction.
  • the any network device will use at least one of the source IP address and the source MAC address in the metadata of any packet, It is determined as the source address information of the data stream, and the any network device determines at least one of the destination IP address, destination MAC address, and destination MAC address in the metadata of any packet as the destination address information of the data stream, and Add the source address information and the destination address information to the transmission information of the data stream; if the information acquisition instruction includes the attribute identifier of the transmission port, the any network device will include the source port identifier in the metadata of any message And the destination port identifier is added to the transmission information of the data stream; if the information acquisition instruction includes the attribute identifier of the transmission protocol, then the any network device adds the transmission protocol type in the metadata of any message to the data.
  • the any network device In the transmission information of the stream; if the information acquisition instruction includes the attribute identifier of the transmission direction, then the any network device is based on the port used for transmitting the any packet in the any network device or the source address information of the data stream, Determine the transmission direction of the data stream, and add the direction identifier used to indicate the transmission direction to the transmission information of the data stream; if the information acquisition instruction carries the attribute identifier of the transmission time, the time The identification of the window is added to the transmission information of the data stream.
  • the process of determining the transmission direction of the data stream by the any network device based on the port used for transmitting the any packet or the source address information of the data stream in the any network device is as follows: : If the port used to input any message in any network device is the first input port, or the port used to output any message in any network device is the first output port, then the data The stream is an upstream data stream, and the transmission direction of the data stream is determined by the any network device as upstream; if the port used to input the any packet in the any network device is the second input port, or the any network device The internal port used to output any message is the second output port, then the data stream is a downlink data stream, and the any network device determines that the transmission direction of the data stream is downlink; if the source address information of the data stream contains If the indicated first device is a terminal, the data stream is an upstream data stream, and any network device determines that the transmission direction of the data stream is upstream; if the second device indicated by the source address information of
  • any network device transmits a total of 3 data streams, which are data streams 1-3, wherein, data stream 1 and data stream 2 are upstream data streams, and data stream 3 is downstream data flow, through this step 3043, any network device can obtain the transmission information of the data flow 1-3 shown in Table 1.
  • the process shown in this step 304 is that any network device obtains the at least one first data stream transmitted by the second terminal.
  • the any network device sends the at least one transmission information to the control device.
  • the control device receives the at least one transmission information.
  • the process shown in this step 306 is also the process of the control device acquiring the transmission information of the at least one second data stream transmitted by the second terminal .
  • the control device stores the at least one transmission information.
  • the control device extracts at least one item of information from each transmission information, and stores the at least one item of information in a configuration table to realize asset identification.
  • the control device stores each item of information in each transmission information in a configuration table, and the configuration table is shown in Table 1.
  • the control device stores part of information in each transmission information into a configuration table, where the part of information includes a direction identifier and a network address of the terminal.
  • the control device associates and stores the network address of the terminal in at least one transmission information including the same direction identifier and the direction identifier. If the network address of the terminal is an IP address, the control device associates and stores the direction identifier with the IP address of the terminal.
  • the direction identifier in a transmission information is an uplink identifier
  • the data stream corresponding to the transmission information is uplink data flow
  • the source IP address in the transmission information is the IP address of the terminal
  • the control device adds the source IP address to the configuration table, and corresponds to the uplink identifier
  • the direction identifier in a transmission information is the downlink identifier
  • the control device adds the destination IP address to the configuration table and corresponds to the downlink identifier.
  • the control device extracts the source IP address corresponding to the uplink identifier and the destination IP address corresponding to the downlink identifier in Table 1, and associates the extracted source IP address with the uplink identifier and stores it In the configuration table shown in Table 2, the extracted destination IP address and the downlink identifier are associated and stored in the configuration table shown in Table 2.
  • the control device associates and stores the any transmission information with the terminal type of the first target terminal, wherein the first target terminal is used to transmit the any transmission The data stream corresponding to the information.
  • Terminal types include camera, audio, printer, IP phone, ATM or inquiry machine.
  • the control device stores at least one piece of information in any transmission information in association with the terminal type of the target terminal.
  • the control device stores the at least one item of information in association with the terminal type of the first target terminal in an asset table, where the asset table is used to store a variety of terminal types, and the asset table is stored in the local or local of the control device. cloud space.
  • the control device obtains at least one piece of information in any of the transmission information from the configuration table, and based on the at least one piece of information, inquires whether there is a terminal type corresponding to the at least one piece of information in the asset table.
  • the terminal type corresponding to the at least one item of information is the terminal type of the first target terminal, and the control device has stored the at least one item of information in association with the terminal type of the first target terminal, so the control device does not need to perform this step 308 again, If not, the control device stores the at least one item of information in the asset table in association with the terminal type of the first target terminal.
  • the control device displays the terminal type addition information in the terminal type addition interface to prompt the user to add the terminal type in the asset table.
  • the terminal type corresponding to the at least one item of information wherein the terminal type addition information is used to indicate that the terminal type corresponding to the at least one item of information is added in the asset table, and the terminal type addition information includes the at least one item of information and an addition identifier , and the added identifier is used to indicate the terminal type corresponding to the at least one item of information added in the asset table.
  • the user determines the terminal type of the first target terminal based on at least one item of information in the terminal type addition information, and adds the terminal type of the first target terminal in the terminal type addition interface.
  • the control device detects that the user is in the terminal After the adding operation is performed on the type interface, the trigger control device stores the at least one item of information in the asset table in association with the terminal type added by the user.
  • the network device will match the terminal type of the terminal indicated by the any IP address with the terminal type of the terminal.
  • This arbitrary IP address association is stored in the asset table. Still taking the configuration table shown in Table 2 as an example, if the terminal indicated by the IP address 192.168.1.240 is the camera, the terminal indicated by the IP address 192.168.1.180 is the camera, and the terminal indicated by the IP address 192.168.1.150 is the speaker, then The control device associates the three IP addresses with the corresponding terminal types and stores them in the asset table to obtain Table 3.
  • the process shown in this step 308 is that the control device associates the at least one transmission information of the at least one second data stream with the data of the second terminal.
  • the network device acquires the transmission information of the transmitted data stream, and sends the acquired transmission information to the control device, and the control device stores the received transmission information, and queries whether the asset table exists. If the terminal type corresponding to the transmission information does not exist, the control device prompts the user to add the terminal type corresponding to the transmission information in the asset table, and the control device then adds the terminal type corresponding to the transmission information in the asset table based on the user's operation.
  • the method shown in the embodiment of the present application uploads the transmission information of the data stream to the control device through the network device, so that the control device associates and stores the transmission information of each data stream transmitted by the terminal in the terminal verification system with the terminal type, so as to reach the statistical terminal The purpose of validating the terminal type of the terminal in the system.
  • the network device uploads the transmission information of the data stream to the control device based on the information acquisition instruction, and the control device stores the transmission information uploaded by the network device.
  • the control device does not send the information acquisition instruction to the network device, but sends the data stream acquisition instruction to the network device.
  • the network device After the network device receives the data stream acquisition instruction, the network device will One copy of multiple messages in at least one data stream transmitted in the window is uploaded to the control device, and the control device obtains the transmission of the at least one data stream based on the multiple messages in the at least one data stream uploaded by the network device information, and perform steps 307-308.
  • step 304, step 307 and step 308 are periodically executed by the network device.
  • the control device acquires from at least one network device multiple transmission characteristics of the terminal under the any terminal type, and performs training based on the acquired multiple transmission characteristics to obtain the any The target model corresponding to the terminal type.
  • the process refer to the flowchart of a model acquisition method provided by an embodiment of the present application shown in FIG. 5 .
  • the control device determines the target terminal type corresponding to the target model to be acquired.
  • the control device determines the target terminal type based on a user operation.
  • the user inputs the target terminal type in the training interface of the control device, and performs a confirmation operation to realize the process of starting the training.
  • the interface obtains the target terminal type.
  • the control device sends a transmission characteristic acquisition instruction to at least one network device, where the transmission characteristic acquisition instruction is used to instruct to acquire transmission characteristics of the terminal under the target terminal type.
  • the transmission characteristic acquisition instruction includes the target terminal type and at least one network address corresponding to the target terminal type, each network address corresponds to a terminal, and the transmission characteristic of a terminal is the overall transmission characteristic of at least one data stream transmitted by the terminal.
  • the control device queries the asset table for at least one network address corresponding to the target terminal type, forms a terminal address list with the at least one network address found, and adds the terminal address list to the terminal address list. in the transfer characteristic acquisition instruction.
  • the terminal address list is used to record the target terminal type and at least one network address corresponding to the target terminal type. Taking at least one network address corresponding to the target terminal type as the IP address of the terminal as an example, refer to the terminal address list shown in Table 4.
  • the user configures the target terminal type to be the terminal type corresponding to the target model to be acquired (that is, is to configure the target terminal type for training), the control device queries the asset table based on the target terminal type configured by the user, obtains at least one network address corresponding to the target terminal type, and sends the target terminal type and the at least one network address to the network device to start the training process.
  • the any network device receives a transmission characteristic acquisition instruction.
  • the any network device acquires multiple transmission characteristics of at least one second terminal under the target terminal type.
  • the at least one second terminal is a terminal indicated by at least one network address in the transmission feature instruction.
  • the terminal type of the at least one second terminal is the target terminal type.
  • a transmission characteristic of a terminal is the overall transmission characteristic of at least one data stream transmitted by the terminal, and the transmission characteristic may include characteristics of multiple dimensions, and does not only refer to the characteristics of one dimension.
  • the transmission characteristic is an overall transmission characteristic of the at least one data stream within a time window.
  • the transmission characteristic includes an uplink transmission characteristic
  • the uplink transmission characteristic is an overall transmission characteristic of at least one uplink data stream in the at least one data stream
  • the uplink data stream is a data stream output by the terminal
  • Upstream data flow consists of upstream packets.
  • the upstream transmission feature includes one of an upstream message feature and an upstream feature
  • the upstream message feature is the overall feature of the upstream message in the at least one upstream data stream
  • the upstream feature is the at least one upstream feature.
  • the uplink packet features include the average transmission interval of uplink packets, the average uplink load, the total uplink load size, the number of uplink packets, the number of uplink target packets, the proportion of uplink target packets, and the uplink packets.
  • At least one of the load fluctuation values the average transmission interval of the upstream packets is the average transmission interval of the upstream packets in the at least one upstream data stream in a time window, and the average uplink load is the at least one transmission interval in the time window.
  • the average size of the load of the target packets in the upstream data stream, the total size of the upstream load is the total size of the load of the target packets in the at least one upstream data stream within the time window
  • the number of upstream packets is the time The number of upstream packets in the at least one upstream data stream in the window, the number of upstream target packets in the time window
  • the number of target packets in the at least one upstream data stream, the proportion of the upstream target packets is The proportion of target packets in the at least one upstream data stream within the time window, and the upstream packet load fluctuation value is used to indicate the size fluctuation of the target packets in the at least one upstream data stream within the time window.
  • the target message is a message with a payload, and optionally, the data in the payload is service data.
  • the uplink packet feature further includes at least one of a first receiving window fluctuation value and an average value of the first receiving window size, where the first receiving window fluctuation value is used to indicate the at least one uplink data within the time window. Fluctuation of the size of the receiving window carried by the upstream packet in the flow, and the average size of the first receiving window is the average size of the receiving window carried by the upstream packet in the at least one upstream data stream within the time window.
  • the receiving window is a TCP sliding window.
  • the first receiving window fluctuation value is the standard deviation of the receiving window size carried by the upstream packets in the at least one upstream data stream within the time window.
  • the upstream characteristics include the upstream terminal port fluctuation value, the total number of upstream data streams, the number of upstream target data streams, the number of upstream data streams under each data stream type in at least one data stream type, at least At least one of the number of upstream data streams under each transport protocol type in a transport protocol type.
  • the upstream terminal port fluctuation value is used to indicate the fluctuation of the output port of at least one upstream data stream in the terminal within the time window, that is, the fluctuation of the output port of the upstream data stream transmission in the terminal within the time window .
  • the upstream target data flow is the upstream data flow whose corresponding server input port belongs to the target port range.
  • the upstream terminal port fluctuation value is a variance value of the port identifier of the output port of the at least one upstream data stream.
  • the total number of upstream data streams is the total number of the at least one upstream data stream, and the number of the upstream target data streams is the total number of upstream target data streams in the at least one upstream data stream.
  • the at least one transport data stream type includes a data stream, a broadcast data stream, and a unicast data stream.
  • the at least one transport protocol type includes TCP, UDP, and ICMP.
  • the target port range can be set according to actual requirements. For example, the target port range is the range of well-known ports, and the embodiment of the present application does not specifically limit the target port range.
  • the transmission characteristic further includes at least one of the total number of the at least one data stream and a downlink transmission characteristic, and the downlink transmission characteristic is the totality of the at least one downlink data stream in the at least one data stream Transmission characteristics, the downlink data stream is the data stream input by the terminal, and consists of downlink packets.
  • the downlink transmission characteristic includes at least one of a downlink packet characteristic and a downlink characteristic, the downlink packet characteristic is an overall characteristic of downlink packets in the at least one downlink data stream, and the downlink characteristic is the at least one Statistical characteristics of downstream data streams.
  • the characteristics of the downlink packets include the average transmission interval of downlink packets, the average value of downlink loads, the size of the total downlink load, the number of downlink packets, the number of downlink target packets, the proportion of downlink target packets, and the number of downlink packets.
  • At least one of the load fluctuation values the average transmission interval of the downlink packets is the average transmission interval of downlink packets in the at least one downlink data stream within a time window, and the average downlink load is the at least one downlink packet in the time window.
  • the average size of the target packet load in the data stream, the total downstream load size is the total size of the target packet load in the at least one downstream data stream within the time window, and the number of downlink packets is the time window.
  • the number of downlink packets in the at least one downlink data stream, the number of downlink target packets is the number of target packets in the at least one downlink data stream within the time window, and the proportion of the downlink target packets is in The proportion of target packets in the at least one downlink data stream in the time window, and the downlink packet load fluctuation value is used to indicate the size fluctuation of the target packets in the one downlink data stream in the time window.
  • the downlink packet feature further includes at least one of a second receiving window fluctuation value and an average value of the second receiving window size, where the second receiving window fluctuation value is used to indicate a downlink packet in the at least one downlink data stream.
  • the average size of the second receive window is the average size of the receive window carried by the downlink packets in the at least one downlink data stream within the time window.
  • the fluctuation value of the second receiving window is the standard deviation of the receiving window size carried by the downlink packet in the at least one downlink data stream within the time window.
  • the downstream characteristics include a downlink terminal port fluctuation value, the total number of downstream data streams, the number of downstream target data streams, the number of downstream data streams under each data stream type in at least one data stream type, at least At least one of the number of downstream data streams under each transmission protocol type in a transmission protocol type.
  • the downstream terminal port fluctuation value is used to indicate the fluctuation of the input port of the at least one downstream data stream in the terminal within the time window, that is, the fluctuation of the input port of the downstream data stream transmission in the terminal within the time window condition.
  • the downlink target data stream is a downlink data stream whose corresponding output port of the server belongs to the target port range.
  • the downlink terminal port fluctuation value is the variance value of the port identifier of the input port of at least one downlink data stream in the terminal within the time window.
  • the total number of downlink data streams is the total number of the at least one downlink data stream, and the number of the downlink target data streams is the total number of downlink target data streams in the at least one downlink data stream.
  • any network device After acquiring the transmission feature acquisition instruction, any network device acquires at least one network address (for example, the IP address of at least one second terminal) corresponding to the target terminal type from the transmission feature instruction, and within multiple time windows In any time window, the any network device determines a transmission of the second terminal based on any network address in the at least one network address and the data stream transmitted by the second terminal indicated by the any network address. feature, so that any network device can obtain a transmission feature of the at least one second terminal within the any time window, and within the multiple time windows, the any network device can obtain the at least one first terminal Multiple transmission characteristics of two terminals.
  • at least one network address for example, the IP address of at least one second terminal
  • the any network device determines the data flow of the second terminal based on any network address in the at least one network address and the data stream transmitted by the second terminal indicated by the any network address.
  • a way of transmitting features includes way 1 or way 2.
  • Manner 1 Within the any time window, the any network device determines a transmission characteristic of the second terminal based on each data stream transmitted by the second terminal.
  • Mode 1 can be implemented by the following steps A-C.
  • Step A Within the any time window, the any network device determines, based on the any network address, at least one third data stream transmitted by the second terminal indicated by the network address.
  • the packets in the at least one third data stream are all WAN packets.
  • the network device determines that the data stream to which the any packet belongs is the one transmitted by the second terminal. Data flow; the any network device determines whether the any packet is a WAN packet by comparing the destination IP address and the source IP address of the any packet. If the any packet is a WAN packet, then the any packet is A network device determines that the data stream to which any of the packets belongs is a third data stream transmitted by the second terminal.
  • Step B The any network device acquires the streaming characteristics of each of the first and third data streams in the at least one third data stream, and the streaming characteristics of a third data stream is the transmission characteristics of the third data stream.
  • the stream transmission characteristics of a data stream include at least one of transmission information, data stream type, target port type, and packet characteristics of the data stream.
  • the transmission information is used to indicate the transmission attribute of the data stream, and the transmission information includes source address information (source IP address and/or source MAC address), source port identifier, destination address information (destination IP address and/or source MAC address) of the data stream or destination MAC address), destination port identifier, and at least one of the transmission protocol type.
  • the transmission information further includes at least one of a direction identifier and an identifier of the time window, where the direction identifier is used to indicate the transmission direction of the data stream.
  • the target port type is the port type of the port that transmits the data stream in the server, wherein the port type of the port is divided into a first port type, a second port type and a third port type.
  • the first port type corresponds to The port range is the well-known port range [0, 1024)
  • the port range corresponding to the second port type is the registered port range [1024, 49152)
  • the port range corresponding to the third port type is the private port range [49152, + ⁇ ). It should be noted that the port ranges corresponding to the first port type, the second port type, and the third port type can all be set according to specific implementation scenarios.
  • the port range corresponding to the first port type is [0, 10001)
  • the port range corresponding to the second port type is [10001, 20000)
  • the port range corresponding to the third port type is [20000, + ⁇ ) .
  • the embodiment of the present application does not specifically limit the port intervals corresponding to the first port type, the second port type, and the third port type.
  • the packet characteristic is the characteristic of the packet in the data flow.
  • the packet characteristics include at least one of the sum of packet transmission intervals, the load size, the sum of squares of loads, the number of target packets, the total number of packets, the fluctuation value of the receiving window, the total size of the window, and the sum of squares of the window.
  • the sum of the packet transmission intervals is the total duration of the transmission intervals between the packets in the data stream in a time window
  • the load size is the total size of the load of the target packets in the data stream in the time window
  • the sum of squares of the load is the sum of the squares of the payload of the target packet
  • the number of target packets is the total number of target packets in the packet within the time window
  • the total number of packets is the total number of packets in the time window
  • the receive window fluctuation value is used to indicate the fluctuation of the receive window size carried by the packets in the data stream within the time window
  • the total size of the window is the time window
  • the sum of the size of the receiving window carried by the packet in the data stream, and the sum of squares of the window is the sum of the squares of the size of the sliding window.
  • any network device acquires at least one of the transmission information, data stream type, destination port type, and packet characteristics of the any third data stream, and uses The acquired information is determined as the streaming characteristics of any third data stream.
  • the process of acquiring the transmission information of any third data stream by any network device is the same as the process of acquiring the transmission information of at least one data stream by any network device in step 304.
  • the process of acquiring the transmission information of any third data stream by a network device will not be repeated.
  • the process of acquiring the data stream type of any third data stream by any network device is as follows: the any network device determines, according to the source IP address and destination IP address carried in any packet in any third data stream, to determine Whether the any third data stream is a broadcast data stream or a multicast data stream, and if the any third data stream is neither a broadcast data stream nor a multicast data stream, then the any network device determines the any third data stream The stream is a unicast data stream.
  • the process of acquiring the target port type of the any third data stream by the any network device the any network device acquires the port identifier of the server in the any packet, and determines the port type to which the port identifier of the server belongs as The destination port type.
  • the port identifier of the server is the source port identifier or the destination port identifier of any packet.
  • the process for the any network device to obtain the packet characteristics of the any third data stream is: within the any time window, the any network device records the first packet transmitted by the any third data stream. a time, and the second time from the transmission of the last message, and the time difference between the second time and the first time is determined as the sum of the message transmission intervals; within the any time window, the any network The device counts the total size of the payload of the target packets in any third data stream, and determines the total size as the payload size; within the any time window, each third data stream transmits one target packet , the any network device calculates the load size square of the target packet, so that any network device can obtain at least one load size square of the target packet within the any time window.
  • the sum of the squares of the loads in the at least one target packet is performed to obtain the sum of the squares of the loads; the any network device counts the total number of target packets in the any third data stream in the any time window, Obtain the number of target packets; the any network device counts the total number of packets in the any third data stream in the any time window to obtain the total number of packets; in the any time window, The any network device sums the receiving window sizes carried by the packets in the any third data stream to obtain the total window size; within the any time window, the any network device receives the any third data stream
  • the receiving window size carried by each packet in the stream is squared, and the squares of the receiving window sizes carried by each packet are summed to obtain the sum of squares of the windows; any network device obtains the sum of squares based on the total size of the window and the sum of squares of the windows
  • the standard deviation of the receiving window size carried by the packet, and the standard deviation is determined as the fluctuation value of the receiving window; the sum of the packet
  • the IP addresses of the terminals transmitting the data stream ae are 192.168.1.2, 192.168.1.2, 192.168.1.5, 192.168.1.5, and 192.168.1.6, respectively.
  • the streaming characteristics are shown in Table 5 below.
  • Step C The any network device acquires a transmission characteristic of the second terminal based on the streaming transmission characteristic of the at least one third data stream.
  • This step C is realized by the processes shown in the following steps C1-C3.
  • Step C1 The any network device acquires the upstream transmission characteristics based on the streaming characteristics of at least one upstream data stream in the at least one third data stream.
  • the any network device acquires at least one of the upstream packet characteristics and the upstream characteristics of the at least one upstream data stream based on the streaming characteristics of the at least one upstream data stream (referred to as “at least one streaming characteristics”), and will obtain At least one of the received upstream packet feature and upstream feature is determined as the upstream transmission feature.
  • the process of acquiring the upstream message feature by the any network device based on the at least one stream transmission feature is: the any network device sums the message transmission interval sums in the at least one stream transmission feature, and obtains the first A sum value, the first sum value is the sum of the packet transmission intervals of the upstream packets in the at least one upstream data stream; the any network device sums the total number of packets in the at least one stream transmission characteristic, Obtain the number of uplink packets; the first sum value is divided by the number of uplink packets by the any network device to obtain the average transmission interval of uplink packets; the load in the at least one stream transmission characteristic of the any network device Sum the size to obtain the total uplink load size; the any network device sums the number of target packets in the at least one stream transmission feature to obtain the number of uplink target packets; Divide the number of target packets by the number of uplink packets to obtain the proportion of uplink target packets; the any network device divides the total uplink load size by the number of uplink target packet
  • the process of acquiring the upstream feature by the any network device based on the at least one streaming feature is: the any network device acquires at least one source port identifier from the transmission information of the at least one streaming feature, and the any network device acquires at least one source port identifier from the transmission information of the at least one streaming feature.
  • the any network calculates the variance of the at least one source port identifier, and determine the variance as the upstream terminal port fluctuation value; the any network set the total number of the at least one stream transmission feature as the total number of upstream data streams; the Any network device counts the total number of upstream target data streams in the at least one upstream data stream, and obtains the number of the upstream target data streams; the any network device acquires the data stream type in the at least one stream transmission feature, and based on Acquire the data stream type, and count the number of upstream data streams under each data stream type in the at least one data stream type; the any network device acquires the transmission protocol type in the at least one stream transmission feature, and based on the acquired transmission Protocol type, count the number of upstream data streams under each transmission protocol type in at least one transmission protocol type; any network device determines the upstream terminal port fluctuation value, the total number of upstream data streams, and the number of upstream target data streams , at least one of the number of upstream data streams under each data stream type in at least one data stream
  • At least one upstream data flow includes data flow 1 and data flow 2, data flow 1 is a multicast data flow, data flow 2 is a broadcast data flow, and at least one data flow type includes multicast data flow, broadcast data flow and unicast data flow data stream, the number of upstream data streams under the at least one data stream type is 1, 1, and 0, respectively.
  • at least one upstream data stream includes data stream 1 and data stream 2
  • the transmission protocol type of data stream 1 is TCP
  • the transmission protocol type of data stream 2 is UDP
  • the at least one transmission protocol type includes TCP, UDP and ICMP
  • the number of upstream data streams under at least one transmission protocol type is 1, 1, and 0, respectively.
  • Step C2 The any network device acquires the downlink transmission characteristic based on the stream transmission characteristic of at least one downlink data stream in the at least one third data stream.
  • This step C2 is the same as the process shown in step C1, and this step C2 is not described repeatedly in this embodiment of the present application.
  • Step C3 The any network device determines the uplink transmission characteristic and the downlink transmission characteristic as one transmission characteristic of the second terminal.
  • the any network device determines the uplink transmission characteristic as a transmission characteristic of the second terminal. In another possible implementation manner, the any network device determines the uplink transmission characteristic, the downlink transmission characteristic and the total number of the at least one third data stream as a transmission characteristic of the second terminal.
  • the data streams a and b in Table 5 are the data streams output by the same second terminal, and the arbitrary device is based on the data streams a and b in Table 5.
  • the stream transmission feature of b to obtain a transmission feature of the second terminal.
  • the source IP addresses of the data streams c and d are the same, then the data streams c and d are the data streams output by the same second terminal, and any device obtains the second data stream based on the streaming characteristics of the data streams c and d in Table 5.
  • a transmission characteristic of the terminal Based on the stream transmission characteristics of data stream e in Table 5 and a transmission characteristic of the second terminal indicated by the destination IP address of data stream e, any device finally obtains the transmission characteristics of the three second terminals shown in Table 6.
  • FIG. 7 a schematic diagram of a transmission feature acquisition provided by an embodiment of the present application shown in FIG. 7 .
  • the target duration of a time window in Figure 7 is 1 hour, and each day from 2016-09-22 to 2016-10-12 is divided into 24 time windows.
  • any network device obtains the first A transmission feature of the two terminals, in each time window, when any third data stream transmitted by the second terminal is input to the any network device, the any network device obtains the stream of the any third data stream Transmission characteristics, then any network device acquires a transmission characteristic of the second terminal based on the transmission characteristics of at least one third data stream transmitted in each time window, and can also acquire the terminal type of the second terminal, for example In the time window of 5:00-5:59 on 2016-09-24, any network device inputs a total of 344 third data streams transmitted by the second terminal, then any network device is based on the 344 third data streams.
  • Stream transmission characteristics of the flow collect a transmission characteristic of the second terminal within the time window, and configure the collected transmission characteristics, such as uplink/downlink packet characteristics, uplink/downlink flow characteristics, and the like.
  • the above example uses 1 hour as a time window to count the transmission characteristics, while in some other embodiments, 1 minute is used as a time window to count the transmission characteristics in every minute in real time, and the time window
  • the window duration (that is, the target duration) of the time window can be set according to the specific implementation scenario.
  • the embodiment of the present application does not specifically limit the window duration of the time window.
  • Manner 2 Within the any time window, the any network device determines a transmission characteristic of the second terminal according to the at least one network address.
  • the any network device acquires at least one of the upstream packet characteristics of multiple upstream packets carrying the at least one network address, and the upstream characteristics of the upstream data streams in which the multiple upstream packets are located, The any network device determines at least one of the acquired upstream packet feature and upstream feature as the upstream transmission feature.
  • the process for the any network device to obtain the uplink packet characteristics of the plurality of uplink packets carrying the at least one network address is as follows: The total number of multiple uplink packets of the address is obtained to obtain the number of uplink packets; the any network device records the arrival time of the first packet and the last packet of the multiple uplink packets in the arbitrary time window. Reach time, and divide the time difference between the two arrival times by the target difference to obtain the average transmission interval of uplink packets, where the first packet is the one that carries any of the at least one network address within the any time window.
  • the average number of the first receiving window size is obtained; the average transmission interval of uplink packets, the average uplink load average, the total uplink load size, the number of uplink packets, the number of uplink target packets, the uplink At least one of the target packet ratio, the load fluctuation value of the uplink packet, the fluctuation value of the first receiving window, and the average value of the size of the first receiving window is determined as the characteristic of the uplink packet.
  • the process for any network device to acquire the upstream characteristics of the upstream data streams in which the multiple upstream packets carrying the at least one network address are located is: in the any time window, the any network device Obtain multiple source port identifiers in multiple uplink messages that carry the at least one network address, and the any network device calculates the variance of the multiple source port identifiers, and determines the variance as the uplink terminal port fluctuation value; in the In any time window, the any network device counts the total number of upstream data streams carrying the at least one network address to obtain the total number of upstream data streams; The total number of upstream target data streams in the upstream data streams of the at least one network address is obtained to obtain the number of upstream target data streams; within the any time window, the any network device counts each of the at least one data stream type The number of upstream data streams carrying the at least one network address under the data stream type, to obtain the number of upstream data streams under each data stream type in the at least one data stream type; within the any time window, the any The network
  • Any of the network devices can also refer to the process of acquiring the uplink transmission characteristics in Mode 2 to acquire the downlink transmission characteristics. And referring to the process shown in step C3, a transmission feature of the second terminal is acquired.
  • the any network device determines a transmission characteristic of the second terminal indicated by each network address based on the foregoing manner 1 or manner 2, so that within the multiple time windows , the any network device can acquire multiple transmission characteristics of the at least one second terminal.
  • the any network device sends a plurality of transmission characteristics of the at least one second terminal to the control device.
  • the control device receives multiple transmission characteristics of the at least one second terminal.
  • the control device can respectively receive the at least one second terminal from the at least one network device of multiple transmission characteristics.
  • the control device uses the multiple third transmission features of the at least one second terminal as the input and output of the initial model, and performs training to obtain the target model.
  • the plurality of third transmission characteristics are part of the plurality of transmission characteristics of the at least one second terminal.
  • One transmission characteristic of each second terminal is a sample data, and each second terminal is a normal terminal, when the number of multiple transmission characteristics of the at least one second terminal received by the control device is greater than or equal to the target number , it indicates that the sample data of the normal terminal is sufficient; the control device divides the received multiple transmission characteristics of the at least one second terminal into three parts, and puts the three parts into the training set, the verification set and the test set respectively, That is, the training set, the verification set and the test set all include part of the transmission characteristics of the multiple transmission characteristics of the at least one second terminal, wherein each transmission characteristic in the training set is the third transmission characteristic, and each transmission characteristic in the verification set is The fourth transmission characteristic, each transmission characteristic in the test set is the sixth transmission characteristic.
  • the target model is used to reconstruct normal transmission characteristics.
  • the target model is an unsupervised deep neural network model, such as a deep autoencoder model.
  • the transmission feature X restored by the decoder where x n is the nth feature included in the transmission feature X, such as the total number of upstream data streams, and n is the total dimension of the transmission feature X or the feature in the transmission feature.
  • the total number, n>1, x' n is the reconstructed x n .
  • the control device inputs a plurality of third transmission features into the i-th model, and the i-th model reconstructs each input third transmission feature, and outputs the reconstructed transmission feature of each third transmission feature.
  • the control device inputs a plurality of third transmission features and corresponding reconstructed transmission features into the target loss function, and calculates the value of the target loss function; if the value of the target loss function is greater than or equal to the preset threshold, and i ⁇ q, the control Based on the optimization algorithm, the device continues to iterate to update the model parameters of the i-th model to obtain the i+1-th model, and the control device enters the i+1-th training process; if the objective loss function value is less than the preset threshold, the control device ends Training, the i-th model is determined as the target model; if the target loss function value is greater than or equal to the preset difference and i ⁇ q, then the control device ends the training, so as to avoid the control device in the case that the target model cannot be trained The iterative training is continued, or if the change of the model parameters of the model between two adjacent iterations is less than a preset model parameter change value, the control device ends the training.
  • the optimization algorithm includes a gradient descent algorithm, and the objective loss function is shown in the following formula (1), where m is the total number of multiple third transmission features, is the value of the i-th dimension in the j-th third transmission feature, is reconstructed from the target model. n ⁇ i>0, m ⁇ j>0.
  • control device can also display training progress information, and the training progress information includes the current number of training times and the value of the target loss function, so that the user can obtain through the training progress information. training progress.
  • the control device inputs multiple fourth transmission features of multiple target terminals under the target terminal type into the target model, and the target model reconstructs each fourth transmission feature based on the inputted multiple fourth transmission features, A plurality of fifth transmission characteristics are output, the plurality of fourth transmission characteristics are in one-to-one correspondence with the plurality of fifth transmission characteristics, and one fourth transmission characteristic is an overall transmission characteristic of at least one data stream transmitted by a target terminal.
  • the multiple target terminals include at least one second terminal and at least one third terminal, wherein the at least one second terminal is a normal terminal, and the at least one third terminal is an abnormal terminal.
  • a fifth transmission characteristic is a reconstructed fourth transmission characteristic.
  • the verification set also includes multiple transmission characteristics of at least one third terminal.
  • the transmission characteristics of the second terminal and the transmission characteristics of the third terminal are both regarded as the first terminal.
  • the control device obtains a plurality of fourth transmission features from the verification set, and inputs the obtained plurality of fourth transmission features into the target model, and the target model outputs the reconstructed plurality of fourth transmission features based on the plurality of input fourth transmission features Fourth transmission characteristics, wherein the plurality of fourth transmission characteristics belong to at least one second terminal and at least one third terminal, and the plurality of reconstructed fourth transmission characteristics are also the plurality of fifth transmission characteristics, and each The fifth transmission characteristic corresponds to a fourth transmission characteristic.
  • the control device acquires the target difference degree based on the plurality of fifth transmission characteristics and the corresponding fourth transmission characteristics.
  • the control device obtains the target difference based on the degree of difference between at least one fifth transmission characteristic among the plurality of fifth transmission characteristics and the corresponding fourth transmission characteristic, and the difference between one fifth transmission characteristic and the corresponding fourth transmission characteristic
  • the degree of difference is used to indicate the difference between the one fifth transmission characteristic and the corresponding fourth transmission characteristic, wherein the fourth transmission characteristic corresponding to the at least one fifth transmission characteristic is the transmission characteristic of the second terminal .
  • this step 509 is implemented by the processes shown in the following steps 5091-5093.
  • Step 5091 For any fourth transmission characteristic of any second terminal in the at least one second terminal, the control device acquires the degree of difference between the any fourth transmission characteristic and the corresponding fifth transmission characteristic.
  • the degree of difference between the any fourth transmission feature and the corresponding fifth transmission feature is the mean square error between the corresponding features in the any fourth transmission feature and the corresponding fifth transmission feature.
  • error, MSE as shown in formula (2), where MSE(X, X') is the mean square error between any of the fourth transmission features and the corresponding feature in the corresponding fifth transmission feature.
  • the control device performs this step 5091 on each fourth transmission feature of each second terminal in the at least one second terminal, so as to obtain a plurality of fourth transmission features and corresponding fifth transmission features of the at least one second terminal difference between.
  • the control device determines a fifth transmission characteristic corresponding to each fourth transmission characteristic of the at least one second terminal as one of the at least one fifth transmission characteristic.
  • the control device further screens multiple fifth transmission characteristics corresponding to multiple fourth transmission characteristics of the at least one second terminal to obtain the at least one fifth transmission characteristic, For details, refer to the following step 5092.
  • Step 5092 Based on the degree of difference between the plurality of fourth transmission characteristics of the at least one second terminal and the corresponding fifth transmission characteristics, the control device selects the fifth transmission characteristics corresponding to the plurality of fourth transmission characteristics of the second terminal. Among the transmission characteristics, at least one fifth transmission characteristic is determined.
  • the degree of difference is If the difference degree is a normal value, the difference degree can participate in the calculation of the target difference degree, and the control device determines the fifth transmission characteristic corresponding to any fourth transmission characteristic as one of the at least one fifth transmission characteristic.
  • the target difference interval includes a plurality of differences, and the plurality of differences are normal values, all of which can participate in the calculation of the target difference.
  • the degree of difference in the target difference interval can be set according to specific conditions.
  • the control device adopts a boxplot or sets a quantile to determine the target difference interval.
  • the determined target difference interval is [0, 0.1]. Therefore, the scope of the target difference interval is not specifically limited in the embodiments of the present application.
  • Step 5093 The control device determines the target difference degree based on the difference degree between the at least one fifth transmission characteristic and the corresponding fourth transmission characteristic.
  • the control device obtains the target degree of difference based on the average value or fraction of at least one degree of difference, wherein the at least one degree of difference includes a degree of difference between the at least one fifth transmission characteristic and the corresponding fourth transmission characteristic,
  • the fractional bits include the median value of the at least one degree of difference.
  • the control device obtains the target degree of difference ⁇ based on the average value of the at least one degree of difference, as shown in formula (3).
  • MSE[ ⁇ ] is the array formed by the at least one difference degree
  • AVE(MSE[ ⁇ ]) is the average value of the at least one difference degree
  • STD(MSE[ ⁇ ]) is the standard of the at least one difference degree
  • the difference is used to indicate the fluctuation of the at least one degree of difference
  • K is the sensitivity coefficient.
  • the target model is not only used to completely reconstruct the normal transmission characteristics, but also used to completely reconstruct the abnormal transmission characteristics. It should be noted that a certain error range is allowed to completely reconstruct the normal transmission characteristics. For example, when the target model is reconstructing an actual transmission feature, if the difference between the reconstructed transmission feature output by the target model and the actual transmission feature is smaller than the target difference, it means that the target model is completely reconstructed If the actual transmission characteristic is a normal transmission characteristic, it means that the reconstruction is accurate; if the actual transmission characteristic is an abnormal transmission characteristic, it means that this reconstruction fails; when the target model is reconstructing a In the case of actual transmission characteristics, if the difference between the reconstructed transmission characteristics output by the target model and the actual transmission characteristics is greater than or equal to the target difference degree, it means that the target model has not completely reconstructed the actual transmission characteristics.
  • the at least one fourth transmission feature is an actual transmission feature
  • the at least one fifth transmission feature is a reconstructed transmission feature
  • the normal transmission feature is a transmission feature of a normal terminal, that is, a transmission feature of at least one second terminal
  • the abnormal transmission characteristic is the transmission characteristic of the abnormal terminal, that is, the transmission characteristic of at least one third terminal.
  • the control device can also reconstruct the exact conditions of the fourth transmission characteristics of the target terminals according to the target model, dynamically adjust the value of K, and determine the target difference degree through the adjusted K. In a possible implementation manner, the control device determines the target difference degree by executing the K value update process r times.
  • the control device K as K j, and based on the formula (3) to calculate a difference degree [theta] j, if the degree of difference in the [theta] j, which corresponds to the first target model condition, the control device determines the degree of difference as the target opening degree ⁇ j difference, K value update ending; otherwise, the control device will be updated to K J K j + 1, and K j + 1 to a K, executed in a first The process of updating the K value for j+1 times.
  • r ⁇ j ⁇ 1 K j is K used in the process of updating the jth K value
  • ⁇ j is the difference degree calculated based on K j and formula (3).
  • the first condition includes at least one of the following: under the degree of difference ⁇ j , the first precision of the target model is greater than or equal to a first precision threshold, and the first recall of the target model is greater than or equal to equal to the first recall threshold; under the difference degree ⁇ j , the second precision of the target model is greater than or equal to the second precision threshold, and the second recall of the target model is greater than or equal to the second Recall threshold.
  • the first precision is the accuracy rate that the target model completely reconstructs the normal transmission features
  • the first recall is the probability that the target model completely reconstructs the normal transmission characteristics
  • the second precision is The target model does not completely reconstruct the accuracy of abnormal transmission features
  • the second recall is the probability that the target model does not completely reconstruct abnormal transmission features.
  • This embodiment of the present application does not specifically limit the first precision rate threshold, the first recall rate threshold, the second precision rate threshold, and the second recall rate threshold.
  • the control device obtains the number of normal transmission characteristics, the first normal number, the first 2. The normal number, the number of abnormal transmission characteristics, the first abnormal number and the second abnormal number; the control device determines the first abnormal number based on the normal transmission characteristic number, the first normal number and the second normal number Precision and first recall; the control device determines the second precision and second recall based on the number of abnormal transmission characteristics, the number of first abnormalities and the number of second abnormalities.
  • An actual transmission characteristic has a first label, and the first label is used to indicate whether the actual transmission characteristic is a normal transmission characteristic. If the first label is the first normal identification, it means that the actual transmission characteristic is a normal transmission characteristic. If the label is the first abnormality identifier, it means that the actual transmission characteristic is an abnormal transmission characteristic.
  • the plurality of fourth transmission characteristics are all actual transmission characteristics, the first labels of the normal transmission characteristics in the plurality of fourth transmission characteristics are the first normal identifiers, and the first labels of the abnormal transmission characteristics in the plurality of fourth transmission characteristics Both are the first exception identifier.
  • the embodiments of the present application do not specifically limit the representations of the first normal identifier and the first abnormal identifier.
  • a reconstructed transmission characteristic has a second label, and the second label is used to indicate whether the actual transmission characteristic corresponding to the reconstructed transmission characteristic is a normal transmission characteristic. If the second label is a second normal identification, it indicates that the reconstruction
  • the actual transmission characteristic corresponding to the transmission characteristic is the normal transmission characteristic, that is, the terminal to which the actual transmission characteristic belongs is a normal terminal; if the second label is the second abnormal identifier, it means that the actual transmission characteristic corresponding to the reconstructed transmission characteristic is Abnormal transmission characteristic, the terminal to which the actual transmission characteristic belongs is an abnormal terminal.
  • the plurality of fifth transmission characteristics are reconstructed transmission characteristics.
  • the degree of difference between the fifth transmission characteristic and the corresponding fourth transmission characteristic is less than the degree of difference ⁇ j , it means that the Under the difference degree ⁇ j , the fourth transmission characteristic corresponding to the fifth transmission characteristic is a normal transmission characteristic, then the second label of the fifth transmission characteristic is the second normal identification, otherwise, it is explained that under the difference degree ⁇ j ,
  • the fourth transmission characteristic corresponding to the fifth transmission characteristic is an abnormal transmission characteristic, and the second label of the fifth transmission characteristic is a second abnormality identifier.
  • the embodiments of the present application do not specifically limit the representations of the second normal identifier and the second abnormal identifier.
  • the number of normal transmission characteristics is the total number of transmission characteristics of the second terminal in the plurality of fourth transmission characteristics, that is, the total number of normal transmission characteristics in the plurality of fourth transmission characteristics.
  • the first normal number is the total number of target normal transmission features in the plurality of fourth transmission features under the difference degree ⁇ j
  • the target normal transmission feature is that the target model in the plurality of fourth transmission features can be accurately reconstructed normal transmission characteristics.
  • the second normal number is the number of the fifth data stream whose second label is the second normal identifier under the difference degree ⁇ j.
  • the number of abnormal transmission characteristics is the total number of transmission characteristics of the third terminal in the plurality of fourth transmission characteristics, that is, the total number of abnormal transmission characteristics in the plurality of fourth transmission characteristics.
  • the first abnormal number is the total number of target abnormal transmission features in the plurality of fourth transmission features under the difference degree ⁇ j
  • the target abnormal transmission feature is the target model reconstruction failure in the plurality of fourth transmission features.
  • Abnormal transmission feature; the second abnormal number is the number of the fifth data stream whose second label is the second abnormal identification under the difference degree ⁇ j.
  • the control device obtains the number of normal transmission characteristics, the first normal number, the first The two normal numbers, the number of abnormal transmission characteristics, the first abnormal number and the second abnormal number include: if the first label of a fourth transmission characteristic is the first normal identification, the fourth transmission characteristic is the normal transmission characteristic, Then the control device determines the number of the first labels of the plurality of fourth transmission characteristics as the number of labels of the first normal identification as the number of normal transmission characteristics; if the first label of a fourth transmission characteristic is the first normal identification, And the second label of the fifth transmission characteristic corresponding to the fourth transmission characteristic is the second normal identification, indicating that the target model reconstructs the fourth transmission characteristic (normal transmission characteristic) is accurate, then the control device
  • the four transmission characteristics are determined as the target normal transmission characteristics, and the control device determines the total number of target normal transmission characteristics in the plurality of fourth transmission characteristics as the first normal number; the control device counts the second label as the second normal identification The number of the fifth data stream, and the counted number is determined as the second normal number; if
  • the characteristic is determined as the target abnormal transmission characteristic, and the control device determines the total number of the target abnormal transmission characteristic in the plurality of fourth transmission characteristics as the first abnormal number; the control device counts the second label as the second abnormal identification. The number of five data streams, and the counted number is determined as the second abnormal number.
  • the control device determines the first precision rate and the first recall rate based on the number of normal transmission characteristics, the first normal number and the second normal number, including: the control device The ratio between the first normal number and the second normal number is determined as the first precision rate, and the control device determines the ratio between the first normal number and the normal transmission characteristic number as The first recall.
  • the control device determining the second precision rate and the second recall rate based on the number of abnormal transmission characteristics, the first abnormal number and the second abnormal number includes: the control device The ratio between the first abnormal number and the second abnormal number is determined as the second precision, and the control device determines the ratio between the first abnormal number and the abnormal transmission feature number as The second recall.
  • the verification set includes 100 fourth transmission characteristics, of which 90 fourth transmission characteristics are normal transmission characteristics, and 10 fourth transmission characteristics are abnormal transmission characteristics, that is, the number of normal data streams is 90, and the number of abnormal data streams is 10.
  • the control device inputs 100 fourth transfer features into the target model, and the target model outputs 100 fifth transfer features.
  • the control device marks the second label of any fifth transmission feature as a second normal identifier to indicate that the fourth transmission feature corresponding to the fifth transmission feature is a normal transmission feature (that is, the predicted first The fourth transmission characteristic corresponding to the fifth transmission characteristic is the normal transmission characteristic), otherwise, the control device marks the fifth label of the fifth transmission characteristic as the second abnormal identification to indicate the fourth transmission corresponding to the fifth transmission characteristic
  • the characteristic is an abnormal transmission characteristic (that is, the fourth transmission characteristic corresponding to the predicted fifth transmission characteristic is an abnormal transmission characteristic).
  • the number of fifth transmission features with the second label as the second normal identifier in the 100 fifth transmission features is 88
  • the number of fifth transmission features with the second label as the second abnormal identifier in the 100 fifth transmission features Take 12 as an example, that is, the second normal number is 88
  • the second abnormal number is 12.
  • the first labels of the 86 fourth transmission characteristics are the first normal identification, indicating that the 86 fourth transmission characteristics are the target normal transmission characteristics
  • the first normal number is 86
  • the first labels of the two fourth transmission characteristics in the 88 fourth transmission characteristics are the first abnormal identification, indicating that the two fourth transmission characteristics are actually abnormal transmission characteristics.
  • the difference degree ⁇ j the result of reconstructing the two fourth transmission features by the target model is inaccurate; among the 12 fourth transmission features corresponding to the 12 fifth transmission features, the first labels of the 8 fourth transmission features are The first abnormality identifier indicates that the 8 fourth transmission characteristics are the target abnormal transmission characteristics, then the number of the first abnormality is 8, and the first label of the 4 fourth transmission characteristics among the 12 fourth transmission characteristics is the first label. A normal flag, indicating that the four fourth transmission characteristics are actually normal transmission characteristics.
  • the first precision rate the first The normal number is 86 / the second normal number is 88
  • the first recall rate the first normal number 86 / the normal data stream number 90
  • the second precision rate the first abnormal number 8 / the second abnormal number Number 12
  • the first recall rate the number of the first abnormality 8/the number of abnormal data streams 10.
  • the predicted fifth transmission feature is also the reconstructed fifth transmission feature.
  • the first precision rate of the target model is greater than or equal to the first precision rate threshold, it means that under the difference degree ⁇ j , the output result of the target model when reconstructing the normal transmission features is basically is accurate.
  • the first recall rate of the target model is greater than or equal to the first recall rate threshold, it means that the target model can completely reconstruct the normal transmission characteristics with a high probability under the difference degree ⁇ j. Therefore, if the The first precision of the target model is greater than or equal to the first precision threshold and the first recall of the target model is greater than or equal to the first recall threshold, indicating that under the degree of difference ⁇ j , the target model It has the function of completely reconstructing the normal transmission characteristics.
  • the second precision rate of the target model is greater than or equal to the second precision rate threshold, it means that under the difference degree ⁇ j , the output result of the target model when reconstructing abnormal transmission features is basically accurate.
  • the second recall rate of the target model is greater than or equal to the second recall rate threshold, indicating that the target model cannot fully reconstruct the abnormal transmission characteristics with a high probability under the difference degree ⁇ j.
  • the control device determines the degree of difference as a target degree [theta] j difference.
  • the target difference degree is set by the user according to experience, and the target difference degree does not need to be determined by the above formula (3) and the updating method.
  • the control device inputs multiple sixth transmission features of multiple target terminals under the target terminal type into the target model, and the target model reconstructs each sixth transmission feature based on the inputted multiple sixth transmission features, A plurality of seventh transmission characteristics are output, the plurality of sixth transmission characteristics are in one-to-one correspondence with the plurality of seventh transmission characteristics, and one sixth transmission characteristic is an overall transmission characteristic of at least one data stream transmitted by a target terminal.
  • the multiple target terminals include at least one second terminal and at least one third terminal, wherein the at least one second terminal is a normal terminal, and the at least one third terminal is an abnormal terminal.
  • a seventh transmission characteristic is a reconstructed sixth transmission characteristic.
  • the test set also includes multiple transmission characteristics of at least one third terminal.
  • the transmission characteristics of the second terminal and the transmission characteristics of the third terminal are both regarded as the first terminal.
  • the control device acquires multiple sixth transmission features from the test set, and inputs the acquired multiple sixth transmission features into the target model, and the target model outputs the reconstructed multiple sixth transmission features based on the input multiple sixth transmission features Sixth transmission characteristics, wherein the multiple sixth transmission characteristics belong to at least one second terminal and at least one third terminal, and the reconstructed multiple sixth transmission characteristics are also the multiple seventh transmission characteristics, and each The seventh transmission characteristic corresponds to a sixth transmission characteristic.
  • the control device determines that the target model has passed the verification.
  • the control device determines whether the target model meets the second condition based on the degree of difference between the plurality of seventh transmission characteristics and the corresponding sixth transmission characteristics, and if the target model meets the second condition, the control device determines the target The model passes the verification, otherwise, the control device determines that the target model fails the verification, and if the target model fails the verification, the control device jumps to execute steps 507-511 until the finally acquired target model can pass the verification.
  • the second condition includes at least one of the following: under the target difference degree, the first precision of the target model is greater than or equal to the third precision threshold, and the first recall of the target model is greater than or equal to The third recall threshold; under the target difference degree, the second recall of the target model is greater than or equal to the fourth recall threshold, and the second recall of the target model is greater than or equal to the fourth recall rate threshold.
  • the embodiments of the present application do not specifically limit the third precision threshold, the third recall threshold, the fourth precision threshold, and the fourth recall threshold.
  • step 5093 The calculation methods of the first precision rate, the first recall rate, the second precision rate, and the second recall rate are described in step 5093, and will not be repeated here.
  • the control device adds the target model to a model library, and assigns a model identifier to the target model, and the model library is used to indicate the target model corresponding to at least one terminal type .
  • Each terminal type corresponds to a target model.
  • the control device can also store the target model in association with the target terminal type.
  • the control device associates the model identifier of the target model with the target terminal type and stores it in a target model list for querying, wherein the target model list uses for recording the target model corresponding to at least one terminal type. For example, if the target terminal type is a camera, and the target model corresponding to the target terminal type is target model 3, the control device stores the camera and target model 3 in the target model list shown in Table 8 in association.
  • the training is performed by the control device to obtain the target model
  • the network device can also be trained to obtain the target model, and the network device performs training to obtain the target model.
  • the process is the same as the process of training the control device, and the process of training the network device is not described here in this embodiment of the present application.
  • the network address corresponding to the target terminal type is delivered to at least one network device through the control device, and each network device collects information indicated by each network address based on each network address delivered.
  • the transmission characteristics of the second terminal are then obtained by the control device based on the multiple transmission characteristics of at least one second terminal uploaded by each network device, and the target model corresponding to the type of the target terminal is obtained by training, so that the subsequent verification based on the target model can be performed. Whether the terminal under the target terminal type is an abnormal terminal.
  • the control device sends the target model to at least one network device, and each network device verifies the terminal under the target terminal type based on the target model.
  • the control device sends a model storage instruction to at least one network device, where the model storage instruction is used to instruct to store the target model.
  • the model storage instruction includes the target model, the target terminal type, and at least one network address corresponding to the target terminal type, the target difference degree, and the storage identifier.
  • the target terminal type is also the terminal type of the second terminal.
  • the at least one network address That is, the network address corresponding to the target terminal type in the configuration table, that is, the network address of at least one second terminal, and the storage identifier is used to indicate storage of the target model.
  • the user inputs the target terminal type in the model update interface of the control device, and when the control device detects that the user performs a confirmation operation on the model update interface, the control device is triggered to execute The following operations: based on the target terminal type input in the model update interface, the control device queries the asset table for the model identifier of the target model corresponding to the target terminal type, and queries the configuration table for at least one network corresponding to the target terminal type address, and obtain the target model corresponding to the model identifier in the model library, the control device adds the target terminal type, at least one network address, target model and storage identifier to the model storage instruction, and sends the information to the at least one network The device sends the model storage command.
  • the any network device receives the model storage instruction.
  • the any network device stores the target model in the model storage instruction.
  • the any network device associates and stores the target model, the target terminal type, the at least one network address, and the target difference degree in the model storage instruction.
  • the any network device acquires a first transmission characteristic of the first terminal, where the first transmission characteristic is an overall transmission characteristic of at least one first data stream transmitted by the first terminal.
  • the terminal type of the first terminal is the target terminal type, and the first terminal is also the verified terminal of the target terminal type.
  • the network address of the first terminal is any one of the at least one network address, that is, the network address of any second terminal in the at least one second terminal is the same as the network address of the first terminal.
  • Each of the at least one first data stream carries the network address of the first terminal.
  • the first transmission characteristic is also a transmission characteristic of the first terminal, that is, the overall transmission characteristic of at least one first data stream transmitted by the first terminal within the one time window.
  • the network address of the first terminal carried in the message transmitted by the first terminal is the network address of a second terminal; if the first terminal is the preset service used terminal, the first terminal is any one of the at least one second terminal.
  • the any network device In order to verify whether the first terminal is an abnormal terminal, the any network device also needs to verify whether the first terminal is an abnormal terminal based on the first transmission characteristic of the first terminal. Therefore, any network device also needs to verify whether the first terminal is an abnormal terminal. Obtain the first transmission characteristic of the first terminal.
  • the any network device determines the at least one data stream as at least one first data stream.
  • a data stream the any network device determines the terminal outputting the at least one first data stream as the first terminal; the any network device determines the first transmission of the first terminal based on the at least one first data stream feature.
  • the process of determining the first transmission feature of the first terminal by any network device based on the at least one first data stream is the same as the process of determining a transmission feature of the second terminal by any network device in the above step 504,
  • the process of determining the first transmission feature of the first terminal based on the at least one first data stream by any network device will not be described repeatedly.
  • the any network device inputs the first transmission characteristic into a target model, and the target model reconstructs the first transmission characteristic based on the inputted first transmission characteristic, and outputs the second transmission characteristic.
  • the target model corresponds to the target terminal type, and the second transmission feature is the first transmission feature reconstructed by the target model.
  • the process shown in this step 805 is also a process in which any network device reconstructs the first transmission characteristic based on the first transmission characteristic to obtain the second transmission characteristic.
  • the any network device determines that the first terminal has not passed the verification, and the target degree of difference is used to indicate the first The difference between the transmission characteristic and the second transmission characteristic.
  • the degree of difference between the first transmission characteristic and the second transmission characteristic is the mean square error between the second transmission characteristic and the corresponding characteristic in the corresponding first transmission characteristic.
  • the any network device calculates, based on the above formula (2), the mean square error between the second transmission characteristic and the corresponding characteristic in the corresponding first transmission characteristic, and determines the mean square error as the difference between the first transmission characteristic and the first transmission characteristic.
  • the degree of difference between the second transmission characteristics the any network device compares the degree of difference between the first transmission characteristic and the second transmission characteristic with the target degree of difference to determine the difference between the first transmission characteristic and the second transmission characteristic.
  • the difference between the two transmission features is greater than or equal to the target difference; if the difference between the first transmission feature and the second transmission feature is greater than or equal to the target difference, it means that the target model has not been completely reconstructed
  • the difference between a transmission characteristic and the second transmission characteristic is less than the target difference, it means that the target terminal completely reconstructs the first transmission characteristic, the first transmission characteristic is a normal transmission characteristic, and the behavior of the first terminal If it is normal, the first terminal is a normal terminal, and the any network device determines that the first terminal has passed the verification.
  • the any network device sends a verification result of the first terminal to the control device, where the verification result is used to indicate whether the first terminal passes the verification.
  • the verification result includes the verification identifier and the network address of the first terminal.
  • the verification identifier is used to indicate whether the first terminal has passed the verification, and the verification identifier includes a first verification identifier or a second verification identifier, wherein the first verification identifier is used to indicate that the first terminal has passed the verification, and the second verification identifier is used for to indicate that the first terminal has not passed the verification. If the first terminal passes the verification, the verification identifier is the first verification identifier, and if the first terminal fails the verification, the verification identifier is the second verification identifier.
  • the control device receives the verification result.
  • control device displays the first prompt message, and sends a disconnection instruction to any network device
  • the first prompt information is used to prompt the user that the first terminal has failed the verification, and the first prompt information includes the second verification identifier, the network address of the first terminal, and a warning identifier, and the warning identifier is used to prompt the user that the first terminal has not passed the verification. If the verification is not passed, it is used to prompt the user that the first terminal is an abnormal terminal.
  • the disconnection instruction is used to instruct any network device to disconnect from the first terminal, the disconnection instruction includes the network address of the first terminal and a disconnection identifier, and the disconnection identifier is used for Instruct any network device to disconnect from the first terminal.
  • the control device is triggered to display the first prompt information, and the disconnection instruction is sent to any network device.
  • the control device is triggered to display the first prompt information, and the user can refer to the first prompt information on the control device.
  • An operation of triggering the disconnection instruction by the user is also performed on the control device, and then the control device is triggered to send the disconnection instruction to any network device.
  • the control device displays second prompt information, and the second prompt information is used to prompt the user that the first terminal has passed the verification, that is, it is used to prompt that the first terminal is a normal terminal, and the second prompt information is used to prompt the user that the first terminal has passed the verification.
  • the second prompt information includes the IP address of the first terminal and the first verification identifier.
  • the any network device receives the disconnection instruction, and disconnects the connection with the first terminal based on the disconnection instruction.
  • the any network device After the any network device receives the disconnection instruction, the any network device obtains the network address from the disconnection instruction, and disconnects the connection with the first terminal indicated by the network address .
  • the any network device can also display the first prompt information to prompt the user that the first terminal has not passed the verification.
  • the any network device if the user performs an operation for triggering the disconnection of the connection with the first terminal on the any network device, the any network device is triggered to directly disconnect the first total segment. There is no need to wait for the control device to issue a disconnection instruction.
  • the terminal is verified by reconstructing the transmission characteristics of the terminal. For example, if the difference between the reconstructed transmission characteristics and the transmission characteristics of the terminal is large, it means that the transmission characteristics of the terminal appear. If the terminal is abnormal, it is determined that the terminal has not passed the verification. Since the terminal has specific normal transmission characteristics, the normal transmission characteristics are not easy to be counterfeited. Therefore, this method can accurately verify various abnormal terminals and improve the verification of the terminal. Therefore, this method can also accurately verify the fake terminal, instead of simply verifying the IP address of the terminal, so as to prevent the fake terminal from passing the verification.
  • the network device When receiving the data stream transmitted by the terminal, the network device obtains the transmission information of the data stream according to the messages in the data stream, and uploads the transmission information to the control device, and the control device extracts the network of the terminal in the transmission information. address (that is, asset identification) to realize the configuration of the terminal, and the user will mark the terminal type of the terminal in the asset table based on the network address of the terminal (for example, associate the network address of the terminal with the terminal type of the terminal) storage).
  • address that is, asset identification
  • the control device sends the network address information (that is, device information) of the terminal corresponding to the target terminal type to the network device, and the network device matches at least one data stream transmitted by the terminal of the corresponding device type based on the delivered network address.
  • the network device obtains the transmission characteristics of the corresponding terminal based on the stream transmission characteristics of the at least one data stream, and sends the transmission characteristics of the corresponding terminal to the sample library in the control device, Based on the transmission characteristics of at least one terminal under the target terminal type in the sample library, the control device trains to obtain a target model corresponding to the target terminal type, and sends the target model to the network device, and the network device based on the target model.
  • the terminal under the terminal type is verified. If the terminal fails the verification, a warning message, which is the first prompt message, is displayed.
  • the user issues an isolation policy to the network device, such as disconnecting the terminal that fails the verification. , after receiving the isolation policy, the network device disconnects the connection with the terminal that has not passed the authentication.
  • FIG. 10 is a schematic structural diagram of an apparatus for realizing terminal verification provided by an embodiment of the present application.
  • the apparatus 1000 includes:
  • a first acquisition module 1001 configured to acquire a first transmission characteristic of a first terminal, where the first transmission characteristic is an overall transmission characteristic of at least one first data stream transmitted by the first terminal;
  • a first reconstruction module 1002 configured to reconstruct the first transmission characteristic based on the first transmission characteristic to obtain a second transmission characteristic, where the second transmission characteristic is the reconstructed first transmission feature;
  • a determination module 1003 configured to determine that the first terminal fails the verification if the degree of difference between the first transmission characteristic and the second transmission characteristic is greater than or equal to a target degree of difference.
  • the first transmission characteristic includes an uplink transmission characteristic
  • the uplink transmission characteristic is an overall transmission characteristic of at least one upstream data stream in the at least one first data stream.
  • the upstream transmission feature includes at least one of an upstream message feature and an upstream feature
  • the upstream message feature is an overall feature of an upstream message in the at least one upstream data stream
  • the upstream feature is a statistical feature of the at least one upstream data stream.
  • the uplink packet characteristics include the average transmission interval of uplink packets, the average uplink load, the total uplink load size, the number of uplink packets, the number of uplink target packets, the proportion of uplink target packets, and the uplink report.
  • at least one of the fluctuation values of the uplink packet, the average transmission interval of the uplink packets is the average transmission interval of the uplink packets in a time window, and the average uplink load is the at least one transmission interval of the uplink packets in the time window.
  • the average size of the load of the target packets in the upstream data stream, the total size of the upstream load is the total size of the load of the target packets in the at least one upstream data stream within the time window
  • the number of the upstream packets is the number of upstream packets in the at least one upstream data stream within the time window
  • the number of upstream target packets is the number of target packets in the at least one upstream data stream within the time window
  • the proportion of the uplink target packets is the proportion of target packets in the at least one upstream data stream within the time window
  • the uplink packet load fluctuation value is used to indicate the Size fluctuations of target packets in at least one upstream data stream
  • the upstream characteristics include the upstream terminal port fluctuation value, the total number of upstream data streams, the number of upstream target data streams, the number of upstream data streams under each data stream type in the at least one data stream type, and at least one type of transmission. At least one of the number of upstream data streams under each transmission protocol type in the protocol type, and the upstream terminal port fluctuation value is used to indicate the at least one upstream data stream in the first terminal within the time window.
  • the fluctuation situation of the output port, the upstream target data flow is the upstream data flow whose corresponding server input port belongs to the target port range.
  • the uplink packet feature further includes at least one of a first receiving window fluctuation value and an average value of the first receiving window size, where the first receiving window fluctuation value is used to indicate the Fluctuation of the receive window size carried by upstream packets.
  • the first receiving window fluctuation value is the standard deviation of the receiving window size carried by the uplink packet within the time window.
  • the first transmission characteristic further includes at least one of the total number of the at least one first data stream and a downlink transmission characteristic
  • the downlink transmission characteristic is at least one of the at least one first data stream The overall transmission characteristics of the downstream data stream.
  • the downlink transmission characteristic includes at least one of a downlink packet characteristic and a downlink characteristic
  • the downlink packet characteristic is an overall characteristic of a downlink packet in the at least one downlink data stream
  • the downlink characteristic is is a statistical feature of the at least one downstream data stream.
  • the characteristics of the downlink packets include the average transmission interval of downlink packets, the average value of downlink loads, the size of the total downlink load, the number of downlink packets, the number of downlink target packets, the proportion of downlink target packets, and the downlink packet size.
  • the average transmission interval of the downlink packets is the average transmission interval of the downlink packets in a time window
  • the average downlink load is the at least one transmission interval of the downlink packets in the time window the average size of the load of the target packets in the downlink data stream
  • the total size of the downlink load is the total size of the load of the target packets in the at least one downlink data stream within the time window
  • the number of the downlink packets is the number of downlink packets in the at least one downlink data stream in the time window
  • the number of downlink target packets is the number of target packets in the at least one downlink data stream in the time window
  • the proportion of downlink target packets is the proportion of target packets in the at least one downlink data stream within the time window
  • the downlink packet load fluctuation value is used to indicate the Size fluctuations of target packets in at least one downstream data stream
  • the downstream characteristics include a downlink terminal port fluctuation value, the total number of downstream data streams, the number of downstream target data streams, the number of downstream data streams under each data stream type in at least one data stream type, and at least one type of transmission. At least one of the number of downlink data streams under each transmission protocol type in the protocol type, and the downlink terminal port fluctuation value is used to indicate the at least one downlink data stream in the first terminal within the time window.
  • the fluctuation situation of the input port, the downlink target data stream is the downlink data stream whose corresponding server output port belongs to the target port range.
  • the downlink message feature further includes at least one of a second receive window fluctuation value and an average value of the second receive window size, where the second receive window fluctuation value is used to indicate the reception carried by the downlink message. Window size fluctuations.
  • the second receiving window fluctuation value is the standard deviation of the receiving window size carried by the downlink packet within the time window.
  • the first obtaining module 1001 is used for:
  • the first transmission characteristics are obtained based on the streaming characteristics of the at least one first data stream.
  • the stream transmission characteristic of a first data stream includes at least one of transmission information, data stream type, destination port type, and packet characteristics of the first data stream, and the transmission information is used to indicate the transmission information of the data stream.
  • the transmission attribute, the target port type is the port type of the port in the server that transmits the first data stream
  • the message feature is the feature of the message in the first data stream.
  • the transmission information includes at least one element in a five-tuple of the first data stream
  • the message characteristics include at least one of the sum of message transmission intervals, the load size, the sum of squares of the loads, the number of target messages, the total number of messages, the fluctuation value of the receiving window, the total size of the window, and the sum of squares of the window.
  • the sum of packet transmission intervals is the total duration of transmission intervals between packets in the first data stream within a time window
  • the load size is the target packet in the first data stream within the time window
  • the total size of the load, the sum of squares of the load is the sum of the squares of the load of the target packet
  • the number of target packets is the total number of target packets in the packets within the time window
  • the total number of packets is the total number of packets in the time window
  • the receiving window fluctuation value is used to indicate the fluctuation of the receiving window size carried by the packet in the time window
  • the total size of the window is the sum of the sizes of the receiving windows carried by the message within the time window
  • the sum of squares of the windows is the sum of the squares of the sizes of the sliding windows.
  • the transmission information further includes at least one of a direction identifier and an identifier of the time window, where the direction identifier is used to indicate the transmission direction of the first data stream.
  • the first reconstruction module 1002 is used for:
  • the first transmission feature is input into a target model, and the target model reconstructs the first transmission feature based on the inputted first transmission feature, and outputs the second transmission feature.
  • the apparatus 1000 further includes:
  • a training module configured to use multiple third transmission features of at least one second terminal under the target terminal type as the input and output of the initial model, and perform training to obtain the target model, where the target terminal type is the first terminal
  • a third transmission characteristic is an overall transmission characteristic of at least one data stream transmitted by a second terminal.
  • the apparatus 1000 further includes:
  • the second reconstruction module is configured to input multiple fourth transmission characteristics of multiple target terminals under the target terminal model into the target model, and the target model reconstructs the multiple fourth transmission characteristics based on the inputted fourth transmission characteristics.
  • the target terminal type is the terminal type of the first terminal
  • the multiple fourth transmission features are in one-to-one correspondence with the multiple fifth transmission features
  • a fourth transmission characteristic is the overall transmission characteristic of at least one data stream transmitted by a target terminal;
  • the second obtaining module is configured to obtain the target difference degree based on the plurality of fifth transmission characteristics and the plurality of fourth transmission characteristics.
  • the second obtaining module is used for:
  • the target difference degree is obtained based on a degree of difference between at least one fifth transmission characteristic among the plurality of fifth transmission characteristics and a corresponding fourth transmission characteristic.
  • the apparatus 1000 further includes:
  • the third reconstruction module is configured to input multiple sixth transmission features of multiple target terminals under the target terminal type into the target model, and the target model reconstructs each sixth transmission feature based on the inputted sixth transmission features.
  • a sixth transmission feature output a plurality of seventh transmission features
  • the target terminal type is the terminal type of the first terminal
  • the plurality of sixth transmission features are in one-to-one correspondence with the plurality of seventh transmission features
  • a sixth transmission characteristic is an overall transmission characteristic of at least one data stream transmitted by a target terminal;
  • the determining module 1003 is further configured to determine that the target model has passed the verification based on the degree of difference between the plurality of seventh transmission characteristics and the corresponding sixth transmission characteristics.
  • the apparatus 1000 further includes:
  • a third acquiring module configured to acquire transmission information of at least one second data stream transmitted by the at least one second terminal, where the transmission information is used to indicate a transmission attribute of the data stream;
  • a storage module configured to store the transmission information in association with the terminal type of the second terminal
  • the third acquiring module is further configured to acquire multiple transmission characteristics of the at least one second terminal based on the transmission information associated with the terminal type, where one transmission characteristic is an overall transmission characteristic of at least one data stream transmitted by the terminal.
  • the apparatus 1000 further includes:
  • the receiving module is used for receiving the target model from the control device.
  • the apparatus 1000 is a control device or a network device.
  • the apparatus 1000 verifies the terminal by reconstructing the transmission characteristics of the terminal. For example, if the difference between the reconstructed transmission characteristics and the transmission characteristics of the terminal is large, it means that the transmission characteristics of the terminal are abnormal, and the terminal is abnormal. If the terminal is an abnormal terminal, it is determined that the terminal has not passed the verification. Since the terminal has specific normal transmission characteristics, the normal transmission characteristics are not easy to be counterfeited. Therefore, the device 1000 can accurately verify various abnormal terminals, which improves the accuracy of terminal verification. , and the counterfeit terminal is a kind of abnormal terminal. Therefore, the device 1000 can also accurately verify the counterfeit terminal, instead of simply verifying the IP address of the terminal to prevent the counterfeit terminal from passing the verification.
  • FIG. 11 is an apparatus for realizing terminal verification provided by an embodiment of the present application, and the apparatus 1100 includes:
  • an acquisition module 1101 configured to acquire a plurality of third transmission characteristics of at least one second terminal of the target terminal type, where a third transmission characteristic is an overall transmission characteristic of at least one data stream transmitted by a second terminal;
  • the training module 1102 is configured to use the plurality of third transmission features as the input and output of the initial model, and perform training to obtain a target model, which is used to reconstruct the transmission characteristics of the verified terminal of the target terminal type , to verify the terminal to be verified, and the transmission characteristic is the overall transmission characteristic of at least one data stream transmitted by the terminal to be verified.
  • the apparatus 1100 further includes:
  • the sending module 1103 is configured to send the target model to the network device.
  • the apparatus 1100 further includes:
  • the first reconstruction module is configured to input a plurality of fourth transmission characteristics of a plurality of target terminals under the target terminal type into the target model, and the target model reconstructs a plurality of fourth transmission characteristics based on the inputted fourth transmission characteristics.
  • Each fourth transmission characteristic is constructed, and a plurality of fifth transmission characteristics are output, and the plurality of fourth transmission characteristics are in one-to-one correspondence with the plurality of fifth transmission characteristics, and one fourth transmission characteristic is at least one transmission characteristic of a target terminal. the overall transmission characteristics of the data stream;
  • the first target obtaining module is further configured to obtain the target difference degree based on the plurality of fifth transmission characteristics and the plurality of fourth transmission characteristics.
  • the first target acquisition module is used for:
  • the target difference degree is obtained based on a degree of difference between at least one fifth transmission characteristic among the plurality of fifth transmission characteristics and a corresponding fourth transmission characteristic.
  • the apparatus 1100 further includes:
  • the second reconstruction module is configured to input a plurality of sixth transmission characteristics of multiple target terminals under the target terminal type into the target model, and the target model reconstructs a plurality of sixth transmission characteristics based on the inputted sixth transmission characteristics.
  • Each sixth transmission feature is constructed, and a plurality of seventh transmission features are output, and the plurality of sixth transmission features are in one-to-one correspondence with the plurality of seventh transmission features, and a sixth transmission feature is at least one transmission feature of a target terminal. the overall transmission characteristics of the data stream;
  • a determination module configured to determine that the target model has passed the verification based on the degree of difference between the plurality of seventh transmission characteristics and the corresponding sixth transmission characteristics.
  • the apparatus 1100 further includes:
  • a second target acquisition module configured to acquire transmission information of at least one second data stream transmitted by the at least one second terminal, where the transmission information is used to indicate a transmission attribute of the data stream;
  • a storage module configured to store the transmission information in association with the terminal type of the second terminal
  • the second target acquisition module is further configured to acquire multiple transmission characteristics of the at least one second terminal based on the transmission information associated with the terminal type, where one transmission characteristic is an overall transmission characteristic of at least one data stream transmitted by the terminal .
  • Embodiments of the present application also provide a computer program product or computer program, where the computer program product or computer program includes computer instructions, where the computer instructions are stored in a computer-readable storage medium, and the processor of the electronic device is obtained from the computer-readable storage medium. After reading the computer instructions, the processor executes the computer instructions, so that the electronic device executes the above-mentioned method for realizing terminal verification.

Landscapes

  • Engineering & Computer Science (AREA)
  • Signal Processing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Computer Security & Cryptography (AREA)
  • Computer Hardware Design (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

一种用于实现终端验证的方法、装置、系统、设备及存储介质,属于通信技术领域。本方法通过重构终端的传输特征,来对终端进行验证,例如若重构出的传输特征与终端的传输特征之间的差异较大,则说明该终端的传输特征出现异常,该终端为异常终端,则确定该终端未通过验证,由于终端具有特定的正常传输特征,正常传输特征不易仿冒,因此,本方法能够准确地验证出各种异常终端,提高了终端验证的准确度,而仿冒终端为异常终端的一种,因此,本方法也能够准确的验证出仿冒终端,而不是通过简单的对该终端的IP地址进行验证的方式,以防对仿冒终端验证通过。

Description

用于实现终端验证的方法、装置、系统、设备及存储介质
本申请要求于2020年07月13日提交的申请号为202010669766.4、发明名称为“实现仿冒终端检测的方法、装置和系统”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
本申请要求于2020年10月31日提交的申请号为202011198953.5、发明名称为“用于实现终端验证的方法、装置、系统、设备及存储介质”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请涉及通信技术领域,特别涉及一种用于实现终端验证的方法、装置、系统、设备及存储介质。
背景技术
在智慧园区、高教、制造、金融等行业,经常出现一些异常终端非法访问服务器的情况。而异常终端的种类有很多,以仿冒终端为例,越来越多的物联网(the internet of things,IOT)终端支持网络接入功能,相比于智能终端(例如电脑、平板、手机等),IOT终端的自身安全防护功能薄弱,极易被仿冒,为了避免仿冒终端带来安全隐患,网络设备一般通过对其接入的IOT终端进行验证,以确定IOT终端是否为仿冒终端。
目前,终端验证的过程一般是:当接收到一个IOT终端输出的数据流后,网络设备从该数据流中的报文内提取该IOT终端的互联网协议(internet protocol,IP)地址,并查询IP地址库,若IP地址库中存储有该IOT终端的IP地址,则该网络设备确定该IOT终端不是仿冒终端,对该IOT终端验证通过,若IP地址库中未存储该IOT终端的IP地址,则网络设备确定该IOT终端是仿冒终端,对该IOT终端不通过验证。
由于IOT终端的IP地址极易被仿冒,若仿冒终端通过仿冒IP地址库中的IP地址,向网络设备输出数据流,那么,网络设备从该仿冒终端输出的数据流中提取到的IP地址为IP地址库中的IP地址,则网络设备就会对该仿冒终端验证通过,可见,网络设备通过上述终端验证的过程,并不能准确地验证出仿冒终端,也即是终端验证的准确度低。
发明内容
本申请提供了一种用于实现终端验证的方法、装置、系统、设备及存储介质,能够提高终端验证的准确度。该技术方案如下:
第一方面,提供了一种用于实现终端验证的方法,所述方法包括:
获取第一终端的第一传输特征;基于所述第一传输特征,对所述第一传输特征进行重构,得到第二传输特征;若所述第一传输特征与所述第二传输特征之间的差异度大于或等于目标差异度,确定所述第一终端未通过验证;其中,所述第一传输特征为所述第一终端传输的至少一个第一数据流的总体传输特征,所述第二传输特征为重构出的所述第一传输特征。
需要说明的是,重构是指重新构造,所述基于所述第一传输特征,对所述第一传输特征进行重构是指:在所述第一传输特征的基础上,通过预设算法再重新构造出一个第二传输特征,重新构造出的所述第二传输特征尽可能与所述第一传输特征一致。
该预设算法用于尽可能重构出正常传输特征,该预设算法包括降维编码和升维解码,其中,降维编码为降低传输特征维度的一种编码方式,升维编码为升高传输特征维度的一种编码方式。可选地,所述基于所述第一传输特征对所述第一传输特征包括:对所述第一传输特征进行降维编码,对降维编码后的所述第一传输特征进行升维解码。可选地,该预设算法由下文的目标模型来表达。
本方法通过重构终端的传输特征,来对终端进行验证,例如若重构出的传输特征与终端的传输特征之间的差异较大,则说明该终端的传输特征出现异常,该终端为异常终端,则确定该终端未通过验证,由于终端具有特定的正常传输特征,正常传输特征不易仿冒,因此,本方法能够准确地验证出各种异常终端,提高了终端验证的准确度,而仿冒终端为异常终端的一种,因此,该本方法也能够准确的验证出仿冒终端,而不是通过简单的对该终端的IP地址进行验证的方式,以防对仿冒终端验证通过。
需要说明的是,正常终端的传输特征也即是终端的正常传输特征,异常终端的传输特征也即是终端的异常传输特征。
在一种可能的实现方式中,所述第一传输特征包括上行传输特征,所述上行传输特征为所述至少一个第一数据流中至少一个上行数据流的总体传输特征。
在一种可能的实现方式中,所述上行传输特征包括上行报文特征、上行流特征中的至少一个,所述上行报文特征为所述至少一个上行数据流中上行报文的总体特征,所述上行流特征为所述至少一个上行数据流的统计特征。
在一种可能的实现方式中,所述上行报文特征包括上行报文平均传输间隔、上行负载平均值、上行总负载大小、上行报文个数、上行目标报文个数、上行目标报文占比、上行报文负载波动值中的至少一个,所述上行报文平均传输间隔为在一个时间窗口内所述上行报文的平均传输间隔,所述上行负载平均值为在所述时间窗口内所述至少一个上行数据流中目标报文的负载的平均大小,所述上行总负载大小为在所述时间窗口内所述至少一个上行数据流中目标报文的负载的总大小,所述上行报文个数为在所述时间窗口内所述至少一个上行数据流中上行报文的个数,所述上行目标报文个数在所述时间窗口内所述至少一个上行数据流中目标报文的个数,所述上行目标报文占比为在所述时间窗口内所述至少一个上行数据流中目标报文的占比,所述上行报文负载波动值用于指示在所述时间窗口内所述至少一个上行数据流中目标报文的大小波动情况;
所述上行流特征包括上行终端端口波动值、上行数据流总个数、上行目标数据流个数、至少一种数据流类型中每种数据流类型下的上行数据流个数、至少一种传输协议类型中每种传输协议类型下的上行数据流个数中的至少一个,所述上行终端端口波动值用于指示在所述时间窗口内所述第一终端中所述至少一个上行数据流的输出端口的波动情况,所述上行目标数据流为所对应的服务器输入端口属于目标端口范围的上行数据流。
在一种可能的实现方式中,所述上行报文特征还包括第一接收窗口波动值、第一接收窗口大小平均值中的至少一个,所述第一接收窗口波动值用于指示在所述时间窗口内所述上行报文携带的接收窗口大小的波动情况。
在一种可能的实现方式中,所述第一接收窗口波动值为在所述时间窗口内所述上行报文携带的接收窗口大小的标准差。
在一种可能的实现方式中,所述第一传输特征还包括所述至少一个第一数据流的总个数、下行传输特征中的至少一个,所述下行传输特征为所述至少一个第一数据流中至少一个下行数据流的总体传输特征。
在一种可能的实现方式中,所述下行传输特征包括下行报文特征、下行流特征中的至少一个,所述下行报文特征为所述至少一个下行数据流中下行报文的总体特征,所述下行流特征为所述至少一个下行数据流的统计特征。
在一种可能的实现方式中,所述下行报文特征包括下行报文平均传输间隔、下行负载平均值、下行总负载大小、下行报文个数、下行目标报文个数、下行目标报文占比、下行报文负载波动值中的至少一个,所述下行报文平均传输间隔为在一个时间窗口内所述下行报文的平均传输间隔,所述下行负载平均值为在所述时间窗口内所述至少一个下行数据流中目标报文的负载的平均大小,所述下行总负载大小为在所述时间窗口内所述至少一个下行数据流中目标报文的负载的总大小,所述下行报文个数为在所述时间窗口内所述至少一个下行数据流中下行报文的个数,所述下行目标报文个数在所述时间窗口内所述至少一个下行数据流中目标报文的个数,所述下行目标报文占比为在所述时间窗口内所述至少一个下行数据流中目标报文的占比,所述下行报文负载波动值用于指示在所述时间窗口内所述至少一个下行数据流中目标报文的大小波动情况;
所述下行流特征包括下行终端端口波动值、下行数据流总个数、下行目标数据流个数、至少一种数据流类型中每种数据流类型下的下行数据流个数、至少一种传输协议类型中每种传输协议类型下的下行数据流个数中的至少一个,所述下行终端端口波动值用于指示在所述时间窗口内所述第一终端中所述至少一个下行数据流的输入端口的波动情况,所述下行目标数据流为对应的服务器输出端口属于目标端口范围的下行数据流。
在一种可能的实现方式中,所述下行报文特征还包括第二接收窗口波动值、第二接收窗口大小平均值中的至少一个,所述第二接收窗口波动值用于指示所述下行报文携带的接收窗口大小的波动情况。
在一种可能的实现方式中,所述第二接收窗口波动值为在所述时间窗口内所述下行报文携带的接收窗口大小的标准差。
在一种可能的实现方式中,所述获取第一终端的第一传输特征包括:
获取所述至少一个第一数据流中每个第一数据流的流传输特征;基于所述至少一个第一数据流的流传输特征,获取所述第一传输特征。
在一种可能的实现方式中,一个第一数据流的流传输特征包括所述第一数据流的传输信息、数据流类型、目标端口类型、报文特征中的至少一个,所述传输信息用于指示数据流的传输属性,所述目标端口类型为服务器中传输所述第一数据流的端口的端口类型,所述报文特征为所述第一数据流中报文的特征。
在一种可能的实现方式中,所述传输信息包括所述第一数据流的五元组中的至少一元;
所述报文特征包括报文传输间隔总和、负载大小、负载平方和、目标报文个数、报文总个数、接收窗口波动值、窗口总大小、窗口平方和中的至少一个,所述报文传输间隔总和为在一个时间窗口内所述第一数据流中报文之间的传输间隔的总时长,所述负载大小为在所述 时间窗内所述第一数据流中目标报文的负载的总大小,所述负载平方和为所述目标报文的负载大小平方和,所述目标报文个数为在所述时间窗口内所述报文中目标报文的总个数,所述报文总个数为在所述时间窗口内所述报文的总个数,所述接收窗口波动值用于指示在所述时间窗口内所述报文携带的接收窗口大小的波动情况,所述窗口总大小为在所述时间窗口内所述报文携带的接收窗口大小的总和,所述窗口平方和为所述滑动窗口的大小的平方和。
在一种可能的实现方式中,所述传输信息还包括方向标识、所述时间窗口的标识中的至少一个,所述方向标识用于指示所述第一数据流的传输方向。
在一种可能的实现方式中,所述基于所述第一传输特征,对所述第一传输特征进行重构,得到第二传输特征包括:
将所述第一传输特征输入目标模型,由所述目标模型基于输入的所述第一传输特征,重构所述第一传输特征,输出所述第二传输特征。
在一种可能的实现方式中,所述将所述第一传输特征输入目标模型之前,所述方法还包括:
将目标终端类型下至少一个第二终端的多个第三传输特征作为初始模型的输入和输出,进行训练,得到所述目标模型,所述目标终端类型为所述第一终端的终端类型,一个第三传输特征为一个第二终端传输的至少一个数据流的总体传输特征。
在一种可能的实现方式中,所述将所述第一传输特征输入目标模型之前,所述方法还包括:
将目标终端类型下多个目标终端的多个第四传输特征输入所述目标模型,由所述目标模型基于输入的所述多个第四传输特征,重构每个第四传输特征,输出多个第五传输特征;基于所述多个第五传输特征与所述多个第四传输特征,获取所述目标差异度;
其中,所述目标终端类型为所述第一终端的终端类型,所述多个第四传输特征与所述多个第五传输特征一一对应,一个第四传输特征为一个目标终端传输的至少一个数据流的总体传输特征。
在一种可能的实现方式中,所述基于所述多个第五传输特征与所述多个第四传输特征,获取所述目标差异度包括:
基于所述多个第五传输特征中至少一个第五传输特征与对应的第四传输特征之间的差异度,获取所述目标差异度。
在一种可能的实现方式中,所述将所述第一传输特征输入目标模型之前,所述方法还包括:
将目标终端类型下多个目标终端的多个第六传输特征输入所述目标模型,由所述目标模型基于输入的所述多个第六传输特征,重构每个第六传输特征,输出多个第七传输特征;基于所述多个第七传输特征与对应的第六传输特征之间的差异度,确定所述目标模型通过验证;
其中,所述目标终端类型为所述第一终端的终端类型,所述多个第六传输特征与所述多个第七传输特征一一对应,一个第六传输特征为一个目标终端传输的至少一个数据流的总体传输特征。
在一种可能的实现方式中,所述将所述第一传输特征输入目标模型之前,所述方法还包括:
获取所述至少一个第二终端传输的至少一个第二数据流的传输信息;将所述传输信息与 所述第二终端的终端类型进行关联存储;基于所述终端类型关联的传输信息,获取所述至少一个第二终端的多个传输特征;
其中,所述传输信息用于指示数据流的传输属性,一个传输特征为终端传输的至少一个数据流的总体传输特征。
在一种可能的实现方式中,所述将所述第一传输特征输入目标模型之前,所述方法还包括:
从控制设备接收所述目标模型。
在一种可能的实现方式中,所述方法的执行主体为控制设备或网络设备。
第二方面,提供了一种用于实现终端验证的方法,所述方法包括:
获取目标终端类型的至少一个第二终端的多个第三传输特征;将所述多个第三传输特征作为初始模型的输入和输出,进行训练,得到目标模型;
其中,一个第三传输特征为一个第二终端传输的至少一个数据流的总体传输特征,所述目标模型用于重构所述目标终端类型的被验证终端的传输特征,以对所述被验证终端进行验证,所述传输特征为所述被验证终端传输的至少一个数据流的总体传输特征。
在一种可能的实现方式中,所述将所述多个第三传输特征作为初始模型的输入和输出,进行训练之后,所述方法还包括:
向网络设备发送所述目标模型。
在一种可能的实现方式中,所述将所述多个第三传输特征作为初始模型的输入和输出,进行训练之前,所述方法还包括:
将所述目标终端类型下多个目标终端的多个第四传输特征输入所述目标模型,由所述目标模型基于输入的所述多个第四传输特征,重构每个第四传输特征,输出多个第五传输特征;基于所述多个第五传输特征与所述多个第四传输特征,获取所述目标差异度;
其中,所述多个第四传输特征与所述多个第五传输特征一一对应,一个第四传输特征为一个目标终端传输的至少一个数据流的总体传输特征。
在一种可能的实现方式中,所述基于所述多个第五传输特征与所述多个第四传输特征,获取所述目标差异度包括:
基于所述多个第五传输特征中至少一个第五传输特征与对应的第四传输特征之间的差异度,获取所述目标差异度。
在一种可能的实现方式中,所述将所述多个第三传输特征作为初始模型的输入和输出,进行训练之前,所述方法还包括:
将所述目标终端类型下多个目标终端的多个第六传输特征输入所述目标模型,由所述目标模型基于输入的所述多个第六传输特征,重构每个第六传输特征,输出多个第七传输特征;基于所述多个第七传输特征与对应的第六传输特征之间的差异度,确定所述目标模型通过验证;
其中,所述多个第六传输特征与所述多个第七传输特征一一对应,一个第六传输特征为一个目标终端传输的至少一个数据流的总体传输特征。
在一种可能的实现方式中,所述获取目标终端类型的至少一个第二终端的多个第三传输特征之前,所述方法还包括:
获取所述至少一个第二终端传输的至少一个第二数据流的传输信息;将所述传输信息与 所述第二终端的终端类型进行关联存储;基于所述终端类型关联的传输信息,获取所述至少一个第二终端的多个传输特征;
其中,所述传输信息用于指示数据流的传输属性,一个传输特征为终端传输的至少一个数据流的总体传输特征。
第三方面,提供了一种用于实现终端验证的系统,所述系统包括控制设备和网络设备;
所述控制设备用于:
获取目标终端类型的至少一个第二终端的多个第三传输特征,一个第三传输特征为一个第二终端传输的至少一个数据流的总体传输特征;
将所述多个第三传输特征作为初始模型的输入和输出,进行训练,得到目标模型;
向所述网络设备发送所述目标模型;
所述网络设备用于:
获取所述目标终端类型的第一终端的第一传输特征,所述第一传输特征为所述第一终端传输的至少一个第一数据流的总体传输特征;
将所述第一传输特征输入所述目标模型,由所述目标模型基于输入的所述第一传输特征,重构所述第一传输特征,输出第二传输特征,所述第二传输特征为重构出的所述第一传输特征;
若所述第一传输特征与所述第二传输特征之间的差异度大于或等于目标差异度,确定所述第一终端未通过验证。
在一种可能的实现方式中,所述控制设备还用于:
将所述目标终端类型下多个目标终端的多个第四传输特征输入所述目标模型,由所述目标模型基于输入的所述多个第四传输特征,重构每个第四传输特征,输出多个第五传输特征;基于所述多个第五传输特征与所述多个第四传输特征,获取所述目标差异度;
其中,所述多个第四传输特征与所述多个第五传输特征一一对应,一个第四传输特征为一个目标终端传输的至少一个数据流的总体传输特征。
在一种可能的实现方式中,所述控制设备还用于:
基于所述多个第五传输特征中至少一个第五传输特征与对应的第四传输特征之间的差异度,获取所述目标差异度。
在一种可能的实现方式中,所述控制设备还用于:
将所述目标终端类型下多个目标终端的多个第六传输特征输入所述目标模型,由所述目标模型基于输入的所述多个第六传输特征,重构每个第六传输特征,输出多个第七传输特征;基于所述多个第七传输特征与对应的第六传输特征之间的差异度,确定所述目标模型通过验证;
其中,所述多个第六传输特征与所述多个第七传输特征一一对应,一个第六传输特征为一个目标终端传输的至少一个数据流的总体传输特征。
在一种可能的实现方式中,所述控制设备还用于:
获取所述至少一个第二终端传输的至少一个第二数据流的传输信息;将所述传输信息与所述第二终端的终端类型进行关联存储;基于所述终端类型关联的传输信息,获取所述至少一个第二终端的多个传输特征;
其中,所述传输信息用于指示数据流的传输属性,一个传输特征为终端传输的至少一个 数据流的总体传输特征。
在一种可能的实现方式中,所述第一传输特征包括上行传输特征,所述上行传输特征为所述至少一个第一数据流中至少一个上行数据流的总体传输特征。
在一种可能的实现方式中,所述上行传输特征包括上行报文特征、上行流特征中的至少一个,所述上行报文特征为所述至少一个上行数据流中上行报文的总体特征,所述上行流特征为所述至少一个上行数据流的统计特征。
在一种可能的实现方式中,所述上行报文特征包括上行报文平均传输间隔、上行负载平均值、上行总负载大小、上行报文个数、上行目标报文个数、上行目标报文占比、上行报文负载波动值中的至少一个,所述上行报文平均传输间隔为在一个时间窗口内所述上行报文的平均传输间隔,所述上行负载平均值为在所述时间窗口内所述至少一个上行数据流中目标报文的负载的平均大小,所述上行总负载大小为在所述时间窗口内所述至少一个上行数据流中目标报文的负载的总大小,所述上行报文个数为在所述时间窗口内所述至少一个上行数据流中上行报文的个数,所述上行目标报文个数在所述时间窗口内所述至少一个上行数据流中目标报文的个数,所述上行目标报文占比为在所述时间窗口内所述至少一个上行数据流中目标报文的占比,所述上行报文负载波动值用于指示在所述时间窗口内所述至少一个上行数据流中目标报文的大小波动情况;
所述上行流特征包括上行终端端口波动值、上行数据流总个数、上行目标数据流个数、至少一种数据流类型中每种数据流类型下的上行数据流个数、至少一种传输协议类型中每种传输协议类型下的上行数据流个数中的至少一个,所述上行终端端口波动值用于指示在所述时间窗口内所述第一终端中所述至少一个上行数据流的输出端口的波动情况,所述上行目标数据流为所对应的服务器输入端口属于目标端口范围的上行数据流。
在一种可能的实现方式中,所述上行报文特征还包括第一接收窗口波动值、第一接收窗口大小平均值中的至少一个,所述第一接收窗口波动值用于指示在所述时间窗口内所述上行报文携带的接收窗口大小的波动情况。
在一种可能的实现方式中,所述第一传输特征还包括所述至少一个第一数据流的总个数、下行传输特征中的至少一个,所述下行传输特征为所述至少一个第一数据流中至少一个下行数据流的总体传输特征。
在一种可能的实现方式中,所述下行传输特征包括下行报文特征、下行流特征中的至少一个,所述下行报文特征为所述至少一个下行数据流中下行报文的总体特征,所述下行流特征为所述至少一个下行数据流的统计特征。
在一种可能的实现方式中,所述下行报文特征包括下行报文平均传输间隔、下行负载平均值、下行总负载大小、下行报文个数、下行目标报文个数、下行目标报文占比、下行报文负载波动值中的至少一个,所述下行报文平均传输间隔为在一个时间窗口内所述下行报文的平均传输间隔,所述下行负载平均值为在所述时间窗口内所述至少一个下行数据流中目标报文的负载的平均大小,所述下行总负载大小为在所述时间窗口内所述至少一个下行数据流中目标报文的负载的总大小,所述下行报文个数为在所述时间窗口内所述至少一个下行数据流中下行报文的个数,所述下行目标报文个数在所述时间窗口内所述至少一个下行数据流中目标报文的个数,所述下行目标报文占比为在所述时间窗口内所述至少一个下行数据流中目标报文的占比,所述下行报文负载波动值用于指示在所述时间窗口内所述至少一个下行数据流 中目标报文的大小波动情况;
所述下行流特征包括下行终端端口波动值、下行数据流总个数、下行目标数据流个数、至少一种数据流类型中每种数据流类型下的下行数据流个数、至少一种传输协议类型中每种传输协议类型下的下行数据流个数中的至少一个,所述下行终端端口波动值用于指示在所述时间窗口内所述第一终端中所述至少一个下行数据流的输入端口的波动情况,所述下行目标数据流为对应的服务器输出端口属于目标端口范围的下行数据流。
在一种可能的实现方式中,所述下行报文特征还包括第二接收窗口波动值、第二接收窗口大小平均值中的至少一个,所述第二接收窗口波动值用于指示所述下行报文携带的接收窗口大小的波动情况。
在一种可能的实现方式中,所述网络设备还用于:
获取所述至少一个第一数据流中每个第一数据流的流传输特征;
基于所述至少一个第一数据流的流传输特征,获取所述第一传输特征。
在一种可能的实现方式中,一个第一数据流的流传输特征包括所述第一数据流的传输信息、数据流类型、目标端口类型、报文特征中的至少一个,所述传输信息用于指示数据流的传输属性,所述目标端口类型为服务器中传输所述第一数据流的端口的端口类型,所述报文特征为所述第一数据流中报文的特征。
在一种可能的实现方式中,所述传输信息包括所述第一数据流的五元组中的至少一元;
所述报文特征包括报文传输间隔总和、负载大小、负载平方和、目标报文个数、报文总个数、接收窗口波动值、窗口总大小、窗口平方和中的至少一个,所述报文传输间隔总和为在一个时间窗口内所述第一数据流中报文之间的传输间隔的总时长,所述负载大小为在所述时间窗内所述第一数据流中目标报文的负载的总大小,所述负载平方和为所述目标报文的负载大小的平方和,所述目标报文个数为在所述时间窗口内所述报文中目标报文的总个数,所述报文总个数为在所述时间窗口内所述报文的总个数,所述接收窗口波动值用于指示在所述时间窗口内所述报文携带的接收窗口大小的波动情况,所述窗口总大小为在所述时间窗口内所述报文携带的接收窗口大小的总和,所述窗口平方和为所述滑动窗口的大小的平方和。
在一种可能的实现方式中,所述传输信息还包括方向标识、所述时间窗口的标识中的至少一个,所述方向标识用于指示所述第一数据流的传输方向。
第四方面,提供了一种用于实现终端验证的装置,用于执行上述用于实现终端验证的方法。具体地,该用于实现终端验证的装置包括用于执行上述第一方面或上述第一方面的任一种可选方式提供的用于实现终端验证的方法的功能模块。
第五方面,提供了一种用于实现终端验证的装置,用于执行上述用于实现终端验证的方法。具体地,该用于实现终端验证的装置包括用于执行上述第二方面或上述第二方面的任一种可选方式提供的用于实现终端验证的方法的功能模块。
第六方面,提供了一种电子设备,该电子设备包括处理器和存储器,该存储器中存储有至少一条程序代码,该程序代码由该处理器加载并执行以实现如上述第一方面或上述第一方面的任一种可选方式提供的用于实现终端验证的方法所执行的操作,或实现如上述第二方面或上述第二方面的任一种可选方式提供的用于实现终端验证的方法所执行的操作。
第七方面,提供了一种计算机可读存储介质,该存储介质中存储有至少一条程序代码,该程序代码由处理器加载并执行以实现如上述第一方面或上述第一方面的任一种可选方式提 供的用于实现终端验证的方法所执行的操作,或实现如上述第二方面或上述第二方面的任一种可选方式提供的用于实现终端验证的方法所执行的操作。
第八方面,提供了一种计算机程序产品或计算机程序,该计算机程序产品或计算机程序包括程序代码,该程序代码存储在计算机可读存储介质中,电子设备的处理器从计算机可读存储介质读取该程序代码,处理器执行该程序代码,使得该电子设备执行上述第一方面或上述第一方面的任一种可选方式提供的用于实现终端验证的方法,或执行上述第二方面或上述第二方面的任一种可选方式提供的用于实现终端验证的方法。
附图说明
图1是本申请实施例提供的一种用于实现终端验证的系统的示意图;
图2是本申请实施例提供的一种电子设备的结构示意图;
图3是本申请实施例提供的一种终端类型识别的流程图;
图4是本申请实施例提供的一种存储终端类型的示意图;
图5是本申请实施例提供的一种模型获取方法的流程图;
图6是本申请实施例提供的一种开启训练的示意图;
图7是本申请实施例提供的一种传输特征获取的示意图;
图8是本申请实施例提供的一种用于实现终端验证的方法的流程图;
图9是本申请实施例提供的一种用于实现终端验证的方法的示意图;
图10是本申请实施例提供的一种用于实现终端验证的装置的结构示意图;
图11是本申请实施例提供的一种用于实现终端验证的装置的结构示意图。
具体实施方式
为使本申请的目的、技术方案和优点更加清楚,下面将结合附图对本申请实施方式作进一步地详细描述。
图1是本申请实施例提供的一种用于实现终端验证的系统的示意图,参见图1,该系统100包括多个终端101、多个网络设备102、多个服务器103以及控制设备104。
一个终端101,用于通过网络设备102向服务器103发送数据流。在一种可能的实现方式中,该终端101向网络设备102输出数据流,由网络设备102向服务器103转发该终端101输出的数据流,此时该终端101也即是输出数据流的设备,服务器103为输入数据流的设备。该终端101还用于通过网络设备102从服务器103接收数据流。在一种可能的实现方式中,服务器103向网络设备102输出数据流,由网络设备102向终端101转发服务器103输出的数据流,此时该终端101为输入数据流的设备,服务器103为输出数据流的设备。
为了便于描述,将输出数据流的设备记为“第一设备”,将输入数据流的设备记为“第二设备”。为了区分数据流的传输方向,将终端101输出的数据流记为“上行数据流”,此时终端101为第一设备,服务器103为第二设备;将终端101输入的数据流记为“下行数据流”此时终端101为第二设备,服务器103为第一设备。其中,终端101的上行数据流和下行数据流均为终端101传输的数据流。在一种可能的实现方式中,终端101为IOT终端,例如摄像头、音响、打印机、IP话机、自动取款机(automatic teller machine,ATM)或查询机等。
网络设备102除了用于转发数据流之外,还用于采集终端101的传输特征,并向控制设 备104上报采集到的终端101的传输特征,由控制设备104基于终端101的传输特征,对该终端101进行验证,以确定该终端101为正常终端还是异常终端。
其中,正常终端为行为正常的终端,终端按照预设规则与服务器进行合法的交互,则视为行为正常。例如,若终端在预设规则所规定的时间段内通过访问服务器,进行预设业务,或者,若终端在预设规则所规定的权限范围内与服务器进行合法交互,则该终端的行为正常,该终端为正常终端。异常终端为行为异常的终端,终端未按照预设规则与服务器交互,或者终端与服务器的交互不合法,则视为行为异常。例如,若终端仿冒具有预设规则所规定的权限的终端,与服务器进行非法交互,或者,若终端在预设规则规定的权限外与服务器之间进行非法交互,则该终端的行为异常,该终端为异常终端。终端与服务器非法交互包括终端非法访问服务器中的特定内容,或者终端在预设规则规定的时间段以外的时间访问服务器,或者终端被黑客攻破作为跳板,恶意攻击网络,或者终端进行了非法业务等各种情况。
网络设备102采集的传输特征为终端101的实际传输特征,且能够体现终端101的行为。可选地,控制设备104具有对终端101的实际传输特征进行重构的能力,控制设备104对终端101的实际传输特征进行重构,得到重构传输特征,其中,重构传输特征也即是重构出的实际传输特征;若该实际传输特征与该重构传输特征之间的差异度小于目标差异度,说明该实际传输特征与该重构传输特征之间的差异较小,该控制设备104确定该实际传输特征所体现的终端行为正常,该实际传输特征为正常传输特征,并确定该终端101是正常终端,则该控制设备104对该终端101通过验证;若该实际传输特征与该重构传输特征之间的差异度大于或等于目标差异度,说明该实际传输特征与该重构传输特征之间的差异较大,该控制设备104确定该实际传输特征所体现的终端行为异常,该实际传输特征为异常传输特征,并确定则该终端101是异常终端,则该控制设备104对该终端101不通过验证。
在一种可能的实现方式中,控制设备104基于目标模型重构实际传输特征。控制设备104基于目标模型重构实际传输特征之前,先通过训练来获取该目标模型。可选地,网络设备102采集同一种终端类型下至少一个正常终端101的多个传输特征,并将该多个传输特征发送至控制设备104,由控制设备104将该多个传输特征作为初始模型的输入和输出,进行训练,得到该目标模型,以便后续控制设备104基于该目标模型重构该终端类型下终端101的实际传输特征。可选地,该控制设备104在进行训练的过程中,还能够显示训练进度,以便用户能够获悉训练进度。
需要说明的是,上述介绍的网络设备102负责采集传输特征,由控制设备104基于网络设备102采集到的传输特征,训练出的目标模型,并基于目标模型对终端101进行验证。而在一种可能的实现方式中,网络设备102负责采集传输特征,由该控制设备104基于网络设备102采集到的传输特征,训练出至少一个目标模型,每个目标模型对应一个终端类型,并将该至少一个目标模型发送至网络设备102,由网络设备102在采集到任一终端101的传输特征后,基于该任一终端101的终端类型,确定该至少一个目标模型中该任一终端101的终端类型所对应的目标模型,该网络设备102并基于确定出的目标模型重构该任一终端101的传输特征,以对该任一终端101进行验证。
而在一种可能的实现方式中,网络设备102将其转发的数据流转发至控制设备104,由控制设备104基于网络设备102转发的数据流,采集终端101的传输特征,并进行训练,并基于训练出的目标模型对终端101进行验证。可选地,网络设备102包括防火墙设备、路由 器和交换机中的一种。
上述介绍的网络设备102和控制设备104是分离的2个电子设备,而在另一种可能的实现方式中,网络设备102和控制设备104为同一个电子设备,该电子设备兼具网络设备102和控制设备104的功能。
图2是本申请实施例提供的一种电子设备的结构示意图,该电子设备200提供为上述的网络设备和/或控制设备,该电子设备200可因配置或性能不同而产生比较大的差异,包括一个或一个以上处理器201和一个或一个以上的存储器202,其中,所述处理器201为中央处理器(central processing units,CPU)或其他类型的处理器,所述存储器202中存储有至少一条程序代码,所述至少一条程序代码由所述处理器201加载并执行以实现下述各个方法实施例提供的网络设备和/或控制设备所执行的步骤。当然,可选地,该电子设备200还具有有线或无线网络接口、键盘以及输入输出接口等部件,以便进行输入输出,该电子设备200还包括其他用于实现设备功能的部件,在此不做赘述。
在示例性实施例中,还提供了一种计算机可读存储介质,例如包括程序代码的存储器,上述程序代码可由终端中的处理器执行以完成下述实施例中的用于实现终端验证的方法。例如,该计算机可读存储介质是非临时计算机可读存储介质,如只读存储器(read-only memory,ROM)、随机存取存储器(random access memory,RAM)、只读光盘(compact disc read-only memory,CD-ROM)、磁带、软盘和光数据存储设备等。
在终端验证系统的初始化过程中,控制设备能够对该终端验证系统中的终端进行类型识别,以确定该终端验证系统中终端的终端类型,为了进一步说明控制设备对该终端验证系统中的终端进行类型识别的过程,参见如图3所示的本申请实施例提供的一种终端类型识别的流程图。
301、控制设备获取类型识别指令,该类型识别指令用于指示识别终端的终端类型。
该类型识别指令包括识别标识,该识别标识用于指示识别终端的终端类型。在一种可能的实现方式中,该类型识别指令由用户操作来触发。当该控制设备检测到用户在该控制设备上执行了用于触发该类型识别指令的操作时,触发该控制设备获取到该类型识别指令。
当终端验证系统的组网较为复杂时,终端验证系统中的控制设备还能够分为多级控制设备,分别为中心控制设备以及多个区域控制设备,其中,中心控制设备用于管理多个区域控制设备,每个区域控制设备用于管理至少一个网络区域中的网络设备。当中心控制设备基于用户操作获取到类型识别指令后,还能够向各个区域控制设备发送获取到的类型识别指令,以便各个区域控制设备均能从该中心控制设备接收到该类型识别指令,此时本步骤301中的该控制设备为区域控制设备。
在一种可能的实现方式中,当该控制设备每获取到一次该类型识别指令后,该控制设备执行一次终端类型识别的过程(也即是下述步骤302-308)。在另一种可能的实现方式中,控制设备仅接收一次类型识别指令,当接收到该类型识别指令后,周期性地执行终端类型识别的过程。本申请实施例对终端类型识别的时机不作具体限定。
302、该控制设备向至少一个网络设备发送信息获取指令,该信息获取指令用于指示网络设备上传数据流的传输信息。
该至少一个网络设备为该控制设备负责管理的网络设备。数据流的传输信息用于指示该数据流的传输属性,例如传输地址、传输端口、传输协议、传输方向、传输时间分别是数据 流的一种传输属性。该信息获取指令包括至少一种传输属性的属性标识,以指示网络设备基于该至少一种传输属性的属性标识,上传数据流的至少一种传输属性。
303、对于该至少一个网络设备中的任一网络设备,该任一网络设备接收该信息获取指令。
304、该任一网络设备基于该信息获取指令,获取至少一个数据流的至少一个传输信息。
该至少一个数据流为该网络设备在一个时间窗口内传输的数据流,该时间窗口的时长为目标时长,该时间窗口为该任一网络设备检测至少一个数据流的检测时间。可选地,该至少一个数据流包括至少一个上行数据流,可选地,该至少一个数据流还包括至少一个下行数据流。一个数据流的传输信息包括该第一数据流的五元组中的至少一元,该第一数据流的五元组包括该数据流的源地址信息、源端口标识、目的地址信息、目的端口标识、传输协议类型。
该源地址信息用于指示输出该数据流的第一设备的网络地址,该第一设备为向该网络设备输出该数据流的设备,例如服务器或者终端,该源地址信息包括该第一设备的IP地址。源端口标识用于指示该第一设备中用于输出该数据流的输出端口。
该目的地址信息用于指示输入该数据流的第二设备的IP地址,该第二设备为输入该数据流的设备,也即是用于接收该网络设备转发的数据流的设备,例如服务器或终端,该目的地址信息包括该第二设备的网络地址。该目的端口标识用于指示该第二设备中用于输入该数据流的输入端口。
需要说明的是,该第一设备和该第二设备中的一个设备为终端,另一个设备为服务器,若该第一设备为终端,则该数据流也即是上行数据流,若该第二设备为终端,则该数据流的也即是下行数据流。可选地,一个设备的网络地址包括该设备的IP地址和媒体存取控制(media access control,MAC)地址中的至少一个,该第一设备的网络地址也即是源网络地址,例如源IP地址、源MAC地址。该第二设备的网络地址也即是目的网络地址,例如目的IP地址、目的MAC地址。需要说明的是,在本申请实施例中以设备的网络地址为IP地址为例进行说明。
该传输协议类型用于指示该数据流所遵循的传输协议,例如传输控制协议(transmission control protocol,TCP)、用户数据报协议(user datagram protocol,UDP)或互联网控制报文协议(internet control message protocol,ICMP)。
在一种可能的实现方式中,该传输信息还包括方向标识、该时间窗口的标识中的至少一个,该方向标识包括上行标识或下行标识,若该方向标识为上行标识,则指示该数据流为上行数据流;若该方向标识为下行标识,则指示该数据流为下行数据流。
该任一网络设备包括多个端口,分别为至少一个第一输入端口、至少一个第一输出端口、至少一个第二输入端口以及至少一个第二输出端口,其中,第一输入端口用于输入从终端输出的数据流,也即是用于输入上行数据流;第一输出端口用于向服务器输出该上行数据流;第二输入端口用于输入从服务器输出的数据流,也即是用于输入下行数据流;第二输出端口用于向终端输出该下行数据流。
该任一网络设备基于该任一网络设备内的输入端口或输出端口传输的数据流,获取该至少一个数据流的传输信息。在一种可能的实现方式中,该任一网络设备基于该任一网络设备内的输入端口以及该信息获取指令,获取该至少一个数据流的传输信息的过程由下述步骤3041-3043来实现。
步骤3041、对于该至少一个第一输入端口或至少一个第二输入端口中的任一输入端口,该任一网络设备在一个时间窗口内获取该任一输入端口输入的多个报文的元数据。
该任一网络设备内设置有目标时长的时间窗口,该任一网络设备在该时间窗口内对该任一网络设备所传输的数据流进行检测,以获取该任一网络设备所传输的数据流在该时间窗口内的流传输特征,其中,数据流的传输信息属于数据流的流传输特征。该时间窗口为该任一网络设备接收到该信息获取指令后的目标时长所在的时间段,例如,该任一网络设备接收到该信息获取指令后的目标时长之内执行本步骤3041,那么,该任一网络设备接收到该信息获取指令后的目标时长也即是该时间窗口。
该多个报文为在该时间窗口内该任一输入端口内输入的数据流中的报文。其中,一个报文的元数据包括该报文所属的数据流的流标识信息、源MAC地址以及目的MAC地址,可选地,数据流的流标识信息为一个五元组,包括该数据流的源IP地址、源端口标识、目的IP地址、目的端口标识以及传输协议类型。
对于该任一输入端口,该任一网络设备在该时间窗口内的多个时间点,从该任一输入端口输入的数据流中分别获取多个报文,并解析该多个报文,得到该多个报文的元数据(metadata)。
步骤3042、该任一网络设备对该多个报文的元数据进行去重,得到至少一个报文的元数据,该至少一个报文的元数据互不相同,且该至少一个报文中的每个报文分别属于不同的数据流。
在该时间窗口内该任一输入端口可能输入一个或多个数据流,则该多个报文中的部分报文可能属于同一个数据流,为了避免上传重复信息,该任一网络设备对该多个报文的元数据进行去重,得到至少一个报文的元数据。例如,该多个报文分别为报文1-3,其中,报文1的元数据与报文2的元数据相同,与报文3的元数据不同,则报文1和报文2均属于数据流1,而报文3属于数据流2,则该任一网络设备通过对报文1-3的元数据进行去重,保留报文1的元数据和报文3的元数据。
步骤3043、对于该至少一个报文中的任一报文,该任一网络设备基于信息获取指令,生成该任一报文所属的数据流的传输信息。
该任一网络设备基于信息获取指令包括的至少一种属性标识,生成该任一报文所属的数据流的传输信息。在一种可能的实现方式中,若该信息获取指令包括传输地址的属性标识,则该任一网络设备将该任一报文的元数据内的源IP地址、源MAC地址中的至少一个,确定为该数据流的源地址信息,该任一网络设备将该任一报文的元数据内的目的IP地址、目的MAC地址、中的至少一个,确定为该数据流的目标地址信息,并将该源地址信息以及该目的地址信息添加至该数据流的传输信息中;若该信息获取指令包括传输端口的属性标识,则该任一网络设备将任一报文的元数据内源端口标识以及目的端口标识添加至该数据流的传输信息中;若该信息获取指令包括传输协议的属性标识,则该任一网络设备将该任一报文的元数据内的传输协议类型添加至该数据流的传输信息中;若该信息获取指令包括传输方向的属性标识,则该任一网络设备基于该任一网络设备内用于传输该任一报文的端口或该数据流的源地址信息,确定该数据流的传输方向,并将用于指示该传输方向的方向标识添加至该数据流的传输信息中;若该信息获取指令携带传输时间的属性标识,则该任一网络设备将该时间窗口的标识添加至该数据流的传输信息中。
在一种可能的实现方式中,该任一网络设备基于该任一网络设备内用于传输该任一报文的端口或该数据流的源地址信息,确定该数据流的传输方向的过程为:若该任一网络设备内 用于输入该任一报文的端口为第一输入端口,或该任一网络设备内用于输出该任一报文的端口为第一输出端口,则该数据流为上行数据流,该任一网络设备确定该数据流的传输方向为上行;若该任一网络设备内用于输入该任一报文的端口为第二输入端口,或该任一网络设备内用于输出该任一报文的端口为第二输出端口,则该数据流为下行数据流,该任一网络设备确定该数据流的传输方向为下行;若该数据流的源地址信息所指示的第一设备为终端,则该数据流为上行数据流,该任一网络设备确定该数据流的传输方向为上行;若该数据流的源地址信息所指示的第二设备为服务器,则该数据流为下行数据流,该任一网络设备确定该数据流的传输方向为下行。
例如,在该时间窗口内,该任一网络设备共传输了3个数据流,分别为数据流1-3,其中,数据流1和数据流2均为上行数据流,数据流3为下行数据流,通过本步骤3043,该任一网络设备能够获取到表1所示的数据流1-3的传输信息。
表1
Figure PCTCN2021105494-appb-000001
若该任一网络设备传输的至少一个数据流包括第二终端传输的至少一个第二数据流,则本步骤304所示的过程也即是任一网络设备获取该第二终端传输的至少一个第二数据流的传输信息的过程。
305、该任一网络设备向该控制设备发送该至少一个传输信息。
306、该控制设备接收该至少一个传输信息。
若该至少一个数据流包括第二终端传输的至少一个第二数据流,则本步骤306所示的过程也即是控制设备获取该第二终端传输的至少一个第二数据流的传输信息的过程。
307、该控制设备存储该至少一个传输信息。
该控制设备提取每个传输信息中的至少一项信息,并该至少一项信息存储在配置表中,以实现资产识别。在一种可能的实现方式中,该控制设备将每个传输信息中的每一项信息均存储至配置表,则配置表如表1所示。
在一种可能的实现方式中,该控制设备将每个传输信息中的部分信息存储至配置表中,该部分信息包括方向标识以及终端的网络地址。可选地,该控制设备将包括同一个方向标识的至少一个传输信息中终端的网络地址与该方向标识进行关联存储。若终端的网络地址为IP地址,则该控制设备将方向标识与终端的IP地址进行关联存储,若一个传输信息中的方向标识为上行标识,则说明该传输信息所对应的数据流为上行数据流,该传输信息中的源IP地址为终端的IP地址,则该控制设备将该源IP地址添加在配置表中,并与上行标识对应,若一个传输信息中的方向标识为下行标识,则说明该传输信息所对应的数据流为下行数据流,该 传输信息中的目的IP地址为终端的IP地址,则该控制设备将该目的IP地址添加在配置表中,并与下行标识对应。
若该至少一个传输信息如表1所示,则该控制设备在表1中提取上行标识对应的源IP地址以及下行标识对应的目的IP地址,并将提取到的源IP地址与上行标识关联存储在表2所示配置表中,并将提取到的目的IP地址与下行标识关联存储在表2所示的配置表中。
表2
上行标识 下行标识
源IP地址 目的IP地址
192.168.1.240 192.168.1.150
192.168.1.180  
308、对于该至少一个传输信息中的任一传输信息,该控制设备将该任一传输信息与第一目标终端的终端类型进行关联存储,其中,该第一目标终端用于传输该任一传输信息所对应的数据流。
终端类型包括摄像头、音响、打印机、IP话机、ATM或查询机。该控制设备将该任一传输信息中的至少一项信息与该目标终端的终端类型进行关联存储。可选地,该控制设备将该至少一项信息与该第一目标终端的终端类型关联存储在资产表中,该资产表用于存储多种终端类型,该资产表存储在控制设备的本地或云空间。
该控制设备从配置表中获取该任一传输信息中的至少一项信息,并基于该至少一项信息,查询资产表中是否存在该至少一项信息所对应的终端类型,若存在,说明该至少一项信息所对应的终端类型为第一目标终端的终端类型,该控制设备已经将该至少一项信息与第一目标终端的终端类型进行关联存储,则控制设备无须再次执行本步骤308,若不存在,则该控制设备将该至少一项信息与第一目标终端的终端类型关联存储在资产表中。
在一种可能的实现方式中,若资产表中不存在该至少一项信息所对应的终端类型,则该控制设备在终端类型添加界面中显示终端类型添加信息,以提示用户在资产表中添加该至少一项信息所对应的终端类型,其中,该终端类型添加信息用于指示在资产表中添加该至少一项信息对应的终端类型,该终端类型添加信息包括该至少一项信息以及添加标识,该添加标识用于指示在资产表中添加该至少一项信息对应的终端类型。
用户基于该终端类型添加信息中的至少一项信息,确定该第一目标终端的终端类型,并在该终端类型添加界面中添加第一目标终端的终端类型,当控制设备检测到用户在该终端类型界面执行了添加操作后,触发控制设备将该至少一项信息与用户所添加的终端类型关联存储在资产表中。
若该至少一项信息包括终端的网络地址,且终端的网络地址若为IP地址,则对于配置表中的任一IP地址,该网络设备将该任一IP地址所指示的终端的终端类型与该任一IP地址关联存储在资产表中。仍以表2所示的配置表为例,若IP地址192.168.1.240所指示的终端为摄像头,IP地址192.168.1.180所指示的终端为摄像头,IP地址192.168.1.150所指示的终端为音响,则该控制设备将这3个IP地址与对应的终端类型关联存储资产表中,得到表3。
表3
终端类型 终端的IP地址
摄像头 192.168.1.240
  192.168.1.180
音响 192.168.1.150
若该至少一个数据流为第二终端传输的至少一个第二数据流,本步骤308所示的过程也即是该控制设备将至少一个第二数据流的至少一个传输信息与该第二终端的终端类型进行关联存储的过程。
为了进一步说明步骤301-308所示的过程,参见图4所示的本申请实施例提供的一种存储终端类型的示意图。对于网络设备传输的数据流,该网络设备获取传输的该数据流的传输信息,并将获取到的传输信息发送至控制设备,控制设备存储接收到的传输信息,并查询资产表中是否存在该传输信息对应的终端类型,若不存在,则控制设备提示用户在资产表中添加该传输信息对应的终端类型,控制设备再基于用户的操作在资产表中添加该传输信息对应的终端类型。
本申请实施例所示的方法,通过网络设备向控制设备上传数据流的传输信息,以便控制设备将终端验证系统中终端传输的各个数据流的传输信息与终端类型进行关联存储,从而到达统计终端验证系统中终端的终端类型的目的。
需要说明的是,上述步骤301-308所示的过程由网络设备基于信息获取指令,向控制设备上传数据流的传输信息,由控制设备存储网络设备上传的传输信息。在另一种可能的实现方式中,控制设备不向网络设备发送该信息获取指令,而是向网络设备发送数据流获取指令,当网络设备接收到数据流获取指令后,该网络设备将在时间窗口内传输的至少一个数据流中的多个报文复制一份,上传至控制设备,由控制设备基于网络设备上传的至少一个数据流中的多个报文,获取该至少一个数据流的传输信息,并执行步骤307-308。其中,控制设备基于网络设备上传的至少一个数据流中的多个报文,获取该至少一个数据流的至少一个传输信息的过程与步骤3041-3043所示的过程同理,在此,本申请实施例对控制设备基于网络设备上传的至少一个数据流中的多个报文,获取该至少一个数据流的至少一个传输信息的过程不做赘述。在一种可能的实现方式中,若终端验证系统中不存在控制设备,则由网络设备周期性的执行步骤304、步骤307以及步骤308。
对于资产表中记录的任一终端类型,该控制设备从至少一个网络设备获取该任一终端类型下终端的多个传输特征,并基于获取到的多个传输特征,进行训练,得到该任一终端类型对应的目标模型。为了进一步说明该过程,参见图5所示的本申请实施例提供的一种模型获取方法的流程图。
501、控制设备确定待获取的目标模型所对应的目标终端类型。
在一种可能的实现方式中,该控制设备基于用户操作,确定该目标终端类型。用户在该控制设备的训练界面内输入该目标终端类型,并进行确认操作,以实现开启训练的过程,当控制设备在该训练界面内检测到用户进行了确认操作后,该控制设备从该训练界面获取该目标终端类型。
502、该控制设备向至少一个网络设备发送传输特征获取指令,该传输特征获取指令用于指示获取该目标终端类型下终端的传输特征。
该传输特征获取指令包括该目标终端类型以及该目标终端类型对应的至少一个网络地址,每个网络地址对应一个终端,一个终端的传输特征为该终端传输的至少一个数据流的总体传输特征。
在一种可能的实现方式中,该控制设备在资产表中查询该目标终端类型所对应的至少一个网络地址,将查询到的至少一个网络地址组成一个终端地址列表,并将该终端地址列表添加在传输特征获取指令中。其中,终端地址列表用于记录目标终端类型、该目标终端类型对应的至少一个网络地址。以目标终端类型对应的至少一个网络地址为终端的IP地址为例,参见表4所示的终端地址列表。
表4
Figure PCTCN2021105494-appb-000002
为了进一步说明步骤501-502所示的过程,参见图6所示的本申请实施例提供的一种开启训练的示意图,用户配置目标终端类型为待获取的目标模型所对应的终端类型(也即是配置目标终端类型进行训练),控制设备基于用户配置的目标终端类型查询资产表,得到该目标终端类型对应的至少一个网络地址,并将该目标终端类型以及该至少一个网络地址下发至网络设备,以开启训练的过程。
503、对于该至少一个网络设备中的任一网络设备,该任一网络设备接收传输特征获取指令。
504、该任一网络设备基于该传输特征获取指令,获取目标终端类型下至少一个第二终端的多个传输特征。
该至少一个第二终端为该传输特征指令中的至少一个网络地址所指示的终端。该至少一个第二终端的终端类型为目标终端类型。一个终端的一个传输特征为该终端传输的至少一个数据流的总体传输特征,该传输特征可能包括多个维度的特征,并不单指一个维度的特征。可选地,该传输特征为该至少一个数据流在一个时间窗口内的总体传输特征。
在一种可能的实现方式中,该传输特征包括上行传输特征,该上行传输特征为该至少一个数据流中至少一个上行数据流的总体传输特征,该上行数据流为该终端输出的数据流,上行数据流由上行报文组成。可选地,该上行传输特征包括上行报文特征、上行流特征中的一个,该上行报文特征为该至少一个上行数据流中上行报文的总体特征,该上行流特征为该至少一个上行数据流的统计特征。
可选地,该上行报文特征包括上行报文平均传输间隔、上行负载平均值、上行总负载大小、上行报文个数、上行目标报文个数、上行目标报文占比、上行报文负载波动值中的至少一个,该上行报文平均传输间隔为在一个时间窗口内该至少一个上行数据流中上行报文的平均传输间隔,该上行负载平均值为在该时间窗口内该至少一个上行数据流中目标报文的负载的平均大小,该上行总负载大小为在该时间窗口内该至少一个上行数据流中目标报文的负载的总大小,该上行报文个数为在该时间窗口内该至少一个上行数据流中上行报文的个数,该上行目标报文个数在该时间窗口内该至少一个上行数据流中目标报文的个数,该上行目标报文占比为在该时间窗口内该至少一个上行数据流中目标报文的占比,该上行报文负载波动值用于指示在该时间窗口内该至少一个上行数据流中目标报文的大小波动情况。该目标报文为具有负载的报文,可选地,该负载中的数据为业务数据。可选地,该上行报文特征还包括第一接收窗口波动值、第一接收窗口大小平均值中的至少一个,该第一接收窗口波动值用于指示在该时间窗口内该至少一个上行数据流中上行报文携带的接收窗口大小的波动情况,该第 一接收窗口大小平均值为在该时间窗口内该至少一个上行数据流中上行报文携带的接收窗口的平均大小。可选地,该接收窗口为TCP滑动窗口。可选地,所该第一接收窗口波动值为在该时间窗口内该至少一个上行数据流中上行报文携带的接收窗口大小的标准差。
可选地,该上行流特征包括上行终端端口波动值、上行数据流总个数、上行目标数据流个数、至少一种数据流类型中每种数据流类型下的上行数据流个数、至少一种传输协议类型中每种传输协议类型下的上行数据流个数中的至少一个。该上行终端端口波动值用于指示在该时间窗口内终端中至少一个上行数据流的输出端口的波动情况,也即是在该时间窗内在该终端中有上行数据流传输的输出端口的波动情况。该上行目标数据流为所对应的服务器输入端口属于目标端口范围的上行数据流。可选地,该上行终端端口波动值为该至少一个上行数据流的输出端口的端口标识的方差值。该上行数据流总个数为该至少一个上行数据流的总个数,该上行目标数据流个数为该至少一个上行数据流中上行目标数据流的总个数。该至少一种传输数据流类型包括数据流、广播数据流和单播数据流。该至少一种传输协议类型包括TCP、UDP和ICMP。该目标端口范围能够根据实际需求进行设置,例如该目标端口范围为知名端口的范围,本申请实施例对该目标端口范围不作具体限定。
在一种可能的实现方式中,该传输特征还包括该至少一个数据流的总个数、下行传输特征中的至少一个,该下行传输特征为该至少一个数据流中至少一个下行数据流的总体传输特征,该下行数据流为该终端输入的数据流,由下行报文组成。可选地,该下行传输特征包括下行报文特征、下行流特征中的至少一个,该下行报文特征为该至少一个下行数据流中下行报文的总体特征,该下行流特征为该至少一个下行数据流的统计特征。
可选地,该下行报文特征包括下行报文平均传输间隔、下行负载平均值、下行总负载大小、下行报文个数、下行目标报文个数、下行目标报文占比、下行报文负载波动值中的至少一个,该下行报文平均传输间隔为在一个时间窗口内该至少一个下行数据流中下行报文的平均传输间隔,下行负载平均值为在该时间窗口内该至少一个下行数据流中目标报文的负载的平均大小,该下行总负载大小为在该时间窗口内该至少一个下行数据流中目标报文的负载的总大小,该下行报文个数为在该时间窗口内该至少一个下行数据流中下行报文的个数,该下行目标报文个数在该时间窗口内该至少一个下行数据流中目标报文的个数,该下行目标报文占比为在该时间窗口内该至少一个下行数据流中目标报文的占比,该下行报文负载波动值用于指示在该时间窗口内所述该一个下行数据流中目标报文的大小波动情况。可选地,该下行报文特征还包括第二接收窗口波动值、第二接收窗口大小平均值中的至少一个,该第二接收窗口波动值用于指示该至少一个下行数据流中下行报文携带的接收窗口大小的波动情况,第二接收窗口大小平均值为在该时间窗口内该至少一个下行数据流中下行报文携带的接收窗口的平均大小。可选地,该第二接收窗口波动值为在该时间窗口内该至少一个下行数据流中下行报文携带的接收窗口大小的标准差。
可选地,该下行流特征包括下行终端端口波动值、下行数据流总个数、下行目标数据流个数、至少一种数据流类型中每种数据流类型下的下行数据流个数、至少一种传输协议类型中每种传输协议类型下的下行数据流个数中的至少一个。该下行终端端口波动值用于指示在该时间窗口内终端中该至少一个下行数据流的输入端口的波动情况,也即是在该时间窗内在该终端中有下行数据流传输的输入端口的波动情况。该下行目标数据流为对应的服务器输出端口属于目标端口范围的下行数据流。可选地,该下行终端端口波动值为在该时间窗口内该 终端中至少一个下行数据流的输入端口的端口标识的方差值。下行数据流总个数为该至少一个下行数据流的总个数,该下行目标数据流个数为该至少一个下行数据流中下行目标数据流的总个数。
当获取到该传输特征获取指令后,该任一网络设备从该传输特征指令中获取该目标终端类型所对应的至少一个网络地址(例如至少一个第二终端的IP地址),在多个时间窗口中的任一时间窗口内,该任一网络设备基于该至少一个网络地址中的任一网络地址、该任一网络地址所指示的第二终端传输的数据流,确定该第二终端的一个传输特征,从而在该任一时间窗口内,该任一网络设备能够获取到该至少一个第二终端的一个传输特征,在该多个时间窗口内,该任一网络设备能够获取到该至少一个第二终端的多个传输特征。
其中,在该任一时间窗口内,该任一网络设备基于该至少一个网络地址中的任一网络地址以及该任一网络地址所指示的第二终端传输的数据流,确定该第二终端的一个传输特征的方式包括方式1或方式2。
方式1、在该任一时间窗口内,该任一网络设备基于该第二终端传输的每个数据流,确定该第二终端的一个传输特征。
该任一网络设备先获取每个数据流的流传输特征,再基于获取到的至少一个流传输特征,获取该第二终端的一个传输特征。在一种可能的实现方式中,方式1能够由下述步骤A-C来实现。
步骤A、在该任一时间窗口内,该任一网络设备基于该任一网络地址,确定该网络地址指示的该第二终端所传输的至少一个第三数据流。
其中,该至少一个第三数据流中的报文均为广域网报文。在该任一时间窗口内,若该任一网络设备所输入的任一报文携带有该任一网络地址,则该网络设备确定该任一报文所属的数据流为该第二终端传输的数据流;该任一网络设备通过比较该任一报文的目的IP地址以及源IP地址,确定该任一报文是否为广域网报文,若该任一报文为广域网报文,则该任一网络设备确定该任一报文所属的数据流为该第二终端传输的一个第三数据流。
步骤B、该任一网络设备获取该至少一个第三数据流中每个第一第三数据流的流传输特征,一个第三数据流的流传输特征为该第三数据流的传输特征。
一个数据流的流传输特征包括该数据流的传输信息、数据流类型、目标端口类型、报文特征中的至少一个。其中,该传输信息用于指示数据流的传输属性,该传输信息包括该数据流的源地址信息(源IP地址和/或源MAC地址)、源端口标识、目的地址信息(目的IP地址和/或目的MAC地址)、目的端口标识以及传输协议类型中的至少一个。可选地,该传输信息还包括方向标识、该时间窗口的标识中的至少一个,该方向标识用于指示该数据流的传输方向。
该目标端口类型为服务器中传输该数据流的端口的端口类型,其中,端口的端口类型分为第一端口类型、第二端口类型以及第三端口类型,可选地,第一端口类型对应的端口区间为知名端口区间[0,1024),第二端口类型对应的端口区间为注册端口区间[1024,49152),第三端口类型对应的端口区间为私有端口区间[49152,+∞)。需要说明的是,第一端口类型、第二端口类型以及第三端口类型所对应的端口区间均可根据具体实施场景进行设置,例如,若用户根据业务需求,将业务所对应的端口配置在区间[10001,20000),则第一端口类型对应的端口区间为[0,10001)第二端口类型对应的端口区间为[10001,20000)第三端口类型 对应的端口区间为[20000,+∞)。在此,本申请实施例对第一端口类型、第二端口类型以及第三端口类型所对应的端口区间不作具体限定。
该报文特征为该数据流中报文的特征。可选地,该报文特征包括报文传输间隔总和、负载大小、负载平方和、目标报文个数、报文总个数、接收窗口波动值、窗口总大小、窗口平方和中的至少一个,该报文传输间隔总和为在一个时间窗口内该数据流中报文之间的传输间隔的总时长,该负载大小为在该时间窗内该数据流中目标报文的负载的总大小,该负载平方和为该目标报文的负载的大小平方和,该目标报文个数为在该时间窗口内该报文中目标报文的总个数,该报文总个数为在该时间窗口内该数据流中报文的总个数,该接收窗口波动值用于指示在该时间窗口内该数据流中报文携带的接收窗口大小的波动情况,该窗口总大小为在该时间窗口内该数据流中报文携带的接收窗口大小的总和,该窗口平方和为该滑动窗口的大小的平方和。
对于该至少一个第三数据流中任一第三数据流,任一网络设备获取该任一第三数据流的传输信息、数据流类型、目标端口类型、报文特征中的至少一个,并将获取到的这些信息确定为该任一第三数据流的流传输特征。
其中,该任一网络设备获取该任一第三数据流的传输信息的过程与步骤304中任一网络设备获取至少一个数据流的传输信息的过程同理,在此本申请实施例对该任一网络设备获取该任一第三数据流的传输信息的过程不做赘述。
该任一网络设备获取该任一第三数据流的数据流类型的过程为:该任一网络设备根据任一第三数据流中任一报文所携带的源IP地址以及目的IP地址,确定该任一第三数据流为广播数据流还是组播数据流,若该任一第三数据流既不是广播数据流也不是组播数据流,则该任一网络设备确定该任一第三数据流为单播数据流。
该任一网络设备获取该任一第三数据流的目标端口类型的过程:该任一网络设备获取该任一报文中服务器的端口标识,并将该服务器的端口标识所属的端口类型确定为该目标端口类型。其中,该服务器的端口标识为该任一报文的源端口标识或目的端口标识。
该任一网络设备获取该任一第三数据流的报文特征的过程为:在该任一时间窗口内,该任一网络设备记录该任一第三数据流传输第一个报文的第一时间,以及从传输最后一个报文的第二时间,并将该第二时间与第一时间之间的时间差,确定为报文传输间隔总和;在该任一时间窗口内,该任一网络设备统计该任一第三数据流中目标报文的负载的总大小,并将该总大小确定为负载大小;在该任一时间窗口内,该任一第三数据流每传输一个目标报文,该任一网络设备计算该目标报文的负载的大小平方,从而在该任一时间窗口内该任一网络设备能够得到至少一个该目标报文的负载的大小平方,该任一网络设备对该至少一个该目标报文中负载的大小平方进行求和,得到负载平方和;该任一网络设备统计在该任一时间窗口内该任一第三数据流中目标报文的总个数,得到目标报文个数;该任一网络设备统计在该任一时间窗口内该任一第三数据流中报文的总个数,得到报文总个数;在该任一时间窗口内,该任一网络设备对该任一第三数据流中报文携带的接收窗口大小进行求和,得到窗口总大小;在该任一时间窗口内,该任一网络设备对该任一第三数据流中各个报文携带的接收窗口大小进行平方,并对各个报文携带的接收窗口大小的平方进行求和,得到窗口平方和;该任一网络设备基于该窗口总大小以及窗口平方和,获取报文携带的接收窗口大小的标准差,并将该标准差确定为该接收窗口波动值;该任一网络设备将该报文传输间隔总和、负载大小、负载平方和、 目标报文个数、报文总个数、接收窗口波动值、窗口总大小、窗口平方和中的至少一个确定为该报文特征。
以该至少一个第三数据流包括数据流a-e为例,传输数据流a-e的终端的IP地址分别为192.168.1.2、192.168.1.2、192.168.1.5、192.168.1.5、192.168.1.6,数据流a-e的流传输特征如下述表5所示。
表5
Figure PCTCN2021105494-appb-000003
步骤C、该任一网络设备基于该至少一个第三数据流的流传输特征,获取该第二终端的一个传输特征。
本步骤C由下述步骤C1-C3所示的过程来实现。
步骤C1、该任一网络设备基于该至少一个第三数据流中至少一个上行数据流的流传输特征,获取上行传输特征。
该任一网络设备基于该至少一个上行数据流的流传输特征(简称“至少一个流传输特征”),获取该至少一个上行数据流的上行报文特征、上行流特征中的至少一个,将获取到的上行报文特征、上行流特征中的至少一个确定为该上行传输特征。
其中,该任一网络设备基于该至少一个流传输特征,获取该上行报文特征的过程为:该任一网络设备将该至少一个流传输特征中的报文传输间隔总和进行求和,得到第一和值,该第一和值为该至少一个上行数据流中上行报文的报文传输间隔总和;该任一网络设备将该至少一个流传输特征中的报文总个数进行求和,得到上行报文个数;该任一网络设备将该第一和值除以该上行报文个数,得到上行报文平均传输间隔;该任一网络设备对该至少一个流传输特征中的负载大小进行求和,得到上行总负载大小;该任一网络设备对该至少一个流传输特征中的目标报文个数进行求和,得到上行目标报文个数;该任一网络设备将该上行目标报文个数除以该上行报文个数,得到上行目标报文占比;该任一网络设备将该上行总负载大小除以上行目标报文个数,得到上行负载平均值;该任一网络设备基于该至少一个流传输特征中的负载大小,计算在该时间窗口内该至少一个上行数据流中目标报文的大小的标准差,并 将该标准差作为上行报文负载波动值;该任一网络设备将该至少一个流传输特征中的接收窗口波动值求平均值,得到第一接收窗口波动值;该任一网络设备对该至少一个流传输特征中的窗口总大小进行求和,得到第二和值,该第二和值为在该时间窗口内该至少一个上行数据流中上行报文携带的接收窗口大小的总和;该任一网络设备将该第二和值除以该至少一个上行数据流中上行报文携带的接收窗口的总个数,得到第一接收窗口大小平均值;该任一网络设备将上行报文平均传输间隔、上行负载平均值、上行总负载大小、上行报文个数、上行目标报文个数、上行目标报文占比、上行报文负载波动值、第一接收窗口波动值、第一接收窗口大小平均值中的至少一个确定为该上行报文特征。
该任一网络设备基于该至少一个流传输特征,获取上行流特征的过程为:该任一网络设备从该至少一个流传输特征的传输信息中获取至少一种源端口标识,该任一网络设备计算该至少一种源端口标识的方差,并将该方差确定为上行终端端口波动值;该任一网络设将该至少一个流传输特征的总个数,确定为上行数据流总个数;该任一网络设备统计该至少一个上行数据流中上行目标数据流的总个数,得到该上行目标数据流个数;该任一网络设备获取该至少一个流传输特征中的数据流类型,并基于获取到数据流类型,统计至少一种数据流类型中每种数据流类型下的上行数据流个数;该任一网络设备获取该至少一个流传输特征中的传输协议类型,并基于获取到传输协议类型,统计至少一种传输协议类型中每种传输协议类型下的上行数据流个数;该任一网络设备确定出上行终端端口波动值、上行数据流总个数、上行目标数据流个数、至少一种数据流类型中每种数据流类型下的上行数据流个数、至少一种传输协议类型中每种传输协议类型下的上行数据流个数中的至少一个,确定为该上行流特征。
例如至少一个上行数据流包括数据流1和数据流2,数据流1为组播数据流,数据流2为广播数据流,而至少一个数据流类型包括组播数据流、广播数据流以及单播数据流,则该至少一个数据流类型下的上行数据流个数分别为1、1、0。例如至少一个上行数据流包括数据流1和数据流2,数据流1的传输协议类型为TCP,数据流2的传输协议类型为UDP,而至少一个传输协议类型包括TCP、UDP以及ICMP,则该至少一个传输协议类型下的上行数据流个数分别为1、1、0。
步骤C2、该任一网络设备基于该至少一个第三数据流中至少一个下行数据流的流传输特征,获取下行传输特征。
本步骤C2与步骤C1所示的过程的同理,在此本申请实施例对本步骤C2不做赘述。
步骤C3、该任一网络设备将该上行传输特征和下行传输特征确定为该第二终端的一个传输特征。
在一种可能的实现方式,该任一网络设备将该上行传输特征确定为该第二终端的一个传输特征。在另一种可能的实现方式中,该任一网络设备将该上行传输特征、下行传输特征以及该至少一个第三数据流的总个数,确定为该第二终端的一个传输特征。
以表5为例,由于表5中数据流a和b的源IP地址相同,则数据流a和b为同一个第二终端输出的数据流,该任一设备基于表5中数据流a和b的流传输特征,获取该第二终端的一个传输特征。数据流c和d的源IP地址相同,则数据流c和d为同一个第二终端输出的数据流,该任一设备基于表5中数据流c和d的流传输特征,获取该第二终端的一个传输特征。该任一设备基于表5中数据流e的流传输特征,数据流e的目的IP地址所指示的第二终端的一个传输特征,最终得到表6所示的3个第二终端的传输特征。
表6
Figure PCTCN2021105494-appb-000004
为了进一步说明上述方式1所示的过程,参见图7所示的本申请实施例提供的一种传输特征获取的示意图。图7中的一个时间窗口的目标时长均为1小时,从2016-09-22至2016-10-12中的每一天划分为24个时间窗口,任一网络设备在每个时间窗口,获取第二终端的一个传输特征,在每个时间窗口内,当该任一网络设备输入该第二终端传输的任一第三数据流时,该任一网络设备获取该任一第三数据流的流传输特征,然后该任一网络设备基于在每个时间窗口内传输的至少一个第三数据流的流传输特征,获取该第二终端的一个传输特征,还能获取第二终端的终端类型,例如任一网络设备在2016-09-24的5:00-5:59所在的时间窗口,共输入第二终端传输的344个第三数据流,则该任一网络设备基于该344个第三数据流的流传输特征,采集在该时间窗口内该第二终端的一个传输特征,并配置采集到的传输特征,例如上/下行报文特征,上/下行流特征等。需要说明的是,上述举例是以1小时为一个时间窗口,统计的传输特征,而在另外一些实施例中,是以1分钟为一个时间窗口,实时统计在每分钟内的传输特征,时间窗口的窗口时长(也即是目标时长)可根据具体实施场景进行设置,在此,本申请实施例对时间窗口的窗口时长不作具体限定。
方式2、在该任一时间窗口内,该任一网络设备根据该至少一个网络地址,确定第二终端的一个传输特征。
在任一时间窗口内,该任一网络设备获取携带该至少一个网络地址的多个上行报文的上行报文特征、该多个上行报文所在的上行数据流的上行流特征中的至少一个,该任一网络设备并将获取到的上行报文特征、上行流特征中的至少一个确定为上行传输特征。
在任一时间窗口,该任一网络设备获取携带该至少一个网络地址的多个上行报文的上行报文特征的过程为:该任一网络设备统计在该任一时间窗口内携带该至少一个网络地址的多个上行报文的总个数,得到上行报文个数;该任一网络设备记录在该任一时间窗口内该多个上行报文中首报文的到达时间以及尾报文的达到时间,并将这2个达到时间之间的时间差除以目标差值,得到上行报文平均传输间隔,其中,首报文为在该任一时间窗口内,携带该至少一个网络地址中任一网络地址且第一个达到该任一网络设备的上行报文,尾报文为在该任一时间窗口内,携带该至少一个网络地址中的任一网络地址且最后一个达到该任一网络设备的上行报文,该目标差值为上行报文个数与1的差值;该任一网络设备统计在该任一时间窗口内携带该至少一个网络地址的目标报文的个数,得到上行目标报文个数;该任一网络设备将该上行目标报文个数除以该上行报文个数,得到上行目标报文占比;该任一网络设备统计在该任一时间窗口内携带该至少一个网络地址的目标报文的负载总大小,得到上行总负载大小,该任一网络设备将该上行总负载大小除以上行目标报文个数,得到上行负载平均值;该任一网络设备统计在该任一时间窗口内携带该至少一个网络地址的多个目标报文的负载大小,并计算该多个目标报文的负载大小的标准差,并将该标准差确定为上行报文负载波动值;该任一网络设备将携带该至少一个网络地址的上行报文所携带的接收窗口大小求标准差,得到 第一接收窗口波动值;该任一网络设备对在该时间窗口内该多个上行报文携带的接收窗口大小进行求和,得到第一接收窗口总大小;该任一网络设备将该第一接收窗口总大小除以该多个上行报文携带的接收窗口大小的总个数,得到第一接收窗口大小平均值;该任一网络设备将上行报文平均传输间隔、上行负载平均值、上行总负载大小、上行报文个数、上行目标报文个数、上行目标报文占比、上行报文负载波动值、第一接收窗口波动值、第一接收窗口大小平均值中的至少一个确定为该上行报文特征。
在任一时间窗口内,该任一网络设备获取携带该至少一个网络地址的多个上行报文所在的上行数据流的上行流特征的过程为:在该任一时间窗口内,该任一网络设备获取携带该至少一个网络地址的多个上行报文中的多种源端口标识,该任一网络设备计算该多种源端口标识的方差,并将该方差确定为上行终端端口波动值;在该任一时间窗口内,该任一网络设备统计携带该至少一个网络地址的上行数据流的总个数,得到上行数据流总个数;在该任一时间窗口内,该任一网络设备统计携带该至少一个网络地址的上行数据流中上行目标数据流的总个数,得到上行目标数据流个数;在该任一时间窗口内,该任一网络设备统计至少一种数据流类型中每种数据流类型下携带该至少一个网络地址的上行数据流的个数,得到至少一种数据流类型中每种数据流类型下的上行数据流个数;在该任一时间窗口内,该任一网络设备统计至少一种传输协议类型中每种传输协议类型下携带该至少一个网络地址的上行数据流的个数,得到至少一种传输协议类型中每种传输协议类型下的上行数据流个数;该任一网络设备确定出上行终端端口波动值、上行数据流总个数、上行目标数据流个数、至少一种数据流类型中每种数据流类型下的上行数据流个数、至少一种传输协议类型中每种传输协议类型下的上行数据流个数中的至少一个,确定为该上行流特征。
该任一网络设备还能够参考方式2中获取上行传输特征的过程,来获取下行传输特征。并参照步骤C3所示的过程,来获取该第二终端的一个传输特征。
在该多个时间窗口中的每个时间窗口内,该任一网络设备基于上述方式1或方式2,确定每个网络地址所指示的第二终端的一个传输特征,从而在多个时间窗口内,该任一网络设备能够获取到该至少一个第二终端的多个传输特征。
505、该任一网络设备向该控制设备发送该至少一个第二终端的多个传输特征。
506、该控制设备接收该至少一个第二终端的多个传输特征。
由于该至少一个网络设备中的任一网络设备均向该控制设备发送该至少一个第二终端的多个传输特征,因此,该控制设备能够从该至少一个网络设备分别接收该至少一个第二终端的多个传输特征。
507、该控制设备将该至少一个第二终端的多个第三传输特征作为初始模型的输入和输出,进行训练,得到目标模型。
该多个第三传输特征为该至少一个第二终端的多个传输特征中的部分传输特征。每个第二终端的一个传输特征为一个样本数据,每个第二终端均正常终端,当该控制设备接收到的该至少一个第二终端的多个传输特征的个数大于或等于目标个数时,说明正常终端的样本数据足够;该控制设备将接收到的该至少一个第二终端的多个传输特征划分为三部分,并将这三部分分别放入训练集、验证集以及测试集,也即是训练集、验证集以及测试集均包括该至少一个第二终端的多个传输特征中的部分传输特征,其中,训练集中的各个传输特征为第三传输特征,验证集中各个传输特征为第四传输特征,测试集中的各个传输特征为第六传输特 征。
可选地,该目标模型用于重构出正常传输特征。可选地,该目标模型为无监督深度神经网络模型,例如深度自编码器模型。该目标模型包括编码器(encoder)和解码器(decoder),传输特征X=[x 1,x 2,...,x n]作为编码器的输入信号输入编码器,编码器学习输入的传输特征X到隐编码,其学习的过程为编码器对传输特征X进行降维编码,得到编码特征Z,编码特征Z作为解码器的输入信号,输入至解码器中;解码器尽量将输入的编码特征Z重构成传输特征X,可选地,解码器对编码特征Z进行升维解码,输出传输特征X’=[x’ 1,x’ 2...,x’ n],传输特征X’为解码器所还原出的传输特征X,其中,x n为传输特征X所包括的第n个特征,例如上行数据流总个数,n为传输特征X的总维度或者是传输特征中特征的总个数,n>1,x’ n为重构出的x n
在对初始模型进行训练的过程中,预设q次训练次数、目标损失函数以及预设阈值,该目标损失函数用于计算模型输入的传输特征与输出的传输特征之间的平均差异情况,在第i次训练过程中,控制设备将多个第三传输特征输入第i模型,该第i模型重构输入的每个第三传输特征,输出每个第三传输特征的重构传输特征,该控制设备将多个第三传输特征与对应的重构传输特征输入该目标损失函数,计算目标损失函数值;若该目标损失函数值大于或等于该预设阈值,且i<q,则该控制设备基于优化算法,继续迭代,更新第i模型的模型参数,得到第i+1模型,控制设备进入第i+1次训练过程;若该目标损失函数值小于该预设阈值,则控制设备结束训练,将第i模型确定为该目标模型;若该目标损失函数值大于或等于预设差值且i≥q,则该控制设备结束训练,以避免控制设备在训练不出目标模型的情况下持续进行迭代训练,或者若模型的模型参数在相邻两次迭代之间的变化小于预设的模型参数变化值,该控制设备结束训练。
其中,q≥i≥1,当i=1时,第i模型为初始模型,当i>1时,第i模型为在第i-1次训练过程中,更新参数后的第i-1模型。该优化算法包括梯度下降算法,该目标损失函数如下述公式(1)所示,其中,m为多个第三传输特征的总个数,
Figure PCTCN2021105494-appb-000005
为第j个第三传输特征中第i个维度的数值,
Figure PCTCN2021105494-appb-000006
是目标模型重构出的
Figure PCTCN2021105494-appb-000007
n≥i>0,m≥j>0。
Figure PCTCN2021105494-appb-000008
可选地,在进行训练的过程中,每进行一次训练,该控制设备还能够显示训练进度信息,该训练进度信息包括当前的训练次数、目标损失函数值,以便用户能够通过该训练进度信息获取训练进度。
508、该控制设备将目标终端类型下多个目标终端的多个第四传输特征输入该目标模型,由该目标模型基于输入的该多个第四传输特征,重构每个第四传输特征,输出多个第五传输特征,该多个第四传输特征与该多个第五传输特征一一对应,一个第四传输特征为一个目标终端传输的至少一个数据流的总体传输特征。
该多个目标终端包括至少一个第二终端以及至少一个第三终端,其中,该至少一个第二终端均为正常终端,该至少一个第三终端均为异常终端。一个第五传输特征为重构出的一个第四传输特征。
验证集除了包括至少一个第二终端的多个传输特征以外,还包括至少一个第三终端的多个传输特征,在该验证集中第二终端的传输特征和第三终端的传输特征均视为第四传输特征。 该控制设备从验证集中获取多个第四传输特征,并将获取到的多个第四传输特征输入该目标模型,由目标模型基于输入的多个第四传输特征,输出重构出的多个第四传输特征,其中,该多个第四传输特征属于至少一个第二终端和至少一个第三终端,重构出的多个第四传输特征也即是该多个第五传输特征,每个第五传输特征对应一个第四传输特征。
509、该控制设备基于该多个第五传输特征与对应的第四传输特征,获取目标差异度。
该控制设备基于该多个第五传输特征中至少一个第五传输特征与对应的第四传输特征之间的差异度,获取该目标差异度,一个第五传输特征与对应的第四传输特征之间的差异度用于指示该一个第五传输特征与对应的第四传输特征之间的差异情况,其中,该至少一个第五传输特征所对应的第四传输特征均为第二终端的传输特征。可选地,本步骤509由下述本步骤5091-5093所示的过程来实现。
步骤5091、对于至少一个第二终端中任一第二终端的任一第四传输特征,该控制设备获取该任一第四传输特征与对应的第五传输特征之间的差异度。
可选地,该任一第四传输特征与对应的第五传输特征之间的差异度为该任一第四传输特征与对应的第五传输特征中相应特征之间的均方误差(mean square error,MSE),如公式(2)所示,其中,MSE(X,X’)为该任一第四传输特征与对应的第五传输特征中相应特征之间的均方误差。
Figure PCTCN2021105494-appb-000009
该控制设备对至少一个第二终端中每个第二终端的每个第四传输特征均执行本步骤5091,从而能够得到至少一个第二终端的多个第四传输特征与对应的第五传输特征之间的差异度。在一种可能的实现方式中,该控制设备将该至少一个第二终端的每个第四传输特征所对应的第五传输特征,确定为该至少一个第五传输特征中的一个。而在另一种可能的实现方式,该控制设备对该至少一个第二终端的多个第四传输特征所对应的多个第五传输特征进行进一步筛选,以得到该至少一个第五传输特征,具体参见下述步骤5092。
步骤5092、该控制设备基于该至少一个第二终端的多个第四传输特征与对应的第五传输特征之间的差异度,从该第二终端的多个第四传输特征所对应的第五传输特征中,确定至少一个第五传输特征。
对于该第二终端的多个第四传输特征中的任一第四传输特征,若该任一第四传输特征与对应的第五传输特征之间的差异度属于目标差异区间,则该差异度为正常值,该差异度能够参与计算目标差异度,该控制设备将该任一第四传输特征所对应的第五传输特征确定为该至少一个第五传输特征中的一个。
其中,该目标差异区间包括多个差异度,该多个差异度均为正常的值,均能够参与计算目标差异度。该目标差异区间中的差异度能够根据具体情况来设置,例如控制设备采取箱线图或者设置分位数,来确定该目标差异区间,再例如确定的目标差异区间为[0,0.1],在此,本申请实施例对该目标差异区间的范围不作具体限定。
步骤5093、该控制设备基于该至少一个第五传输特征与对应的第四传输特征之间的差值度,确定该目标差异度。
该控制设备基于至少一个差异度的平均值或分数位,获取该目标差异度,其中,该至少一个差异度包括该至少一个第五传输特征与对应的第四传输特征之间的差值度,分数位包括该至少一个差异度的中值。
该控制设备基于该至少一个差异度的平均值,获取该目标差异度θ,如公式(3)所示。其中,MSE[·]为该至少一个差异度所组成的数组,AVE(MSE[·])为该至少一个差异度的平均值,STD(MSE[·])为该至少一个差值度的标准差,用于指示该至少一个差异度的波动情况,K为灵敏度系数。
θ=AVE(MSE[·])+K×STD(MSE[·])        (3)
该目标模型除了用于完全重构出正常传输特征以外,还用于不能完全重构出异常传输特征,需要说明的是,完全重构出正常传输特征是允许有一定的误差范围的。例如,当该目标模型在重构一个实际传输特征时,若该目标模型所输出的重构传输特征与该实际传输特征之间的差异度小于该目标差异度,则说明该目标模型完全重构出了该实际传输特征,若该实际传输特征为正常传输特征,则说明本次重构准确,若该实际传输特征为异常传输特征,则说明本次重构失败;当目标模型在重构一个实际传输特征时,若该目标模型所输出的重构传输特征与该实际传输特征之间的差异度大于或等于该目标差异度,则说明该目标模型未完全重构出该实际传输特征,若该实际传输特征为正常传输特征,则说明本次重构失败,若该实际传输特征为异常传输特征,则说明本次重构准确。其中,该至少一个第四传输特征均为实际传输特征,该至少一个第五传输特征均为重构传输特征,正常传输特征为正常终端的传输特征,也即是至少一个第二终端的传输特征;异常传输特征为异常终端的传输特征,也即是至少一个第三终端的传输特征。
该控制设备还能够根据该目标模型重构多个目标终端的多个第四传输特征的准确情况,动态调整K的取值,并通过调整后的K,来确定该目标差异度。在一种可能的实现方式中,该控制设备通过执行r次K值更新过程,以确定该目标差异度。其中,在第j次K值更新的过程中,该控制设备将K j作为K,并基于公式(3)计算出一个差异度θ j,若在差异度θ j下,该目标模型符合第一条件,则该控制设备将该差异度θ j确定为目标差异度,结束K值更新;否则,该控制设备将K j更新为K j+1,并以K j+1作为K,执行在第j+1次K值更新的过程。其中,r≥j≥1,K j为第j次K值更新的过程中所使用的K,θ j为基于K j以及公式(3)所计算出的差异度。
该第一条件包括下述至少一项:在该差异度θ j下,该目标模型的第一查准率大于或等于第一查准率阈值,且该目标模型的第一查全率大于或等于第一查全率阈值;在该差异度θ j下,该目标模型的第二查准率大于或等于第二查准率阈值,且该目标模型的第二查全率大于或等于第二查全率阈值。其中,该第一查准率为该目标模型完全重构出正常传输特征的准确率;该第一查全率为该目标模型完全重构出正常传输特征的概率;该第二查准率为该目标模型未完全重构出异常传输特征的准确率;该第二查全率为该目标模型未完全重构出异常传输特征的概率。本申请实施例对第一查准率阈值、第一查全率阈值、第二查准率阈值和第二查全率阈值的不作具体限定。
在一种可能的实现方式中,该控制设备通过该多个第四传输特征的第一标签以及对应的第五传输特征的第二标签,获取正常传输特征个数、第一正常个数、第二正常个数、异常传输特征个数、第一异常个数以及第二异常个数;该控制设备基于该正常传输特征个数、第一正常个数以及第二正常个数,确定该第一查准率以及第一查全率;该控制设备基于该异常传输特征个数、第一异常个数以及第二异常个数,确定该第二查准率以及第二查全率。
一个实际传输特征具有一个第一标签,该第一标签用于指示该实际传输特征是否为正常 传输特征,若第一标签是第一正常标识,则说明该实际传输特征为正常传输特征,若第一标签为第一异常标识,则说明该实际传输特征是异常传输特征。该多个第四传输特征均为实际传输特征,该多个第四传输特征中正常传输特征的第一标签均为第一正常标识,该多个第四传输特征中异常传输特征的第一标签均为第一异常标识。本申请实施例对第一正常标识和第一异常标识的表示方式不作具体限定。
一个重构传输特征具有一个第二标签,该第二标签用于指示该重构传输特征所对应的实际传输特征是否为正常传输特征,若第二标签是第二正常标识,则说明该重构传输特征所对应的实际传输特征是正常传输特征,也即是该实际传输特征所属终端为正常终端;若第二标签为第二异常标识,则说明该重构传输特征所对应的实际传输特征是异常传输特征,该实际传输特征所属终端为异常终端。该多个第五传输特征均为重构传输特征,对于一个第五传输特征,若该第五传输特征与对应的第四传输特征之间的差异度小于该差异度θ j,则说明在该差异度θ j下,该第五传输特征所对应的第四传输特征为正常传输特征,则该第五传输特征的第二标签为第二正常标识,否则,说明在该差异度θ j下,该第五传输特征所对应的第四传输特征为异常传输特征,则该第五传输特征的第二标签为第二异常标识。本申请实施例对第二正常标识和第二异常标识的表示方式不作具体限定。
该正常传输特征个数为该多个第四传输特征中第二终端的传输特征的总个数,也即是该多个第四传输特征中正常传输特征的总个数。第一正常个数为在该差异度θ j下该多个第四传输特征中目标正常传输特征的总个数,该目标正常传输特征为该多个第四传输特征中目标模型能够准确重构的正常传输特征。第二正常个数为在差异度θ j下第二标签为第二正常标识的第五数据流的个数。
该异常传输特征个数为该多个第四传输特征中第三终端的传输特征的总个数,也即是该多个第四传输特征中异常传输特征的总个数。第一异常个数为在该差异度θ j下该多个第四传输特征中目标异常传输特征的总个数,该目标异常传输特征为该多个第四传输特征中目标模型重构失败的异常传输特征;第二异常个数为在差异度θ j下第二标签为第二异常标识的第五数据流的个数。
在一种可能的实现方式中,该控制设备通过该多个第四传输特征的第一标签以及对应的第五传输特征的第二标签,获取正常传输特征个数、第一正常个数、第二正常个数、异常传输特征个数、第一异常个数以及第二异常个数包括:若一个第四传输特征的第一标签为第一正常标识,该第四传输特征为正常传输特征,则该控制设备将该多个第四传输特征的第一标签为第一正常标识的标签个数,确定为正常传输特征个数;若一个第四传输特征的第一标签为第一正常标识,且该第四传输特征所对应的第五传输特征的第二标签为第二正常标识,说明该目标模型重构该第四传输特征(正常传输特征)是准确的,则该控制设备将该第四传输特征确定为目标正常传输特征,该控制设备将该多个第四传输特征中目标正常传输特征的总个数确定为第一正常个数;该控制设备统计第二标签为第二正常标识的第五数据流的个数,并将统计出的个数确定为第二正常个数;若一个第四传输特征的第一标签为第一异常标识,该第四传输特征为异常传输特征,则该控制设备将该多个第四传输特征的第一标签为第一异常标识的标签个数,确定为异常传输特征个数;若一个第四传输特征的第一标签为第一异常标识,且该第四传输特征所对应的第五传输特征的第二标签为第二异常标识,说明该目标模型重构该第四传输特征(异常传输特征)失败,则该控制设备将该第四传输特征确定为目标 异常传输特征,该控制设备将该多个第四传输特征中目标异常传输特征的总个数确定为第一异常个数;该控制设备统计第二标签为第二异常标识的第五数据流的个数,并将统计出的个数确定为第二异常个数。
在一种可能的实现方式中,该控制设备基于该正常传输特征个数、第一正常个数以及第二正常个数,确该第一查准率以及第一查全率包括:该控制设备将该第一正常个数与该第二正常个数之间的比值确定为该第一查准率,该控制设备将该第一正常个数与该正常传输特征个数之间的比值确定为该第一查全率。
在一种可能的实现方式中,该控制设备基于该异常传输特征个数、第一异常个数以及第二异常个数,确定该第二查准率以及第二查全率包括:该控制设备将该第一异常个数与该第二异常个数之间的比值确定为该第二查准率,该控制设备将该第一异常个数与该异常传输特征个数之间的比值确定为该第二查全率。
以下述表7为例,验证集包括100个第四传输特征,其中,90个第四传输特征为正常传输特征,10个第四传输特征为异常传输特征,也即是正常数据流个数为90,异常数据流个数为10。控制设备将100个第四传输特征输入该目标模型,该目标模型输出100个第五传输特征。在执行r次K值更新过程中,对于100个第五传输特征中的任一第五传输特征,若该任一第五传输特征与对应的第四传输特征之间的差异度小于该差异度θ j,则该控制设备将任一第五传输特征的第二标签标记为第二正常标识,以指示该第五传输特征所对应的第四传输特征为正常传输特征(也即是预测的第五传输特征对应的第四传输特征为正常传输特征),否则,该控制设备将该第五传输特征的第五标签标记为第二异常标识,以指示该第五传输特征所对应的第四传输特征为异常传输特征(也即是预测的第五传输特征对应的第四传输特征为异常传输特征)。以该100个第五传输特征中第二标签为第二正常标识的第五传输特征个数为88、该100个第五传输特征中第二标签为第二异常标识的第五传输特征个数为12为例,也即是第二正常个数为88,第二异常个数为12。其中,这88个第五传输特征所对应的88个第四传输特征中,86个第四传输特征的第一标签为第一正常标识,说明这86个第四传输特征为目标正常传输特征,则该第一正常个数为86,而这88个第四传输特征中2个第四传输特征的第一标签为第一异常标识,说明这2个第四传输特征实际是异常传输特征,在差异度θ j,目标模型重构这个2个第四传输特征的结果不准确;这12个第五传输特征所对应的12个第四传输特征中,8个第四传输特征的第一标签为第一异常标识,说明这8个第四传输特征为目标异常传输特征,则该第一异常个数为8,而这12个第四传输特征中4个第四传输特征的第一标签为第一正常标识,说明这4个第四传输特征实际是正常传输特征,在差异度θ j,目标模型重构这4个第四传输特征的结果不准确,则该第一查准率=第一正常个数86/第二正常个数88,第一查全率=第一正常个数86/正常数据流个数90,该第二查准率=第一异常个数8/第二异常个数12,第一查全率=第一异常个数8/异常数据流个数10。需要说明的是,预测的第五传输特征也即是重构出的第五传输特征。
表7
Figure PCTCN2021105494-appb-000010
Figure PCTCN2021105494-appb-000011
需要是说明的是,若该目标模型的第一查准率大于或等于第一查准率阈值,则说明在差异度θ j下,该目标模型在重构正常传输特征时的输出结果基本上是准确的,若该目标模型的第一查全率大于或等于第一查全率阈值,说明该目标模型在差异度θ j下能够大概率的完全重构出正常传输特征,因此,若该目标模型的第一查准率大于或等于第一查准率阈值且该目标模型的第一查全率大于或等于第一查全率阈值,说明在该在差异度θ j下,该目标模型具有完全重构出正常传输特征的功能。若该目标模型的第二查准率大于或等于第二查准率阈值,则说明在差异度θ j下,该目标模型在重构异常传输特征时的输出结果基本上是准确的,若该目标模型的第二查全率大于或等于第二查全率阈值,说明该目标模型在差异度θ j下大概率的不能完全重构出异常传输特征,因此,若该目标模型的第二查准率大于或等于第二查准率阈值且该目标模型的第二查全率大于或等于第二查全率阈值,说明在该在差异度θ j下,该目标模型不具有完全重构出异常传输特征的功能,所以,若在差异度θ j下该目标模型符合第一条件,则认为该差异度θ j为可靠的,该控制设备将该差异度θ j确定为目标差异度。
在一种可能的实现方式中,由用户根据经验来设定该目标差异度,无须通过上述公式(3)以及更新的方式,确定该目标差异度。
510、该控制设备将目标终端类型下多个目标终端的多个第六传输特征输入该目标模型,由该目标模型基于输入的该多个第六传输特征,重构每个第六传输特征,输出多个第七传输特征,该多个第六传输特征与该多个第七传输特征一一对应,一个第六传输特征为一个目标终端传输的至少一个数据流的总体传输特征。
该多个目标终端包括至少一个第二终端以及至少一个第三终端,其中,该至少一个第二终端均为正常终端,该至少一个第三终端均为异常终端。一个第七传输特征为重构出的一个第六传输特征。
测试集除了包括至少一个第二终端的多个传输特征以外,还包括至少一个第三终端的多个传输特征,在该测试集中第二终端的传输特征和第三终端的传输特征均视为第六传输特征。该控制设备从测试集中获取多个第六传输特征,并将获取到的多个第六传输特征输入该目标模型,由目标模型基于输入的多个第六传输特征,输出重构出的多个第六传输特征,其中,该多个第六传输特征属于至少一个第二终端和至少一个第三终端,重构出的多个第六传输特征也即是该多个第七传输特征,每个第七传输特征对应一个第六传输特征。
511、该控制设备基于该多个第七传输特征与对应的第六传输特征之间的差异度,确定该目标模型通过验证。
该控制设备基于该多个第七传输特征与对应的第六传输特征之间的差异度,确定该目标模型是否符合第二条件,若该目标模型符合第二条件,则该控制设备确定该目标模型通过验证,否则,该控制设备确定该目标模型未通过验证,若该目标模型未通过验证,则该控制设备跳转执行步骤507-511,直至最终获取的目标模型能够通过验证。
其中,第二条件包括下述至少一项:该目标差异度下,该目标模型的第一查准率大于或等于第三查准率阈值,且该目标模型的第一查全率大于或等于第三查全率阈值;在该目标差异度下,该目标模型的第二查准率大于或等于第四查准率阈值,且该目标模型的第二查全率大于或等于第四查全率阈值。本申请实施例对第三查准率阈值、第三查全率阈值、第四查准率阈值和第四查全率阈值的不作具体限定。
其中,第一查准率、第一查全率、第二查准率、第二查全率的计算方式在步骤5093中有相关介绍,在此不做赘述。
需要说明的是,当该目标模型通过验证时,该控制设备将该目标模型添加至模型库中,并为该目标模型分配模型标识,该模型库用于指示至少一个终端类型所对应的目标模型。每个终端类型对应一个目标模型。
该控制设备还能将该目标模型与该目标终端类型进行关联存储。在一种可能的实现方式中,当该目标模型通过验证时,该控制设备将该目标模型的模型标识与该目标终端类型进行关联存储至目标模型列表,以便查询,其中,该目标模型列表用于记录至少一个终端类型所对应的目标模型。例如,该目标终端类型为摄像头,该目标终端类型对应的目标模型为目标模型3,则该控制设备将摄像头和目标模型3关联存储在表8所示的目标模型列表。
表8
模型标识 终端类型
目标模型1 音响
目标模型2 ATM
目标模型3 摄像头
需要说明的是,在本申请实施例中,是由控制设备进行训练,以获取目标模型,而在另外的一些实施例中,网络设备也能进行训练,以获取目标模型,网络设备进行训练的过程与控制设备进行训练的过程同理,在此,本申请实施例对网络设备进行训练的过程不做赘述。
本申请实施例提供的方法,通过该控制设备将该目标终端类型对应的网络地址下发到至少一个网络设备,由每个网络设备基于下发的每个网络地址,收集每个网络地址所指示的第二终端的传输特征,再由控制设备基于每个网络设备上传的至少一个第二终端的多个传输特征,训练得到该目标终端类型对应的目标模型,以便后续基于该目标模型,验证该目标终端类型下的终端是否为异常终端。
该控制设备将该目标模型发送至少一个网络设备,由各个网络设备基于该目标模型对该目标终端类型下的终端进行验证,为了进一步说明该过程,参见图8所示的本申请实施例提供的一种用于实现终端验证的方法的流程图。
801、该控制设备向至少一个网络设备发送模型存储指令,该模型存储指令用于指示存储该目标模型。
该模型存储指令包括该目标模型、目标终端类型以及该目标终端类型对应的至少一个网络地址、目标差异度以及存储标识,该目标终端类型也即是第二终端的终端类型,该至少一 个网络地址也即是配置表中该目标终端类型所对应的网络地址,也即是至少一个第二终端的网络地址,该存储标识用于指示存储该目标模型。
在一种可能的实现方式中,用户在该控制设备的模型更新界面内输入该目标终端类型,当该控制设备在该模型更新界面检测到用户执行了确认操作时,则该触发该控制设备执行以下操作:该控制设备基于该模型更新界面中输入的该目标终端类型,在资产表中查询该目标终端类型对应的目标模型的模型标识,在配置表中查询该目标终端类型对应的至少一个网络地址,并在模型库中获取该模型标识对应的目标模型,该控制设备并将该目标终端类型、至少一个网络地址、目标模型以及存储标识添加在该模型存储指令中,并向该至少一个网络设备发送该模型存储指令。
802、对于该至少一个网络设备中的任一网络设备,该任一网络设备接收该模型存储指令。
803、该任一网络设备存储该模型存储指令中的目标模型。
该任一网络设备将该模型存储指令中的该目标模型、目标终端类型、该至少一个网络地址以及目标差异度进行关联存储。
804、该任一网络设备获取第一终端的第一传输特征,该第一传输特征为该第一终端传输的至少一个第一数据流的总体传输特征。
该第一终端的终端类型为该目标终端类型,该第一终端也即是目标终端类型的被验证终端。该第一终端的网络地址为该至少一个网络地址中的任一个,也即是该至少一个第二终端中任一第二终端的网络地址与该第一终端的网络地址相同。该至少一个第一数据流均携带该第一终端的网络地址。该第一传输特征也即是该第一终端的一个传输特征,也即是该一个时间窗口内该第一终端传输的至少一个第一数据流的总体传输特征。
若该第一终端不是预设业务所使用的终端,则该第一终端传输的报文所携带的第一终端的网络地址为一个第二终端的网络地址,若该第一终端是预设业务所使用的终端,则该第一终端为该至少一个第二终端中的任一个。而为了验证该第一终端是否是异常终端,该任一网络设备还需基于该第一终端的第一传输特征,来验证该第一终端是否为异常终端,因此,该任一网络设备还需获取该第一终端的第一传输特征。
对于该至少一个网络地址中的任一网络地址,当该任一网络设备接收到携带该任一网络地址的至少一个数据流时,该任一网络设备将该至少一个数据流确定为至少一个第一数据流,该任一网络设备将输出该至少一个第一数据流的终端,确定为第一终端;该任一网络设备基于该至少一个第一数据流,确定该第一终端的第一传输特征。
其中,该任一网络设备基于该至少一个第一数据流,确定该第一终端的第一传输特征的过程与上述步骤504中任一网络设备确定第二终端的一个传输特征的过程同理,在此,本申请实施例对该任一网络设备基于该至少一个第一数据流,确定该第一终端的第一传输特征的过程不做赘述。
805、该任一网络设备将该第一传输特征输入目标模型,由该目标模型基于输入的该第一传输特征,重构该第一传输特征,输出第二传输特征。
其中,该目标模型与目标终端类型对应,第二传输特征也即是该目标模型重构出的第一传输特征。本步骤805所示的过程也即是该任一网络设备基于该第一传输特征,重构该第一传输特征,得到第二传输特征的过程。
806、若该第一传输特征与该第二传输特征之间的差异度大于或等于目标差异度,该任一 网络设备确定该第一终端未通过验证,该目标差异度用于指示该第一传输特征与该第二传输特征之间的差异情况。
该第一传输特征与该第二传输特征之间的差异度为该第二传输特征与对应该第一传输特征中相应特征之间的均方误差。
该任一网络设备基于上述公式(2)计算,该第二传输特征与对应该第一传输特征中相应特征之间的均方误差,并将该均方误差确定为该第一传输特征与该第二传输特征之间的差异度;该任一网络设备将该第一传输特征与该第二传输特征之间的差异度与该目标差异度进行比较,以确定该第一传输特征与该第二传输特征之间的差异度是否大于或等于目标差异度;若该第一传输特征与该第二传输特征之间的差异度大于或等于目标差异度,则说明该目标模型没有完全重构出该第一传输特征,该第一传输特征为异常传输特征,该第一终端的行为异常,该第一终端是异常终端,则该任一网络设备确定该第一终端未通过验证;若该第一传输特征与该第二传输特征之间的差异度小于目标差异度,则说明该目标终端完全重构出该第一传输特征,该第一传输特征为正常传输特征,该第一终端的行为正常,该第一终端是正常终端,则该任一网络设备确定该第一终端通过验证。
807、该任一网络设备向该控制设备发送对该第一终端的验证结果,该验证结果用于指示该第一终端是否通过验证。
该验证结果包括验证标识以及该第一终端的网络地址。该验证标识用于指示该第一终端是否通过验证,该验证标识包括第一验证标识或第二验证标识,其中,第一验证标识用于指示该第一终端通过验证,该第二验证标识用于指示该第一终端未通过验证。若该第一终端通过验证,则该验证标识为第一验证标识,若该第一终端未通过验证,则该验证标识为第二验证标识。
808、该控制设备接收该验证结果。
809、若该第一终端未通过验证,则该控制设备显示第一提示信息,并向该任一网络设备发送断开接入指令
该第一提示信息用于提示用户该第一终端未通过验证,该第一提示信息包括该第二验证标识、第一终端的网络地址以及警告标识,该警告标识用于提示用户该第一终端未通过验证,也即是用于提示用户该第一终端为异常终端。该断开接入指令用于指示任一网络设备断开与该第一终端之间的连接,该断开接入指令包括该第一终端的网络地址以及断开标识,该断开标识用于指示任一网络设备断开与该第一终端之间的连接。
在一种可能的实现方式中,若该验证结果中的验证标识为第二验证标识,则触发该控制设备显示第一提示信息,并向该任一网络设备发送该断开接入指令。
在另一种可能的实现方式中,若该验证结果中的验证标识为第二验证标识,则触发该控制设备显示第一提示信息,用户可在该控制设备查阅该第一提示信息,若用户在该控制设备上还执行了用户触发该断开接入指令的操作,则触发该控制设备向该任一网络设备发送该断开接入指令。
若该第一终端通过验证,则该控制设备显示第二提示信息,该第二提示信息用于提示用户该第一终端通过验证,也即是用于提示该第一终端是正常终端,该第二提示信息包括该第一终端的IP地址以及第一验证标识。
810、该任一网络设备接收该断开接入指令,并基于该断开接入指令断开与该第一终端的 之间连接。
当该任一网络设备接收到该断开接入指令后,该任一网络设备从该断开接入指令中获取网络地址,并断开与该网络地址所指示的第一终端之间的连接。
在一种可能的实现中,该任一网络设备也能够显示该第一提示信息提示用户该第一终端未通过验证。在一种可能的实现中,若用户在该任一网络设备上进行了用于触发断开与该第一终端的连接的操作,则触发该任一网络设备直接断开与第一总段之间连接,而无须等待控制设备下发断开接入指令。
本申请实施例提供的方法,通过重构终端的传输特征,来对终端进行验证,例如若重构出的传输特征与终端的传输特征之间的差异较大,则说明该终端的传输特征出现异常,该终端为异常终端,则确定该终端未通过验证,由于终端具有特定的正常传输特征,正常传输特征不易仿冒,因此,该方法能够准确地验证出各种异常终端,提高了终端验证的准确度,而仿冒终端为异常终端的一种,因此,该方法也能够准确的验证出仿冒终端,而不是通过简单的对该终端的IP地址进行验证的方式,以防对仿冒终端验证通过。
为了进一步说明图3、图5和图8所示的过程,参见图9所示的本申请实施例提供的一种用于实现终端验证的方法的示意图。网络设备在接收到终端传输的数据流时,根据该数据流中的报文,获取该数据流的传输信息,并将传输信息上传至控制设备,由该控制设备提取该传输信息中终端的网络地址(也即是资产识别),以实现配置终端配置,并由用户基于该终端的网络地址,在资产表中标注该终端的终端类型(例如将该终端的网络地址与该终端的终端类型关联存储)。,该控制设备将该目标终端类型对应的终端的网络地址信息(也即是设备信息)下发至网络设备,由网络设备基于下发的网络地址匹配对应设备类型的终端传输的至少一个数据流,并获取该至少一个数据流的流传输特征,该网络设备基于该至少一个数据流的流传输特征,获取对应终端的传输特征,并将对应终端的传输特征发送至控制设备中的样本库,由控制设备基于样本库中该目标终端类型下至少一个终端的传输特征,训练得到该目标终端类型对应的目标模型,并将该目标模型下发至网络设备,由网络设备基于该目标模型对该终端类型下的终端进行验证,若终端未通过验证,则显示警告信息也即是第一提示信息,由用户向该网络设备下发隔离策略,例如断开与未通过验证的终端之间的连接,则该网络设备接收到该隔离策略后,断开与未通过验证的终端之间的连接。
图10是本申请实施例提供的一种用于实现终端验证的装置的结构示意图,该装置1000包括:
第一获取模块1001,用于获取第一终端的第一传输特征,所述第一传输特征为所述第一终端传输的至少一个第一数据流的总体传输特征;
第一重构模块1002,用于基于所述第一传输特征,对所述第一传输特征进行重构,得到第二传输特征,所述第二传输特征为重构出的所述第一传输特征;
确定模块1003,用于若所述第一传输特征与所述第二传输特征之间的差异度大于或等于目标差异度,确定所述第一终端未通过验证。
可选地,所述第一传输特征包括上行传输特征,所述上行传输特征为所述至少一个第一数据流中至少一个上行数据流的总体传输特征。
可选地,所述上行传输特征包括上行报文特征、上行流特征中的至少一个,所述上行报文特征为所述至少一个上行数据流中上行报文的总体特征,所述上行流特征为所述至少一个 上行数据流的统计特征。
可选地,所述上行报文特征包括上行报文平均传输间隔、上行负载平均值、上行总负载大小、上行报文个数、上行目标报文个数、上行目标报文占比、上行报文负载波动值中的至少一个,所述上行报文平均传输间隔为在一个时间窗口内所述上行报文的平均传输间隔,所述上行负载平均值为在所述时间窗口内所述至少一个上行数据流中目标报文的负载的平均大小,所述上行总负载大小为在所述时间窗口内所述至少一个上行数据流中目标报文的负载的总大小,所述上行报文个数为在所述时间窗口内所述至少一个上行数据流中上行报文的个数,所述上行目标报文个数在所述时间窗口内所述至少一个上行数据流中目标报文的个数,所述上行目标报文占比为在所述时间窗口内所述至少一个上行数据流中目标报文的占比,所述上行报文负载波动值用于指示在所述时间窗口内所述至少一个上行数据流中目标报文的大小波动情况;
所述上行流特征包括上行终端端口波动值、上行数据流总个数、上行目标数据流个数、至少一种数据流类型中每种数据流类型下的上行数据流个数、至少一种传输协议类型中每种传输协议类型下的上行数据流个数中的至少一个,所述上行终端端口波动值用于指示在所述时间窗口内所述第一终端中所述至少一个上行数据流的输出端口的波动情况,所述上行目标数据流为所对应的服务器输入端口属于目标端口范围的上行数据流。
可选地,所述上行报文特征还包括第一接收窗口波动值、第一接收窗口大小平均值中的至少一个,所述第一接收窗口波动值用于指示在所述时间窗口内所述上行报文携带的接收窗口大小的波动情况。
可选地,所述第一接收窗口波动值为在所述时间窗口内所述上行报文携带的接收窗口大小的标准差。
可选地,所述第一传输特征还包括所述至少一个第一数据流的总个数、下行传输特征中的至少一个,所述下行传输特征为所述至少一个第一数据流中至少一个下行数据流的总体传输特征。
可选地,所述下行传输特征包括下行报文特征、下行流特征中的至少一个,所述下行报文特征为所述至少一个下行数据流中下行报文的总体特征,所述下行流特征为所述至少一个下行数据流的统计特征。
可选地,所述下行报文特征包括下行报文平均传输间隔、下行负载平均值、下行总负载大小、下行报文个数、下行目标报文个数、下行目标报文占比、下行报文负载波动值中的至少一个,所述下行报文平均传输间隔为在一个时间窗口内所述下行报文的平均传输间隔,所述下行负载平均值为在所述时间窗口内所述至少一个下行数据流中目标报文的负载的平均大小,所述下行总负载大小为在所述时间窗口内所述至少一个下行数据流中目标报文的负载的总大小,所述下行报文个数为在所述时间窗口内所述至少一个下行数据流中下行报文的个数,所述下行目标报文个数在所述时间窗口内所述至少一个下行数据流中目标报文的个数,所述下行目标报文占比为在所述时间窗口内所述至少一个下行数据流中目标报文的占比,所述下行报文负载波动值用于指示在所述时间窗口内所述至少一个下行数据流中目标报文的大小波动情况;
所述下行流特征包括下行终端端口波动值、下行数据流总个数、下行目标数据流个数、至少一种数据流类型中每种数据流类型下的下行数据流个数、至少一种传输协议类型中每种 传输协议类型下的下行数据流个数中的至少一个,所述下行终端端口波动值用于指示在所述时间窗口内所述第一终端中所述至少一个下行数据流的输入端口的波动情况,所述下行目标数据流为对应的服务器输出端口属于目标端口范围的下行数据流。
可选地,所述下行报文特征还包括第二接收窗口波动值、第二接收窗口大小平均值中的至少一个,所述第二接收窗口波动值用于指示所述下行报文携带的接收窗口大小的波动情况。
可选地,所述第二接收窗口波动值为在所述时间窗口内所述下行报文携带的接收窗口大小的标准差。
所述第一获取模块1001用于:
获取所述至少一个第一数据流中每个第一数据流的流传输特征;
基于所述至少一个第一数据流的流传输特征,获取所述第一传输特征。
可选地,一个第一数据流的流传输特征包括所述第一数据流的传输信息、数据流类型、目标端口类型、报文特征中的至少一个,所述传输信息用于指示数据流的传输属性,所述目标端口类型为服务器中传输所述第一数据流的端口的端口类型,所述报文特征为所述第一数据流中报文的特征。
可选地,所述传输信息包括所述第一数据流的五元组中的至少一元;
所述报文特征包括报文传输间隔总和、负载大小、负载平方和、目标报文个数、报文总个数、接收窗口波动值、窗口总大小、窗口平方和中的至少一个,所述报文传输间隔总和为在一个时间窗口内所述第一数据流中报文之间的传输间隔的总时长,所述负载大小为在所述时间窗内所述第一数据流中目标报文的负载的总大小,所述负载平方和为所述目标报文的负载的大小平方和,所述目标报文个数为在所述时间窗口内所述报文中目标报文的总个数,所述报文总个数为在所述时间窗口内所述报文的总个数,所述接收窗口波动值用于指示在所述时间窗口内所述报文携带的接收窗口大小的波动情况,所述窗口总大小为在所述时间窗口内所述报文携带的接收窗口大小的总和,所述窗口平方和为所述滑动窗口的大小的平方和。
可选地,所述传输信息还包括方向标识、所述时间窗口的标识中的至少一个,所述方向标识用于指示所述第一数据流的传输方向。
可选地,所述第一重构模块1002用于:
将所述第一传输特征输入目标模型,由所述目标模型基于输入的所述第一传输特征,重构所述第一传输特征,输出所述第二传输特征。
可选地,所述装置1000还包括:
训练模块,用于将目标终端类型下至少一个第二终端的多个第三传输特征作为初始模型的输入和输出,进行训练,得到所述目标模型,所述目标终端类型为所述第一终端的终端类型,一个第三传输特征为一个第二终端传输的至少一个数据流的总体传输特征。
可选地,所述装置1000还包括:
第二重构模块,用于将目标终端模型下的多个目标终端的多个第四传输特征输入所述目标模型,由所述目标模型基于输入的所述多个第四传输特征,重构每个第四传输特征,输出多个第五传输特征,所述目标终端类型为所述第一终端的终端类型,所述多个第四传输特征与所述多个第五传输特征一一对应,一个第四传输特征为一个目标终端传输的至少一个数据流的总体传输特征;
第二获取模块,用于基于所述多个第五传输特征与所述多个第四传输特征,获取所述目 标差异度。
可选地,所述第二获取模块用于:
基于所述多个第五传输特征中至少一个第五传输特征与对应的第四传输特征之间的差异度,获取所述目标差异度。
可选地,所述装置1000还包括:
第三重构模块,用于将目标终端类型下多个目标终端的多个第六传输特征输入所述目标模型,由所述目标模型基于输入的所述多个第六传输特征,重构每个第六传输特征,输出多个第七传输特征,所述目标终端类型为所述第一终端的终端类型,所述多个第六传输特征与所述多个第七传输特征一一对应,一个第六传输特征为一个目标终端传输的至少一个数据流的总体传输特征;
所述确定模块1003,还用于基于所述多个第七传输特征与对应的第六传输特征之间的差异度,确定所述目标模型通过验证。
可选地,所述装置1000还包括:
第三获取模块,用于获取所述至少一个第二终端传输的至少一个第二数据流的传输信息,所述传输信息用于指示数据流的传输属性;
存储模块,用于将所述传输信息与所述第二终端的终端类型进行关联存储;
所述第三获取模块,还用于基于所述终端类型关联的传输信息,获取所述至少一个第二终端的多个传输特征,一个传输特征为终端传输的至少一个数据流的总体传输特征。
可选地,所述装置1000还包括:
接收模块,用于从控制设备接收所述目标模型。
可选地,所述装置1000为控制设备或网络设备。
所述装置1000通过重构终端的传输特征,来对终端进行验证,例如若重构出的传输特征与终端的传输特征之间的差异较大,则说明该终端的传输特征出现异常,该终端为异常终端,则确定该终端未通过验证,由于终端具有特定的正常传输特征,正常传输特征不易仿冒,因此,所述装置1000能够准确地验证出各种异常终端,提高了终端验证的准确度,而仿冒终端为异常终端的一种,因此,所述装置1000也能够准确的验证出仿冒终端,而不是通过简单的对该终端的IP地址进行验证的方式,以防对仿冒终端验证通过。
图11是本申请实施例提供的一种用于实现终端验证的装置,所述1100装置包括:
获取模块1101,用于获取目标终端类型的至少一个第二终端的多个第三传输特征,一个第三传输特征为一个第二终端传输的至少一个数据流的总体传输特征;
训练模块1102,用于将所述多个第三传输特征作为初始模型的输入和输出,进行训练,得到目标模型,所述目标模型用于重构所述目标终端类型的被验证终端的传输特征,以对所述被验证终端进行验证,所述传输特征为所述被验证终端传输的至少一个数据流的总体传输特征。
可选地,所述装置1100还包括:
发送模块1103,用于向网络设备发送所述目标模型。
可选地,所述装置1100还包括:
第一重构模块,用于将所述目标终端类型下多个目标终端的多个第四传输特征输入所述目标模型,由所述目标模型基于输入的所述多个第四传输特征,重构每个第四传输特征,输 出多个第五传输特征,所述多个第四传输特征与所述多个第五传输特征一一对应,一个第四传输特征为一个目标终端传输的至少一个数据流的总体传输特征;
第一目标获取模块,还用于基于所述多个第五传输特征与所述多个第四传输特征,获取所述目标差异度。
可选地,所述第一目标获取模块用于:
基于所述多个第五传输特征中至少一个第五传输特征与对应的第四传输特征之间的差异度,获取所述目标差异度。
可选地,所述装置1100还包括:
第二重构模块,用于将所述目标终端类型下多个目标终端的多个第六传输特征输入所述目标模型,由所述目标模型基于输入的所述多个第六传输特征,重构每个第六传输特征,输出多个第七传输特征,所述多个第六传输特征与所述多个第七传输特征一一对应,一个第六传输特征为一个目标终端传输的至少一个数据流的总体传输特征;
确定模块,用于基于所述多个第七传输特征与对应的第六传输特征之间的差异度,确定所述目标模型通过验证。
可选地,所述装置1100还包括:
第二目标获取模块,用于获取所述至少一个第二终端传输的至少一个第二数据流的传输信息,所述传输信息用于指示数据流的传输属性;
存储模块,用于将所述传输信息与所述第二终端的终端类型进行关联存储;
所述第二目标获取模块,还用于基于所述终端类型关联的传输信息,获取所述至少一个第二终端的多个传输特征,一个传输特征为终端传输的至少一个数据流的总体传输特征。
上述所有可选技术方案,可以采用任意结合形成本公开的可选实施例,在此不再一一赘述。
需要说明的是:上述实施例提供的用于实现终端验证的装置在验证终端时,仅以上述各功能模块的划分进行举例说明,实际应用中,可以根据需要而将上述功能分配由不同的功能模块完成,即将装置的内部结构划分成不同的功能模块,以完成以上描述的全部或者部分功能。另外,上述实施例提供的用于实现终端验证的方法实施例属于同一构思,其具体实现过程详见方法实施例,这里不再赘述。
本申请实施例还提供了一种计算机程序产品或计算机程序,该计算机程序产品或计算机程序包括计算机指令,该计算机指令存储在计算机可读存储介质中,电子设备的处理器从计算机可读存储介质读取该计算机指令,处理器执行该计算机指令,使得该电子设备执行上述用于实现终端验证的方法。
本领域普通技术人员可以理解实现上述实施例的全部或部分步骤可以通过硬件来完成,也可以通过程序来指令相关的硬件完成,所述的程序可以存储于一种计算机可读存储介质中,上述提到的存储介质可以是只读存储器,磁盘或光盘等。
以上所述仅为本申请的示例性实施例,并不用以限制本申请,凡在本申请的精神和原则之内,所作的任何修改、等同替换、改进等,均应包含在本申请的保护范围之内。

Claims (37)

  1. 一种用于实现终端验证的方法,其特征在于,所述方法包括:
    获取第一终端的第一传输特征,所述第一传输特征为所述第一终端传输的至少一个第一数据流的总体传输特征;
    基于所述第一传输特征,对所述第一传输特征进行重构,得到第二传输特征,所述第二传输特征为重构出的所述第一传输特征;
    若所述第一传输特征与所述第二传输特征之间的差异度大于或等于目标差异度,确定所述第一终端未通过验证。
  2. 根据权利要求1所述的方法,其特征在于,所述第一传输特征包括上行传输特征,所述上行传输特征为所述至少一个第一数据流中至少一个上行数据流的总体传输特征。
  3. 根据权利要求2所述的方法,其特征在于,所述上行传输特征包括上行报文特征、上行流特征中的至少一个,所述上行报文特征为所述至少一个上行数据流中上行报文的总体特征,所述上行流特征为所述至少一个上行数据流的统计特征。
  4. 根据权利要求3所述的方法,其特征在于,所述上行报文特征包括上行报文平均传输间隔、上行负载平均值、上行总负载大小、上行报文个数、上行目标报文个数、上行目标报文占比、上行报文负载波动值中的至少一个,所述上行报文平均传输间隔为在一个时间窗口内所述上行报文的平均传输间隔,所述上行负载平均值为在所述时间窗口内所述至少一个上行数据流中目标报文的负载的平均大小,所述上行总负载大小为在所述时间窗口内所述至少一个上行数据流中目标报文的负载的总大小,所述上行报文个数为在所述时间窗口内所述至少一个上行数据流中上行报文的个数,所述上行目标报文个数在所述时间窗口内所述至少一个上行数据流中目标报文的个数,所述上行目标报文占比为在所述时间窗口内所述至少一个上行数据流中目标报文的占比,所述上行报文负载波动值用于指示在所述时间窗口内所述至少一个上行数据流中目标报文的大小波动情况;
    所述上行流特征包括上行终端端口波动值、上行数据流总个数、上行目标数据流个数、至少一种数据流类型中每种数据流类型下的上行数据流个数、至少一种传输协议类型中每种传输协议类型下的上行数据流个数中的至少一个,所述上行终端端口波动值用于指示在所述时间窗口内所述第一终端中所述至少一个上行数据流的输出端口的波动情况,所述上行目标数据流为所对应的服务器输入端口属于目标端口范围的上行数据流。
  5. 根据权利要求4所述的方法,其特征在于,所述上行报文特征还包括第一接收窗口波动值、第一接收窗口大小平均值中的至少一个,所述第一接收窗口波动值用于指示在所述时间窗口内所述上行报文携带的接收窗口大小的波动情况。
  6. 根据权利要求2-5任一项权利要求所述的方法,其特征在于,所述第一传输特征还包 括所述至少一个第一数据流的总个数、下行传输特征中的至少一个,所述下行传输特征为所述至少一个第一数据流中至少一个下行数据流的总体传输特征。
  7. 根据权利要求6所述的方法,其特征在于,所述下行传输特征包括下行报文特征、下行流特征中的至少一个,所述下行报文特征为所述至少一个下行数据流中下行报文的总体特征,所述下行流特征为所述至少一个下行数据流的统计特征。
  8. 根据权利要求7所述的方法,其特征在于,所述下行报文特征包括下行报文平均传输间隔、下行负载平均值、下行总负载大小、下行报文个数、下行目标报文个数、下行目标报文占比、下行报文负载波动值中的至少一个,所述下行报文平均传输间隔为在一个时间窗口内所述下行报文的平均传输间隔,所述下行负载平均值为在所述时间窗口内所述至少一个下行数据流中目标报文的负载的平均大小,所述下行总负载大小为在所述时间窗口内所述至少一个下行数据流中目标报文的负载的总大小,所述下行报文个数为在所述时间窗口内所述至少一个下行数据流中下行报文的个数,所述下行目标报文个数在所述时间窗口内所述至少一个下行数据流中目标报文的个数,所述下行目标报文占比为在所述时间窗口内所述至少一个下行数据流中目标报文的占比,所述下行报文负载波动值用于指示在所述时间窗口内所述至少一个下行数据流中目标报文的大小波动情况;
    所述下行流特征包括下行终端端口波动值、下行数据流总个数、下行目标数据流个数、至少一种数据流类型中每种数据流类型下的下行数据流个数、至少一种传输协议类型中每种传输协议类型下的下行数据流个数中的至少一个,所述下行终端端口波动值用于指示在所述时间窗口内所述第一终端中所述至少一个下行数据流的输入端口的波动情况,所述下行目标数据流为对应的服务器输出端口属于目标端口范围的下行数据流。
  9. 根据权利要求8所述的方法,其特征在于,所述下行报文特征还包括第二接收窗口波动值、第二接收窗口大小平均值中的至少一个,所述第二接收窗口波动值用于指示所述下行报文携带的接收窗口大小的波动情况。
  10. 根据权利要求1-9任一项权利要求所述的方法,其特征在于,所述获取第一终端的第一传输特征包括:
    获取所述至少一个第一数据流中每个第一数据流的流传输特征;
    基于所述至少一个第一数据流的流传输特征,获取所述第一传输特征。
  11. 根据权利要求10所述的方法,其特征在于,一个第一数据流的流传输特征包括所述第一数据流的传输信息、数据流类型、目标端口类型、报文特征中的至少一个,所述传输信息用于指示数据流的传输属性,所述目标端口类型为服务器中传输所述第一数据流的端口的端口类型,所述报文特征为所述第一数据流中报文的特征。
  12. 根据权利要求11所述的方法,其特征在于,所述传输信息包括所述第一数据流的五元组中的至少一元;
    所述报文特征包括报文传输间隔总和、负载大小、负载平方和、目标报文个数、报文总个数、接收窗口波动值、窗口总大小、窗口平方和中的至少一个,所述报文传输间隔总和为在一个时间窗口内所述第一数据流中报文之间的传输间隔的总时长,所述负载大小为在所述时间窗内所述第一数据流中目标报文的负载的总大小,所述负载平方和为所述目标报文的负载大小的平方和,所述目标报文个数为在所述时间窗口内所述报文中目标报文的总个数,所述报文总个数为在所述时间窗口内所述报文的总个数,所述接收窗口波动值用于指示在所述时间窗口内所述报文携带的接收窗口大小的波动情况,所述窗口总大小为在所述时间窗口内所述报文携带的接收窗口大小的总和,所述窗口平方和为所述滑动窗口的大小的平方和。
  13. 根据权利要求12所述的方法,其特征在于,所述传输信息还包括方向标识、所述时间窗口的标识中的至少一个,所述方向标识用于指示所述第一数据流的传输方向。
  14. 根据权利要求1-13任一项权利要求所述的方法,其特征在于,所述基于所述第一传输特征,对所述第一传输特征进行重构,得到第二传输特征包括:
    将所述第一传输特征输入目标模型,由所述目标模型基于输入的所述第一传输特征,重构所述第一传输特征,输出所述第二传输特征。
  15. 根据权利要求14所述的方法,其特征在于,所述将所述第一传输特征输入目标模型之前,所述方法还包括:
    将目标终端类型下至少一个第二终端的多个第三传输特征作为初始模型的输入和输出,进行训练,得到所述目标模型,所述目标终端类型为所述第一终端的终端类型,一个第三传输特征为一个第二终端传输的至少一个数据流的总体传输特征。
  16. 根据权利要求14或15所述的方法,其特征在于,所述将所述第一传输特征输入目标模型之前,所述方法还包括:
    将目标终端类型下多个目标终端的多个第四传输特征输入所述目标模型,由所述目标模型基于输入的所述多个第四传输特征,重构每个第四传输特征,输出多个第五传输特征,所述目标终端类型为所述第一终端的终端类型,所述多个第四传输特征与所述多个第五传输特征一一对应,一个第四传输特征为一个目标终端传输的至少一个数据流的总体传输特征;
    基于所述多个第五传输特征与所述多个第四传输特征,获取所述目标差异度。
  17. 根据权利要求16所述的方法,其特征在于,所述基于所述多个第五传输特征与所述多个第四传输特征,获取所述目标差异度包括:
    基于所述多个第五传输特征中至少一个第五传输特征与对应的第四传输特征之间的差异度,获取所述目标差异度。
  18. 根据权利要求14-17任一项权利要求所述的方法,其特征在于,所述将所述第一传输特征输入目标模型之前,所述方法还包括:
    将目标终端类型下多个目标终端的多个第六传输特征输入所述目标模型,由所述目标模 型基于输入的所述多个第六传输特征,重构每个第六传输特征,输出多个第七传输特征,所述目标终端类型为所述第一终端的终端类型,所述多个第六传输特征与所述多个第七传输特征一一对应,一个第六传输特征为一个目标终端传输的至少一个数据流的总体传输特征;
    基于所述多个第七传输特征与对应的第六传输特征之间的差异度,确定所述目标模型通过验证。
  19. 根据权利要求14-18任一项权利要求所述的方法,其特征在于,所述将所述第一传输特征输入目标模型之前,所述方法还包括:
    获取所述至少一个第二终端传输的至少一个第二数据流的传输信息,所述传输信息用于指示数据流的传输属性;
    将所述传输信息与所述第二终端的终端类型进行关联存储;
    基于所述终端类型关联的传输信息,获取所述至少一个第二终端的多个传输特征,一个传输特征为终端传输的至少一个数据流的总体传输特征。
  20. 根据权利要求14所述的方法,其特征在于,所述将所述第一传输特征输入目标模型之前,所述方法还包括:
    从控制设备接收所述目标模型。
  21. 根据权利要求1-19任一项权利要求所述的方法,其特征在于,所述方法的执行主体为控制设备或网络设备。
  22. 一种用于实现终端验证的方法,其特征在于,所述方法包括:
    获取目标终端类型的至少一个第二终端的多个第三传输特征,一个第三传输特征为一个第二终端传输的至少一个数据流的总体传输特征;
    将所述多个第三传输特征作为初始模型的输入和输出,进行训练,得到目标模型,所述目标模型用于重构所述目标终端类型的被验证终端的传输特征,以对所述被验证终端进行验证,所述传输特征为所述被验证终端传输的至少一个数据流的总体传输特征。
  23. 一种用于实现终端验证的装置,其特征在于,所述装置包括:
    第一获取模块,用于获取第一终端的第一传输特征,所述第一传输特征为所述第一终端传输的至少一个第一数据流的总体传输特征;
    第一重构模块,用于基于所述第一传输特征,对所述第一传输特征进行重构,得到第二传输特征,所述第二传输特征为重构出的所述第一传输特征;
    确定模块,用于若所述第一传输特征与所述第二传输特征之间的差异度大于或等于目标差异度,确定所述第一终端未通过验证。
  24. 根据权利要求23所述的装置,其特征在于,所述第一传输特征包括上行传输特征,所述上行传输特征为所述至少一个第一数据流中至少一个上行数据流的总体传输特征。
  25. 根据权利要求24所述的装置,其特征在于,所述上行传输特征包括上行报文特征、上行流特征中的至少一个,所述上行报文特征为所述至少一个上行数据流中上行报文的总体特征,所述上行流特征为所述至少一个上行数据流的统计特征。
  26. 根据权利要求24或25所述的装置,其特征在于,所述第一传输特征还包括所述至少一个第一数据流的总个数、下行传输特征中的至少一个,所述下行传输特征为所述至少一个第一数据流中至少一个下行数据流的总体传输特征。
  27. 根据权利要求26所述的装置,其特征在于,所述下行传输特征包括下行报文特征、下行流特征中的至少一个,所述下行报文特征为所述至少一个下行数据流中下行报文的总体特征,所述下行流特征为所述至少一个下行数据流的统计特征。
  28. 根据权利要求23-27任一项权利要求所述的装置,其特征在于,所述第一获取模块用于:
    获取所述至少一个第一数据流中每个第一数据流的流传输特征;
    基于所述至少一个第一数据流的流传输特征,获取所述第一传输特征。
  29. 根据权利要求28所述的装置,其特征在于,一个第一数据流的传输特征流传输特征包括所述第一数据流的传输信息、数据流类型、目标端口类型、报文特征中的至少一个,所述传输信息用于指示数据流的传输属性,所述目标端口类型为服务器中传输所述第一数据流的端口的端口类型,所述报文特征用于指示特征为所述第一数据流中报文的特点特征。
  30. 根据权利要求23-29任一项权利要求所述的装置,其特征在于,所述第一重构模块用于:
    将所述第一传输特征输入目标模型,由所述目标模型基于输入的所述第一传输特征,重构所述第一传输特征,输出所述第二传输特征。
  31. 根据权利要求30所述的装置,其特征在于,所述装置还包括:
    训练模块,用于将目标终端类型下至少一个第二终端的多个第三传输特征作为初始模型的输入和输出,进行训练,得到所述目标模型,所述目标终端类型为所述第一终端的终端类型,一个第三传输特征为一个第二终端传输的至少一个数据流的总体传输特征。
  32. 根据权利要求30或31所述的装置,其特征在于,所述装置还包括:
    第二重构模块,用于将目标终端类型下多个目标终端的多个第四传输特征输入所述目标模型,由所述目标模型基于输入的所述多个第四传输特征,重构每个第四传输特征,输出多个第五传输特征,所述目标终端类型为所述第一终端的终端类型,所述多个第四传输特征与所述多个第五传输特征一一对应,一个第四传输特征为一个目标终端传输的至少一个数据流的总体传输特征;
    第二获取模块,用于基于所述多个第五传输特征与所述多个第四传输特征,获取所述目 标差异度。
  33. 根据权利要求30-32任一项权利要求所述的装置,其特征在于,所述装置还包括:
    第三重构模块,用于将目标终端类型下多个目标终端的多个第六传输特征输入所述目标模型,由所述目标模型基于输入的所述多个第六传输特征,重构每个第六传输特征,输出多个第七传输特征,所述目标终端类型为所述第一终端的终端类型,所述多个第六传输特征与所述多个第七传输特征一一对应,一个第六传输特征为一个目标终端传输的至少一个数据流的总体传输特征;
    所述确定模块,还用于基于所述多个第七传输特征与对应的第六传输特征之间的差异度,确定所述目标模型通过验证。
  34. 一种用于实现终端验证的装置,其特征在于,所述装置包括:
    获取模块,用于获取目标终端类型的至少一个第二终端的多个第三传输特征,一个第三传输特征为一个第二终端传输的至少一个数据流的总体传输特征;
    训练模块,用于将所述多个第三传输特征作为初始模型的输入和输出,进行训练,得到目标模型,所述目标模型用于重构所述目标终端类型的被验证终端的传输特征,以对所述被验证终端进行验证,所述传输特征为所述被验证终端传输的至少一个数据流的总体传输特征。
  35. 一种用于实现终端验证的系统,其特征在于,所述系统包括控制设备和网络设备;
    所述控制设备用于:
    获取目标终端类型的至少一个第二终端的多个第三传输特征,一个第三传输特征为一个第二终端传输的至少一个数据流的总体传输特征;
    将所述多个第三传输特征作为初始模型的输入和输出,进行训练,得到目标模型;
    向所述网络设备发送所述目标模型;
    所述网络设备用于:
    获取所述目标终端类型的第一终端的第一传输特征,所述第一传输特征为所述第一终端传输的至少一个第一数据流的总体传输特征;
    将所述第一传输特征输入所述目标模型,由所述目标模型基于输入的所述第一传输特征,重构所述第一传输特征,输出第二传输特征,所述第二传输特征为重构出的所述第一传输特征;
    若所述第一传输特征与所述第二传输特征之间的差异度大于或等于目标差异度,确定所述第一终端未通过验证。
  36. 一种电子设备,其特征在于,所述电子设备包括处理器和存储器,所述存储器中存储有至少一条程序代码,所述程序代码由所述处理器加载并执行以实现如权利要求1至权利要求22任一项所述的方法所执行的操作。
  37. 一种计算机可读存储介质,其特征在于,所述存储介质中存储有至少一条程序代码,所述程序代码由处理器加载并执行以实现如权利要求1至权利要求22任一项所述的方法所 执行的操作。
PCT/CN2021/105494 2020-07-13 2021-07-09 用于实现终端验证的方法、装置、系统、设备及存储介质 WO2022012429A1 (zh)

Priority Applications (4)

Application Number Priority Date Filing Date Title
JP2023502830A JP2023533354A (ja) 2020-07-13 2021-07-09 端末検証を実現するための方法、装置、システム、デバイス、および記憶媒体
EP21842651.8A EP4171095A4 (en) 2020-07-13 2021-07-09 METHOD FOR IMPLEMENTING TERMINAL DEVICE VERIFICATION, APPARATUS, SYSTEM, APPARATUS AND STORAGE MEDIUM
CA3186107A CA3186107A1 (en) 2020-07-13 2021-07-09 Method, apparatus, system, device, and storage medium for implementing terminal verification
US18/154,263 US20230171264A1 (en) 2020-07-13 2023-01-13 Method, Apparatus, System, Device, and Storage Medium for Implementing Terminal Verification

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
CN202010669766.4 2020-07-13
CN202010669766 2020-07-13
CN202011198953.5 2020-10-31
CN202011198953.5A CN114006714A (zh) 2020-07-13 2020-10-31 用于实现终端验证的方法、装置、系统、设备及存储介质

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US18/154,263 Continuation US20230171264A1 (en) 2020-07-13 2023-01-13 Method, Apparatus, System, Device, and Storage Medium for Implementing Terminal Verification

Publications (1)

Publication Number Publication Date
WO2022012429A1 true WO2022012429A1 (zh) 2022-01-20

Family

ID=79555046

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/105494 WO2022012429A1 (zh) 2020-07-13 2021-07-09 用于实现终端验证的方法、装置、系统、设备及存储介质

Country Status (5)

Country Link
US (1) US20230171264A1 (zh)
EP (1) EP4171095A4 (zh)
JP (1) JP2023533354A (zh)
CA (1) CA3186107A1 (zh)
WO (1) WO2022012429A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114915502A (zh) * 2022-07-15 2022-08-16 北京六方云信息技术有限公司 资产异常行为检测方法、装置、终端设备以及存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20060116968A1 (en) * 2004-11-26 2006-06-01 Shigeru Arisawa Method and system for transmitting electronic value information
CN104410982A (zh) * 2014-11-19 2015-03-11 南京邮电大学 一种无线异构网络中终端聚合与重构方法
CN108683901A (zh) * 2018-05-10 2018-10-19 Oppo广东移动通信有限公司 一种数据处理方法、mec服务器及计算机可读存储介质
CN110769008A (zh) * 2019-11-05 2020-02-07 长沙豆芽文化科技有限公司 一种数据安全防护方法、装置及服务设备
CN111325451A (zh) * 2020-02-02 2020-06-23 贾海芳 智能楼宇多级调度方法、智能楼宇调度中心及系统

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2003088532A1 (en) * 2002-04-11 2003-10-23 The Johns Hopkins University Intrusion detection system for wireless networks
JP4763819B2 (ja) * 2009-05-22 2011-08-31 株式会社バッファロー 無線lanアクセスポイント装置、不正マネジメントフレーム検出方法
KR102000159B1 (ko) * 2013-12-18 2019-07-16 한국전자통신연구원 불법 위장 단말 식별 장치 및 방법
EP3442191B1 (en) * 2017-08-07 2020-09-23 Nokia Solutions and Networks Oy Prevention of identity spoofing in a communications network

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20060116968A1 (en) * 2004-11-26 2006-06-01 Shigeru Arisawa Method and system for transmitting electronic value information
CN104410982A (zh) * 2014-11-19 2015-03-11 南京邮电大学 一种无线异构网络中终端聚合与重构方法
CN108683901A (zh) * 2018-05-10 2018-10-19 Oppo广东移动通信有限公司 一种数据处理方法、mec服务器及计算机可读存储介质
CN110769008A (zh) * 2019-11-05 2020-02-07 长沙豆芽文化科技有限公司 一种数据安全防护方法、装置及服务设备
CN111325451A (zh) * 2020-02-02 2020-06-23 贾海芳 智能楼宇多级调度方法、智能楼宇调度中心及系统

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP4171095A4

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114915502A (zh) * 2022-07-15 2022-08-16 北京六方云信息技术有限公司 资产异常行为检测方法、装置、终端设备以及存储介质
CN114915502B (zh) * 2022-07-15 2022-10-04 北京六方云信息技术有限公司 资产异常行为检测方法、装置、终端设备以及存储介质

Also Published As

Publication number Publication date
US20230171264A1 (en) 2023-06-01
EP4171095A1 (en) 2023-04-26
JP2023533354A (ja) 2023-08-02
CA3186107A1 (en) 2022-01-20
EP4171095A4 (en) 2023-12-27

Similar Documents

Publication Publication Date Title
US20230128061A1 (en) Unsupervised encoder-decoder neural network security event detection
US9769190B2 (en) Methods and apparatus to identify malicious activity in a network
US9369479B2 (en) Detection of malware beaconing activities
CN112235264B (zh) 一种基于深度迁移学习的网络流量识别方法及装置
US20220174008A1 (en) System and method for identifying devices behind network address translators
US11848943B2 (en) Centralized threat intelligence
CN110648180B (zh) 一种调整投放渠道的方法、装置和电子设备
He et al. Deep-feature-based autoencoder network for few-shot malicious traffic detection
US10992972B1 (en) Automatic identification of impermissable account sharing
US10802937B2 (en) High order layer intrusion detection using neural networks
CN110417747A (zh) 一种暴力破解行为的检测方法及装置
WO2022012429A1 (zh) 用于实现终端验证的方法、装置、系统、设备及存储介质
WO2022034405A1 (en) Low-latency identification of network-device properties
Bartos et al. IFS: Intelligent flow sampling for network security–an adaptive approach
WO2023001053A1 (zh) 设备验证的方法、装置和系统
Kim et al. A novel approach to detection of mobile rogue access points
US9455895B2 (en) Data link layer switch frame forwarding analysis
CN110648181B (zh) 基于监测投放效率控制投放渠道的方法、装置和电子设备
CN114006714A (zh) 用于实现终端验证的方法、装置、系统、设备及存储介质
Cai et al. E‐Replacement: Efficient scanner data collection method in P4‐based software‐defined networks
Xu et al. Supervised learning framework for covert channel detection in LTE‐A
US11343241B2 (en) Multi-connectivity communication
US12001584B2 (en) Privacy-preserving contact tracing
WO2023098222A1 (zh) 多业务场景的识别方法和决策森林模型的训练方法
CN111049944B (zh) 一种id发现方法和装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21842651

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2023502830

Country of ref document: JP

Kind code of ref document: A

Ref document number: 3186107

Country of ref document: CA

ENP Entry into the national phase

Ref document number: 2021842651

Country of ref document: EP

Effective date: 20230118

NENP Non-entry into the national phase

Ref country code: DE